affiliate-onboarding-sequence

{
  "skill_name": "affiliate-onboarding-sequence",
  "evals": [
    {
      "id": 1,
      "prompt": "I run the affiliate program at Lumaflor, a DTC haircare brand. We approve around 400 affiliates a month - mostly coupon sites, a few Instagram creators, and some newsletter people. Right now the only thing they get is an approval email saying 'welcome, you're in.' My VA generates their tracking link and coupon code by hand and emails it over, usually 2 or 3 business days later. I want a proper welcome email series, and I'd like white-glove treatment: a personal call with each new affiliate in week one. I don't have anyone dedicated to the program, it's me plus the VA. Can you write me the sequence?",
      "expected_output": "A sequence plan that first fixes day-0 provisioning and pins an activation definition, removes the requested per-affiliate call tier outright given no AM and 400 approvals a month, routes the whole intake self-serve with per-archetype first actions, and delivers the SEQUENCE PLAN artifact with a behavior-trigger layer.",
      "files": [],
      "expectations": [
        "Fixes the activation event and its measurement window (for example first tracked sale within 30 days of approval) before designing any touch",
        "Flags the 2-3 business day manual link and code delivery as a day-0 provisioning failure that must be fixed before the sequence launches, not as a touch inside the sequence",
        "Contrasts auto-provisioning at approval (affiliate ready to promote in under an hour) with manual provisioning taking 1-3 business days",
        "Declines the requested personal call with every new affiliate",
        "Removes the high-touch tier from the menu entirely rather than listing it as a lower-priority or later option",
        "Justifies that removal by the absence of a named affiliate manager and the roughly 400 approvals per month volume",
        "Routes the whole intake to a self-serve tier (portal plus automated drip) as the floor",
        "States the tier efficiency ordering explicitly, with self-serve ahead of hybrid ahead of high-touch on activation bought per manager hour",
        "Notes that self-serve leads on that ratio despite converting worst per affiliate, not because it converts best",
        "Places the tracking link and coupon code in the first lines of touch 1",
        "Keeps touch 1 to one action and roughly 300 words or fewer",
        "Gives coupon-deal, social creator, and email-list owner archetypes different first actions and leading assets, without ranking those archetypes against each other",
        "Delivers a SEQUENCE PLAN artifact carrying at least Activation, Routing, Touches, Activation nudge, Never-activated, Exits, and KPIs lines"
      ]
    },
    {
      "id": 2,
      "prompt": "Our affiliate program shows 4% activation and my CMO is furious. Program's been live 8 months, roughly 90 approvals a month, B2C home goods. I've been told the welcome email is the problem - it's honestly pretty bland. Can you rewrite it and the two follow-ups? We're on a standard affiliate platform with cookie-based tracking only, no server-side postback. Affiliates do show clicks in the dashboard though.",
      "expected_output": "A refusal to rewrite copy first, replaced by an ordered diagnosis that starts with an end-to-end conversion trace (justified by clicks-visible / activation-low plus cookie-only tracking), then the approval-email check, per-touch engagement, cohort split and nudge/branch audit, with the rewrite placed last.",
      "files": [],
      "expectations": [
        "Refuses to rewrite the welcome email as the first action and says so explicitly",
        "Gives the reason that a rewrite is a change rather than a check, and that it destroys the baseline that would have shown whether it worked",
        "Starts with an end-to-end trace of one conversion: click, recorded conversion, pending commission, reversal state",
        "Ties that first step to the stated symptom of clicks visible in the dashboard while activation reports low, as a sign the reported number may be fiction",
        "States the diagnosis ordering as information bought per unit of effort rather than cheapest check first",
        "Includes opening the approval email as a newly approved affiliate to confirm a working link or code is present, described as near-zero effort",
        "Includes reading the per-touch open and click rates the platform already records, and notes this localizes where the sequence loses people without naming the cause",
        "Includes splitting activation by archetype and by tier as a separate check",
        "Includes auditing whether a deadline-bound nudge inside the window and a designed never-activated branch exist",
        "Flags cookie-only attribution as undercounting because of browser tracking prevention, and recommends server-side postback carrying a click ID and an order or transaction ID",
        "Places the rewrite of touch 1 or the asset kit last, only after the preceding checks have run",
        "Does not conclude that recruitment or affiliate quality is the cause"
      ]
    },
    {
      "id": 3,
      "prompt": "Draft the day-45 email for affiliates in our program who still haven't sent a single click. I want it firm - something like 'accounts with no activity within 60 days of approval will be removed from the program, so get your first link out now.' We're a B2C pet supplies store, 30-day activation window, and honestly the dead weight in the account list is making our numbers look terrible. Also tell me how many times to chase them and when I should just delete these accounts.",
      "expected_output": "A rejection of the termination-threat wording on motivation grounds, replaced with an opportunity-framed deadline incentive, plus a never-activated branch that diagnoses first, re-engages once or twice, sends a housekeeping-framed final notice at roughly six months, and archives rather than deletes.",
      "files": [],
      "expectations": [
        "Rejects the proposed removal-threat wording and does not produce a version of it",
        "Gives the reason as motivation research finding threats counterproductive, rather than as a tone or brand preference",
        "Reframes the message as an opportunity with a finish line, for example an incentive that expires on a stated date",
        "Recommends archiving or suppressing never-activated accounts rather than deleting them",
        "States that suppression keeps the active-rate metric honest and leaves the door open",
        "Places the final-notice email far later than the proposed 60 days, at roughly six months of zero clicks and zero replies",
        "Frames that final notice as housekeeping rather than punishment",
        "Separates never-activated affiliates (never clicked) from dormant ones (were active, then stopped) and states they get different messages",
        "Limits re-engagement to roughly one or two touches before the final notice",
        "Puts a diagnostic checkpoint at the end of the activation window, before any re-engagement message is sent",
        "Equips the re-engagement touch with one concrete incentive and one ready-to-use asset, not encouragement alone"
      ]
    },
    {
      "id": 4,
      "prompt": "We sell an enterprise data-catalog platform, average sales cycle is about five months, ACV around $60k. We just opened a partner referral program - 18 approvals last month, a mix of boutique data consultancies and two systems integrators. I've got one partner manager who can do calls. I've written a 14-day welcome drip and I'm going to measure activation as 'first referral within 30 days of approval.' Sanity check my plan and fill in the gaps?",
      "expected_output": "A correction of the 30-day window to roughly 90 days on B2B deal physics, a deep-funnel activation event associated to the partner in CRM, day-0 sandbox and deal-registration provisioning, a trigger layer stretched across the window, and hybrid/high-touch routing with a kickoff-call agenda.",
      "files": [],
      "expectations": [
        "Rejects the 30-day activation window for this program and recommends roughly 90 days instead",
        "Ties that window to the stated five-month sales cycle rather than to a generic default",
        "Requires the activation event to be pinned to one deep-funnel milestone (lead, trial, accepted demo, qualified opportunity, or closed-won) rather than left as 'first referral'",
        "Requires the chosen conversion event to be recorded and associated to the partner in the CRM",
        "Keeps a 4-6 touch, 10-14 day calendar skeleton but stretches the behavior-trigger layer across the full 90-day window",
        "Names at least two B2B-appropriate behavior triggers, such as training or certification completed, or no deal registered by a given day",
        "Requires sandbox, demo, or NFR access to be provisioned at approval rather than granted on request",
        "Requires the deal-registration link to be available at day 0",
        "Recommends a hybrid or high-touch tier as viable given roughly 18 approvals a month and a named partner manager with call capacity",
        "Includes a kickoff-call agenda covering the concrete activation plan, a first-deal date, and next-step owners",
        "Adds B2B-specific kit items such as competitor-comparison or buyer-question material, beyond logos, banners and swipe copy",
        "Notes that recurring commission carries clawback or reversal risk on every invoice, so payout holds run longer than in B2C",
        "Places the time-boxed activation nudge inside the 90-day window rather than at day 30"
      ]
    },
    {
      "id": 5,
      "prompt": "Quick one. Our 30-day affiliate activation rate came in at 19% for the August cohort - B2C, fashion accessories. My head of growth pulled a number off a blog saying 30% is the industry standard and now wants to pause the program and go re-recruit better affiliates. Is 19% actually bad? What should I tell her, and what number should we be holding ourselves to going forward? Also our all-time activation rate across every affiliate we've ever approved is 11%, which looks worse - which one do I report?",
      "expected_output": "A verdict that 19% sits at or above the ~18% median B2C gate, a correction that the 30% figure is a diagnostic trigger rather than a pass gate, an honest reliability label on all these vendor figures, a rejection of re-recruiting as the response, and a switch to monthly-cohort reporting.",
      "files": [],
      "expectations": [
        "States that 19% sits at or above the roughly 18% median 30-day B2C activation figure, so the cohort is not failing",
        "Identifies the 30% figure as a diagnostic trigger for investigation, not a pass or ship gate",
        "Notes that holding the program to roughly 30% on day one demands top-quartile performance",
        "Labels the benchmark numbers as vendor or practitioner figures rather than audited studies",
        "Gives the wider context that only roughly 5-20% of approved affiliates ever produce a tracked conversion",
        "Rejects re-recruiting as the response and directs any diagnosis at the sequence itself (provisioning, first touch, nudge)",
        "Recommends reporting by monthly approval cohort rather than the blended all-time rate",
        "Explains that a blended all-time rate hides whether the sequence is improving",
        "States that a cohort is judged only once its activation window has closed",
        "Names time-to-first-conversion, tracked as a cohort median, as a KPI alongside activation rate",
        "Names never-activated share of the cohort at the cutoff as a KPI",
        "Does not present any benchmark figure as an industry standard the user can quote as established fact"
      ]
    },
    {
      "id": 6,
      "prompt": "Setting up partner onboarding for Reyvant, we sell a clinical trial management suite. Our whole partner roster is 14 CROs and consultancies - we added 3 this quarter and expect maybe 6 more next year. Each one that lands a deal is worth north of $200k in first-year contract value. I have a partner manager with open calendar time. I was going to copy the onboarding drip our marketing team uses for trial signups: build out the automated email sequence, the portal, the self-serve asset library, the works. What should the sequence look like?",
      "expected_output": "An inversion of the default tier ordering - high-touch as primary because near-zero volume makes the per-partner hour cheap and the drip's build never pays back - with a kickoff-call agenda, day-0 provisioning kept as a pre-condition, and the plan delivered as the SEQUENCE PLAN artifact.",
      "files": [],
      "expectations": [
        "Inverts the default tier ordering and recommends high-touch as the primary tier for this roster",
        "Gives the reason as near-zero partner volume making the per-partner manager hour cheap while the self-serve drip's build never pays back",
        "Declines to build the full automated drip, portal and self-serve asset library as the centerpiece of the plan",
        "States explicitly that the default efficiency ordering (self-serve ahead of hybrid ahead of high-touch) is a default to re-rank against the program, not a rule",
        "Notes that high-touch requires almost no upfront build but becomes a standing recurring job",
        "Includes a kickoff-call agenda covering terms and commission recap, targets with a concrete activation plan and next-step owners, a product walkthrough with sandbox access, and a first-deal date",
        "Keeps day-0 provisioning as a pre-condition: portal credentials, deal-registration link, and sandbox or NFR access at approval",
        "Keeps terms and disclosure attestation, plus tax and payout collection gating payout eligibility, at onboarding",
        "Keeps the one-action-per-touch rule even inside the high-touch plan",
        "Defines the activation event and its window before laying out the touch plan",
        "Delivers the plan in the SEQUENCE PLAN artifact format rather than as an undifferentiated list of emails"
      ]
    },
    {
      "id": 7,
      "prompt": "Looking at our affiliate onboarding numbers. Touch 1 gets 61% opens and 28% clicks, touch 2 gets 54% opens, touch 3 gets 9% opens and almost no clicks, touch 4 and 5 are back around 40%. Touch 3 is the 'how do you plan to promote?' check-in and it goes out from our program@ address. I want to scrap the whole 5-email drip and replace it with a single onboarding hub page - recorded walkthrough video, all assets in one place, fully self-serve. Our dev team can get it built this quarter. Thoughts?",
      "expected_output": "A rejection of the hub swap because opens recover at touches 4 and 5, a copy-plane diagnosis of touch 3 pinned to the sender mismatch on a reply-seeking email, the trim-beats-hub ordering with its effort and reversibility numbers, and the condition under which a hub would actually be right.",
      "files": [],
      "expectations": [
        "Rejects replacing the full drip with a hub for this engagement pattern",
        "Points out that opens recover at touches 4 and 5, so touch volume itself is not the friction",
        "Diagnoses touch 3 as a copy-plane problem rather than a structural one",
        "Identifies the sender mismatch: a touch asking for a reply is sent from an automated program address instead of a named person",
        "States the ordering that trimming the drip beats swapping in a hub on activation bought per unit of build",
        "Sizes the hub as roughly a quarter of build plus a standing job to keep current, against roughly an hour to trim",
        "Notes that reverting a hub means rebuilding the drip that was deleted, so the swap is far less reversible",
        "States the condition under which the hub would be right: opens decaying across the whole sequence, not one touch underperforming",
        "If a hub is adopted at all, keeps touch 1 and the behavior-trigger layer and collapses only the middle touches",
        "Keeps the one-action, one-CTA rule for the rewritten touch 3",
        "Does not recommend a rebuild before reading what the existing per-touch data already says"
      ]
    },
    {
      "id": 8,
      "prompt": "I'm building the affiliate activation emails for Hoymar (B2C outdoor gear, 30-day window, 60-day cookie, coupon and content affiliates). My boss wants the sequence structured on AIDA so it maps to how the rest of our marketing is built, and he also forwarded a guide referencing something called the '60/40 principle' for affiliate onboarding that he wants applied. Can you build it out on those two?",
      "expected_output": "A refusal to force-fit AIDA and a refusal to apply the unverified 60/40 principle, replaced by the day-offset skeleton plus behavior-trigger layer, with any named re-engagement framework labelled as one practitioner's coinage.",
      "files": [],
      "expectations": [
        "Declines to structure the sequence on AIDA and states that general marketing funnels are not built for affiliate activation",
        "Declines to apply the '60/40 principle' and states it is unverified, with no named originator or citation behind it",
        "Does not invent content or a definition for the 60/40 principle",
        "Offers a day-offset skeleton plus a behavior-trigger layer as the structure instead",
        "States the skeleton as roughly 4-6 touches over 10-14 days, each asking one action",
        "Presents that skeleton as a defensible starting point rather than a fixed industry calendar",
        "States that a behavior trigger overrides the calendar touch it collides with",
        "If a named framework is used for the re-engagement branch, labels it as one practitioner's coinage rather than an industry-standard method",
        "Keeps the activation window at 30 days and puts a deadline-bound nudge inside it",
        "Does not present RACE, AARRR or AIDA as an accepted affiliate-onboarding framework"
      ]
    },
    {
      "id": 9,
      "prompt": "Two messes in our affiliate onboarding automation at Bellfort (B2C kitchenware). First, people who already made a sale on day 3 keep getting the rest of the welcome drip, including the day-21 'make your first sale' bonus email - one of them replied 'do you people not see my orders?' Second, we're reactivating about 40 affiliates out of the archive next month for the holiday push and I don't know whether to drop them back at the top of the sequence. Oh, and a few affiliates are mid-thread with me over email while the automation keeps firing at them. How do I fix the flow control?",
      "expected_output": "Flow-control rules: exit on activation with a celebration touch, re-entry at the tools touch rather than the welcome touch, suppression while a reply thread is open, a trigger-overrides-calendar rule, branch-entry exit, and an explicit flag on transactional-versus-marketing unsubscribe handling.",
      "files": [],
      "expectations": [
        "Exits an affiliate from the calendar sequence the moment the activation event tracks",
        "Fires a distinct celebration touch on first sale and hands over to ongoing program communications",
        "States that continuing the onboarding drip past activation reads as unattended automation",
        "Has the first-sale celebration state when the commission approves and when it pays",
        "Re-enters reactivated archived affiliates at the tools touch (T2), not at the welcome touch (T1)",
        "Gives the reason that reactivated affiliates need a working asset and one action, not re-approval",
        "Suppresses calendar touches while a reply thread with the affiliate manager is open",
        "Specifies that entry into the never-activated branch ends the main sequence",
        "Adds a trigger-overrides-calendar rule: when a behavior trigger fires within a day of a scheduled calendar touch, send the trigger touch and skip or delay the calendar one",
        "Gives the reason that two emails in one day from the same program trains the skim reflex",
        "Honors unsubscribes program-wide",
        "Flags whether transactional notices (commission approved, payout sent) may still send after a marketing unsubscribe as a question for the program's terms and local law, rather than assuming they may"
      ]
    },
    {
      "id": 10,
      "prompt": "I've got 19 days before our next affiliate cohort gets approved and I need to fix what I can before then. Known problems, all real: the tracking link sits below a 400-word block of brand guidelines in our welcome email; that same email also attaches the terms PDF, 14 banners and a certification invite; there's no follow-up at all for people who never click; we send one identical sequence to coupon sites, cashback sites and Instagram creators alike; and we've been paying commissions for four months without ever collecting a W-9 or payout details from anyone. No budget for design work and nobody to add to the team. What order do I do these in?",
      "expected_output": "An explicit fix ordering by activation bought per hour, led by link placement then the welcome dump, with the tax and payout gap pulled out of that ordering onto a compliance clock, the routing-versus-branch tie broken by which window each pays in, and design work deleted rather than ranked last.",
      "files": [],
      "expectations": [
        "Puts moving the tracking link into the first lines of touch 1 first, as a near-zero-effort fix",
        "Puts breaking up the welcome dump into one action per touch second",
        "States the fix ordering as activation bought per hour of fixing, explicitly, not as an implied row order",
        "Pulls the missing tax and payout collection out of that activation ordering entirely",
        "Justifies that separation because it buys avoided payout disputes and reduced legal exposure rather than activation",
        "Tells the user to fix the tax and payout gap on the compliance clock, before further commissions are earned, whatever the activation work is doing",
        "Notes that commissions already earned against uncollected tax forms become a negotiation with the affiliate rather than a build task",
        "Treats archetype routing and the never-activated branch as roughly equal on ratio, each about a week of work",
        "Breaks that tie on timing: routing pays inside the window the user is already in, while the never-activated branch only pays on the next cohort",
        "Uses the 19-day deadline to promote the near-zero fixes and demote or drop anything requiring a build",
        "Deletes the banner and design work from the list outright given the stated no-design-budget constraint, rather than ranking it last",
        "Does not delay the cohort's launch until a complete asset kit exists"
      ]
    },
    {
      "id": 11,
      "prompt": "Write me the actual welcome email for Norrsken Supply's affiliate program - B2C, we sell camping gear. 15% commission on the order total, 45-day cookie, $75 payout threshold, commissions approve after our 30-day return window closes and pay net-30. Links and codes are auto-generated at approval. The affiliate I'm writing to is a newsletter writer with about 20k subscribers. I want it to feel personal and I want them to reply to me - I'm Tomas, I run the program. Send it over and I'll paste it straight into our ESP.",
      "expected_output": "A touch-1 email under 300 words with the link and code in the first lines, one action and one CTA, all money mechanics stated plainly, a concrete subject line, the named human as sender, disclosure and terms surfaced, and an explicit instruction that the draft goes through a humanizer pass before it is treated as final.",
      "files": [],
      "expectations": [
        "Places the tracking link and coupon code in the first lines of the email",
        "Asks exactly one action, with one call to action",
        "Keeps the email under roughly 300 words",
        "Writes the call to action as action plus outcome rather than 'learn more' or equivalent",
        "States the money mechanics plainly: the 15% commission, the 45-day cookie, the $75 payout threshold, approval after the 30-day return window, and net-30 payment",
        "Sets the subject line on the concrete thing inside the email (the link or code) rather than on program excitement such as 'Welcome aboard'",
        "Sends the email from Tomas by name rather than from an automated or no-reply program address",
        "Notes that a reply-seeking email sent from a no-reply or automated address undercuts its own ask",
        "Uses the affiliate's own archetype context, for example their subscribers or their newsletter, rather than the program's framing",
        "Surfaces the disclosure obligation and terms acceptance in this first touch",
        "States that the draft must go through a humanizer pass before it is treated as final copy, rather than handing over raw first-draft output as paste-ready",
        "Specifies that the humanizer pass rewrites tone and phrasing only, leaving the one-action structure and the money mechanics unchanged"
      ]
    }
  ],
  "trigger_queries": [
    { "query": "Design our affiliate onboarding sequence", "should_trigger": true },
    { "query": "Write the welcome emails for affiliates we just approved", "should_trigger": true },
    { "query": "What should new partners get in their first 30 days?", "should_trigger": true },
    { "query": "Our approved affiliates never send a click", "should_trigger": true },
    { "query": "nobody activates after the welcome email", "should_trigger": true },
    { "query": "We approve about 40 affiliates a month and almost none of them ever send traffic", "should_trigger": true },
    { "query": "Map the first 30 days for a newly approved B2B SaaS affiliate", "should_trigger": true },
    { "query": "Design the never-activated branch for our affiliate program", "should_trigger": true },
    { "query": "affiliate welcome email series", "should_trigger": true },
    { "query": "affiliate activation emails", "should_trigger": true },
    { "query": "new affiliate ramp plan", "should_trigger": true },
    { "query": "partner activation emails", "should_trigger": true },
    { "query": "How do I get affiliates who signed up to actually start promoting?", "should_trigger": true },
    { "query": "We've got 300 approved affiliates sitting there doing nothing. What do I send them?", "should_trigger": true },
    { "query": "What's a good drip for people who just got accepted into our partner program?", "should_trigger": true },
    { "query": "how many emails should a new affiliate get and over how long", "should_trigger": true },
    { "query": "Should we put an expiring bonus in front of new affiliates in their first month?", "should_trigger": true },
    { "query": "when do we stop chasing affiliates who never produce", "should_trigger": true },
    { "query": "Time-boxed incentive for a first sale in the first 30 days - how do I structure the touches around it?", "should_trigger": true },
    { "query": "Our partner managers sign people up and then nothing happens for months. Fix the handoff.", "should_trigger": true },
    { "query": "Build the ramp plan from approval to first tracked conversion", "should_trigger": true },
    { "query": "I need an email sequence for people we just let into our commission program", "should_trigger": true },
    { "query": "what goes in the asset kit a new affiliate gets on day one", "should_trigger": true },
    { "query": "Should the welcome email contain everything or should I split it up?", "should_trigger": true },
    { "query": "our first affiliate email has the link at the bottom, does that matter", "should_trigger": true },
    { "query": "Kickoff call agenda for a new B2B referral partner", "should_trigger": true },
    { "query": "We just approved 20 consultancies into the partner program. What happens next week?", "should_trigger": true },
    { "query": "Draft the day-45 follow-up for affiliates with zero clicks", "should_trigger": true },
    { "query": "How long do I wait before archiving an affiliate who never promoted?", "should_trigger": true },
    { "query": "affiliate activation rate is 6%, what do I change", "should_trigger": true },
    { "query": "Our program has a 10% activation rate. Is that normal, and what fixes it?", "should_trigger": true },
    { "query": "Design the touch plan and behavior triggers for new partner onboarding", "should_trigger": true },
    { "query": "what should trigger the second email to a new affiliate - a date, or something they did?", "should_trigger": true },
    { "query": "I want a sequence that reacts to first click and first sale instead of just firing on a calendar", "should_trigger": true },
    { "query": "Set up the flow control so affiliates who already sold stop getting the welcome drip", "should_trigger": true },
    { "query": "Do reactivated affiliates start the sequence over from the top?", "should_trigger": true },
    { "query": "we're bringing 40 dormant affiliates back for Q4, what do we send them", "should_trigger": true },
    { "query": "onboarding plan for newly signed referral partners", "should_trigger": true },
    { "query": "The first thing our new affiliates get is a 14-banner zip file. Is that a problem?", "should_trigger": true },
    { "query": "Sanity check my 14-day affiliate welcome drip", "should_trigger": true },
    { "query": "How do I split the sequence between coupon sites and content reviewers?", "should_trigger": true },
    { "query": "Different first email for cashback partners vs Instagram creators?", "should_trigger": true },
    { "query": "Should I give every new partner a personal call, or automate it?", "should_trigger": true },
    { "query": "self-serve vs high touch for new partner ramp", "should_trigger": true },
    { "query": "we have 12 partners each worth six figures - is an automated drip the right call for onboarding them?", "should_trigger": true },
    { "query": "When should I collect tax forms from affiliates?", "should_trigger": true },
    { "query": "Do I ask for the W-9 at signup or after they earn something?", "should_trigger": true },
    { "query": "we started paying commissions and now we can't get payout details out of anyone", "should_trigger": true },
    { "query": "Our affiliate links take three days to issue by hand. How much is that costing us?", "should_trigger": true },
    { "query": "Write the first-sale celebration email for a new affiliate", "should_trigger": true },
    { "query": "Is it OK to tell inactive affiliates we'll remove them if they don't produce?", "should_trigger": true },
    { "query": "Threatening to kick out inactive affiliates - good idea or not?", "should_trigger": true },
    { "query": "Replace our 5-email partner drip with a single onboarding hub page?", "should_trigger": true },
    { "query": "How do I know whether the problem is my emails or my tracking?", "should_trigger": true },
    { "query": "clicks look fine in the dashboard but activation is reported at 3%", "should_trigger": true },
    { "query": "What KPIs tell me whether my new-affiliate emails are working?", "should_trigger": true },
    { "query": "Should I report affiliate activation by monthly cohort or all-time?", "should_trigger": true },
    { "query": "Our board wants 30% activation in the first 30 days. Is that achievable?", "should_trigger": true },
    { "query": "B2B partner first referral usually takes months - how do I set the window?", "should_trigger": true },
    { "query": "sequence plan for taking an approved affiliate to their first conversion", "should_trigger": true },
    { "query": "Just approved a bunch of creators into our program. What's the plan for week one?", "should_trigger": true },
    { "query": "Write the cold email we send to bloggers to get them to join our affiliate program", "should_trigger": false },
    { "query": "How do I find and recruit affiliates in the pet niche?", "should_trigger": false },
    { "query": "Draft the pitch and follow-up sequence for recruiting new affiliates", "should_trigger": false },
    { "query": "Build a prospect list of content sites to invite into our program", "should_trigger": false },
    { "query": "Breakup email for affiliate prospects who never replied to our invite", "should_trigger": false },
    { "query": "What commission rate should we pay affiliates on a $49/month SaaS plan?", "should_trigger": false },
    { "query": "Revenue share vs flat CPA for our affiliate program", "should_trigger": false },
    { "query": "How long should the recurring commission last - 12 months or lifetime?", "should_trigger": false },
    { "query": "Set our cookie window and clawback period", "should_trigger": false },
    { "query": "Design a tiered commission ladder that pays more above 20 sales a month", "should_trigger": false },
    { "query": "Draft the affiliate program terms and conditions", "should_trigger": false },
    { "query": "Write the prohibited-tactics clause banning brand bidding", "should_trigger": false },
    { "query": "What clauses do I need in an affiliate agreement for EU partners?", "should_trigger": false },
    { "query": "Write the termination clause for our affiliate contract", "should_trigger": false },
    { "query": "Check this Instagram caption for FTC disclosure compliance", "should_trigger": false },
    { "query": "Does '#ad' at the end of a YouTube description meet the rules?", "should_trigger": false },
    { "query": "Review this affiliate blog post against UK CAP rules before it goes live", "should_trigger": false },
    { "query": "Build the rules that flag cookie stuffing in our affiliate data", "should_trigger": false },
    { "query": "One affiliate's conversion rate is 40% - is this fraud?", "should_trigger": false },
    { "query": "How do I tell real affiliate fraud from low-incrementality coupon traffic?", "should_trigger": false },
    { "query": "Audit this month's affiliate commission payout file before we send money", "should_trigger": false },
    { "query": "Check our payout run for double-counted conversions", "should_trigger": false },
    { "query": "What metrics belong on an affiliate performance dashboard?", "should_trigger": false },
    { "query": "Define the formula for affiliate-attributed revenue in our reporting", "should_trigger": false },
    { "query": "Set alert thresholds for when a top affiliate's traffic drops", "should_trigger": false },
    { "query": "Build the partner certification curriculum", "should_trigger": false },
    { "query": "What training modules do our resellers need before they can sell?", "should_trigger": false },
    { "query": "Design the enablement tracks for solution architects at our SI partners", "should_trigger": false },
    { "query": "Design silver/gold/platinum tiers for our partner program", "should_trigger": false },
    { "query": "What criteria promote a partner from registered to gold?", "should_trigger": false },
    { "query": "Build the per-partner scorecard and QBR structure", "should_trigger": false },
    { "query": "Underperformance ladder for partners missing targets two quarters running", "should_trigger": false },
    { "query": "How much should we give customers for referring a friend?", "should_trigger": false },
    { "query": "Cash or account credit for our refer-a-friend rewards?", "should_trigger": false },
    { "query": "Stop customers from creating fake accounts to farm referral rewards", "should_trigger": false },
    { "query": "Set velocity limits and reward caps on our referral program", "should_trigger": false },
    { "query": "Write the campaign brief for a creator we just signed", "should_trigger": false },
    { "query": "Negotiate usage rights and exclusivity with an influencer", "should_trigger": false },
    { "query": "Shortlist criteria for finding creators in the skincare niche", "should_trigger": false },
    { "query": "First DM to a creator we want to work with", "should_trigger": false },
    { "query": "Set the KPIs for a one-off influencer campaign before it launches", "should_trigger": false },
    { "query": "Write the deal registration policy for our co-sell motion", "should_trigger": false },
    { "query": "Rules of engagement between our direct reps and channel partners", "should_trigger": false },
    { "query": "Plan the joint go-to-market launch with our new integration partner", "should_trigger": false },
    { "query": "Model the P&L of our largest reseller relationship", "should_trigger": false },
    { "query": "Should we list on AWS Marketplace or Azure first?", "should_trigger": false },
    { "query": "Map our whole partner ecosystem and find the coverage gaps", "should_trigger": false },
    { "query": "Rank these five alliance candidates by expected value", "should_trigger": false },
    { "query": "Which partner categories should we add next, and in what order?", "should_trigger": false },
    { "query": "Design our B2B partner program from scratch - tiers, motions, benefits", "should_trigger": false },
    { "query": "I'm starting partnerships work at a new company, where do I begin?", "should_trigger": false },
    { "query": "What does a partner manager get asked in interviews?", "should_trigger": false },
    { "query": "Write the job description for our first affiliate manager hire", "should_trigger": false },
    { "query": "Improve our SaaS user onboarding flow so more trials convert", "should_trigger": false },
    { "query": "Write the welcome email series for new newsletter subscribers", "should_trigger": false },
    { "query": "Design the in-app onboarding checklist for first-time users", "should_trigger": false },
    { "query": "Onboarding CRO - where are we losing people in signup?", "should_trigger": false },
    { "query": "New employee onboarding plan for their first 30 days", "should_trigger": false },
    { "query": "Set up a drip campaign for free trial users who haven't logged in", "should_trigger": false },
    { "query": "Shopify merchant onboarding walkthrough", "should_trigger": false },
    { "query": "Onboard a new customer onto our platform after they sign the contract", "should_trigger": false }
  ]
}

# Asset Kit and Day-0 Provisioning

The pre-condition checklist: what must exist before the first touch sends. Vendor names appear only as sourced evidence, never as required tools.

## Day-0 provisioning checklist (both B2B and B2C)

Verify each item fires automatically at approval, or has a named owner with a same-day SLA:

- [ ] Portal/dashboard credentials issued.
- [ ] Tracking link generated (deep links + sub-ID/pass-through parameter conventions documented).
- [ ] Unique coupon/promo code issued, where the program uses codes.
- [ ] Asset library access granted (one persistent hub beats attachments).
- [ ] Terms acceptance + disclosure attestation captured (drafting the terms is the terms skill's job; capturing acceptance here is this skill's).
- [ ] Tax form and payout method collected, gating payout eligibility (below).
- [ ] Archetype + tier tag applied, so the routing fires the right sequence.

Speed matters: with auto-approval and auto-provisioning, an affiliate can be ready to promote in under an hour; manual provisioning is documented at 1-3 business days per affiliate - dead time inside the highest-intent window.

## Tracking readiness

If the activation event cannot be recorded, the sequence measures nothing:

- Server-side (postback) conversion tracking live, carrying a click ID and an order/transaction ID for deduplication. Cookie-only attribution undercounts: browser tracking prevention (per WebKit's published ITP documentation) caps JavaScript-set landing-page cookies at 24 hours under link decoration and clears script-writeable storage after 7 days without interaction.
- Trace the full path once, manually: click → recorded conversion → pending commission → refund/reversal state. If any hop is dark, reported activation understates reality and the KPI block is fiction.
- Coupon-code attribution tested where codes exist - a code redeemed without a click must still credit the affiliate.
- B2B: conversion events defined deep-funnel (trial, qualified opportunity, closed-won) and associated to the partner in the CRM; long multi-session cycles make cookie attribution least reliable exactly where deals are largest.

## Asset kit contents

**B2C ecommerce kit** - built for placement volume:

- Pre-coded tracking links and the unique coupon code.
- Banner set, product images, logo pack with usage rules.
- Swipe copy: 1-2 ready-cut emails, short social captions.
- Product one-sheet: positioning, top sellers, seasonal angles.
- Product feed access (loyalty/cashback and sub-network archetypes), with feed freshness and deep-link format documented.
- Disclosure requirement stated plainly next to the assets, not only in the terms.

**B2B SaaS kit** - built for depth per partner; one published program comparison (Rewardful) notes a standard kit of "logos, an approved tagline, three hero images, and a talking points document" is not enough for B2B, which needs:

- Everything above, plus:
- Sandbox, demo, or NFR (not-for-resale) account access - provisioned day 0, not on request.
- Use-case angles, competitor-comparison material, buyer-question material - fuel for long-form reviews and comparison posts.
- Sales deck and demo script for agency/consultant archetypes.
- Deal-registration link and its conflict rules.
- Pointer to the certification module when one exists (the curriculum itself belongs to the enablement skill).

## Sequencing asset creation

When interview question 5 reveals missing assets, don't delay launch for the full kit. Build in this order:

- efficiency: `link/code provisioning > attestation and tax collection > one leading asset per present archetype > banners, feeds, decks`
- effort: `banners, feeds, decks > leading asset per archetype > attestation and tax collection == link/code provisioning`
- compliance cost: `attestation and tax collection > everything else in the kit`

1. **Blocks touch 1** - working link/code provisioning, terms + disclosure attestation flow, tax/payout collection. Without these the sequence cannot start at all, which is why they lead on efficiency despite costing about a week each when the platform doesn't provide them (an hour when it auto-provisions). That identical shape - a week to wire by hand, an hour when the platform ships it - is what makes provisioning and attestation tie on effort; only the compliance axis separates them.
2. **Blocks touch 2** - the one leading asset per archetype actually present in the intake (swipe email, code page, caption pack, sandbox), a day or so each. Skip archetypes with zero approved affiliates - that is a deletion from the list, not a low-priority row.
3. **Everything else** - banners, feeds, decks, certification pointers, a week or more each and never finished - ships when ready, announced by a later touch or the asset hub. A late banner set costs little; a late link costs the whole window.

The compliance axis separates rung 1 from the rest of the kit, and it is why attestation and tax collection cannot be deferred into rung 2 even though a banner set costs more hours. The disclosure obligation and the tax-form requirement each need a sign-off you cannot apply retroactively: once a commission has been earned against an unattested placement or an uncollected tax form, the fix is a negotiation with the affiliate rather than a build task. Banners, swipe copy and decks carry no such exposure and stay fully reversible.

## Payout and tax gating

Collect payment and tax setup at onboarding, as a gate on payout eligibility - not after the first commission is earned:

- Tax form (per the affiliate's jurisdiction) and payout method captured inside the onboarding flow; payout threshold and hold period stated in touch 1.
- Chasing tax forms after paying is a documented recurring operational mistake in practitioner material; the reverse order costs nothing at onboarding and blocks nothing but payout.
- Precedent that unclaimed setups expire: one large program's published implementation voids commissions if the payment account isn't completed within 120 days of the first commission (ClickFunnels, via its payout provider) - tell affiliates the rule upfront rather than surprising them.
- Match commission state to risk: create commissions as pending, approve after the refund/clawback window closes. B2C risk ends with the return window; B2B recurring commissions carry reversal risk on every invoice, so holds run longer.
- Regulatory and tax specifics change - verify current forms and thresholds against the program's platform documentation and counsel; do not hardcode form names into the sequence copy.

## Compliance placement

Surface the disclosure obligation and restricted-terms policy (e.g. brand bidding) inside touch 1, alongside terms acceptance. No regulator or network prescribes a specific touch number or day for this, so the exact placement is a design choice, not an industry rule - but the FTC's 2023 Endorsement Guides update do require the brand running the affiliate program to train affiliates on disclosure requirements and provide compliant disclosure language before they publish any content, which puts a real compliance floor under placing it no later than touch 1, at approval.

# Benchmarks and Frameworks - with Reliability Flags

Read this before quoting any number or framework to the user. Every figure below carries its source quality; never present a flagged number as measured fact, and never invent one that isn't here.

## Activation benchmarks

**Practitioner-convergent, not controlled measurement** (multiple independent practitioners and vendors say roughly the same thing, each from their own book of programs):

- 5-20% of approved affiliates ever produce a tracked conversion; ~10% is the commonly cited average. Matt McWilliams: in a typical program "95% of affiliates who sign up never make a single sale."
- Revenue concentration: "as much as 90 percent of activity is attributed [to] as little as 10 percent of onboarded affiliates" (Geno Prussakov, AM Navigator) - repeated by its author without external citation; treat the 90/10 shape as folklore that matches everyone's experience, not as data.

**Vendor-published, self-reported platform data - directional only**:

- Rewardful "State of SaaS Affiliate Programs" (2026, 2,847 programs, no third-party audit): 7.6% of affiliates generate at least one referral; 1.28% ever produce a paid sale; of those who refer traffic, 16.8% convert at least one sale; 12-18% confirm a signup within 7 days.
- TrackRev synthesis of network reports (no linked primaries): ~18% median 30-day activation; top quartile ~31%, bottom quartile ~6%; top 10% of affiliates capture ~71% of revenue.
- Track360: affiliates who send no traffic within 30 days rarely become productive; below-30% 30-day activation "signals onboarding problems."

**Archetype context for B2C expectations** (network/vendor-published, directional): affiliate click-to-sale conversion runs roughly 0.5-1% (Awin) to 1-3% (Partnero); by archetype, one managed-program dataset (Partnerize) puts content affiliates near 1.2% and loyalty/cashback near 4.8% - reason to set per-archetype expectations rather than one program-wide number. A 60-day cookie window is reported as the most common configuration (~41% of programs, TrackRev - same vendor-synthesis caveat as above).

**Time-to-first-conversion, B2B**:

- PartnerStack: median 2 days to first partnership _joining_ - not revenue; average ~45-50 days to first sale is a CSM's stated estimate, not a published statistic.
- Reditus: first paid B2B referral commonly lands 3-6 months after launch.
- No claimed "onboarding SLA makes partners X% more likely to transact" statistic exists in published PartnerStack material - do not cite one.

## The diagnostic gates behind the pass threshold

- **B2C: ≥30% activation at 30 days.** Below it, the diagnosis is onboarding, not recruitment (Track360's published framing).
- **B2B: ≥20% activation at 90 days.** Vendor blogs place healthy B2B partner activation at 30-40% within 90 days and call below-20% structural friction (U4IA) - vendor assertions, not analyst-verified.
- Both gates derive from vendor/practitioner data. Use them to trigger iteration; qualify them whenever the user asks "is that the industry standard?" - the honest answer is that these are vendor figures, not audited benchmarks.
- Measurement honesty: where server-side tracking is not deployed, cookie loss (browser tracking prevention, ad blockers) makes reported activation understate true activity - fix tracking before concluding the sequence failed.

## Sequence-shape evidence

- 4-6 emails over 10-14 days, one action each: independently stated by McWilliams ("four to six emails over 10 to 14 days is the right range"), Pyzyk ("five emails is the minimum viable sequence"), and Hunter.io practice (six emails over two weeks).
- Practitioners sequence by trigger and milestone rather than by a fixed day-0/2/7/14/30 calendar. Hence the skill's rule: skeleton as starting point, behavior layer on top.
- Time-boxed goal bonuses outperform flat rate increases because they give "a finish line" (Tapfiliate's published framing) - directional vendor claim, consistent with the deadline-bound bonus in Prussakov's approval-email format.

## Frameworks - honest verdict

- **"3 Es" - Excite, Encourage, Equip** (Prussakov, AM Navigator, 2018). Real, traceable, purpose-built for re-activating stalled affiliates. A single author's practitioner coinage, not an industry-standard methodology - usable for the re-engagement branch if labelled that way.
- **RITS - Resources, Incentives, Tools, Support** (same author, 2015). Same status; overlaps heavily with the 3 Es. Use one or the other, labelled as a practitioner mnemonic.
- **"60/40 Principle"** - appears in one vendor guide with no named originator or citation. Unverified; do not use.
- **AIDA / RACE / AARRR** - general marketing funnels, none of them built for affiliate activation; do not force-fit them.

## Sourced practitioner findings worth quoting (with attribution)

- "Recruiting an affiliate into your program is only half the job done. You need to motivate them to get active." - Geno Prussakov, AM Navigator.
- "Threats are generally counterproductive"; motivate "by presenting them with opportunities to succeed." - motivation researcher Alexander Hiam (1999), cited by Prussakov against activation-by-termination-threat. This is the basis for the never-threats rule; do not embellish it further.
- "Don't just brain dump onto partners. Make it excessively easy to get information and assets they need and lay out the path to success." - Emma Sinai, PartnerStack. The source of the welcome-dump failure mode.
- "People think that a partnership is active once you sign the deal. But if a partner doesn't follow through, they aren't truly active." - Joe Corcione, Trainual.
- "Don't sign an agreement and let it sit. The biggest thing for us has been setting up a concrete activation plan and expectations upfront." - Paige Berger, Secureframe.

# Touch Copy and Sequence Logic

How to draft each touch's copy, and the flow-control rules that keep the sequence honest once affiliates start behaving differently from the plan.

## Copy rules per touch

- Ask one action per touch, with one CTA. A second ask halves the first's chance; move it to the next touch.
- Put the tracking link or coupon code in the first two lines of touch 1 - the single highest-leverage placement decision in the sequence.
- Keep touch 1 under ~300 words. Later touches: 50-125 words for action-only touches (T4-style nudges), 150-300 for enablement touches (T2/T3-style).
- Write CTAs as action + outcome ("Copy your link and place it in your top review post"), never "learn more".
- Name the affiliate's archetype context, not the program's ("your listeners", "your comparison page"), so the ask lands in their world.
- State money mechanics plainly wherever they appear: commission, cookie window, payout threshold, hold period, clawback state. Vague money copy reads as evasive to professional affiliates.
- Frame the activation nudge as an opportunity with a finish line ("earn X by <date>"), never as a condition for staying in the program - the threat framing is the documented counterproductive version of the same message.
- Match sender to job: automated program address for provisioning and mechanics; the named AM (when one exists) for the connection touch and anything asking for a reply. A reply-seeking email from a no-reply address answers itself.

## Subject lines

- Lead with the concrete thing inside: "Your GlowKit link + code", "Your sandbox is live", "One email to send this week".
- Keep the affiliate's action or asset in the subject, not the program's excitement ("Welcome to the family!" says nothing to act on).
- Milestone-trigger subjects name the milestone: "Your first sale just tracked".

## The humanizer pass

Run every drafted email through your preferred humanizer skill before presenting it as final; reject raw first-draft model output. Affiliates are professional recipients of partner email - template-scented copy trains them to skim, and the sequence's entire mechanism is getting one action read and taken. The pass must preserve the one-action structure and the money mechanics verbatim; it rewrites tone, rhythm, and phrasing only.

## Exit, suppression, and re-entry rules

Define these before launch; a sequence without exits keeps selling activation to people who already activated.

- **Exit on activation**: the moment the activation event tracks, leave the calendar track. Fire the celebration touch, then hand over to ongoing program communications - continuing the onboarding drip past activation reads as nobody-watching automation.
- **Exit on the never-activated branch**: entry into the branch ends the main sequence; the branch's own final-notice/archive rules take over.
- **Suppression**: pause calendar touches while a reply thread with the AM is open, and honor unsubscribes program-wide immediately. Behavior-triggered transactional notices (commission approved, payout sent) survive marketing unsubscribes only where the program's terms and local law say so - flag the question, don't assume.
- **Trigger-overrides-calendar**: when a behavior trigger fires within a day of a scheduled calendar touch, send the trigger touch and skip or delay the calendar one - two emails in a day from the same program is the fastest route to the skim reflex.
- **Re-entry**: an affiliate reactivated from the archive re-enters at T2 (tools), not T1 - they don't need re-approval, they need a working asset and one action.

## Diagnosing copy vs structure

This separation runs after the ordered checks in the skill's diagnosis section, never instead of them - it names which plane a confirmed fault sits in, and carries no ordering of its own to compete with theirs. Ranking these four branches would be false precision: they are mutually exclusive readings of one engagement pattern, so you land on exactly one and never choose between them.

When a running sequence underperforms, separate the two failure planes before rewriting anything:

- Opens fine, clicks poor → copy plane: ask unclear, link buried, wrong asset for the archetype.
- Opens poor from touch 2 onward → structure plane: touch 1 overloaded or mis-targeted; the welcome dump trains ignoring.
- Clicks fine, activation poor → neither: check tracking first (the click may be converting invisibly), then the asset kit's fit, then the nudge's deadline.
- Everything fine until the nudge, then silence → the incentive's economics, not its copy - route to the commission skill.

# Worked Sequence Plans

Two complete artifacts in the deliverable format - one B2C ecommerce, one B2B SaaS. Adapt the reasoning, not the literal day numbers.

## Table of Contents

- [Example 1 - B2C ecommerce (skincare DTC brand)](#example-1---b2c-ecommerce-skincare-dtc-brand)
- [Example 2 - B2B SaaS (compliance-automation platform)](#example-2---b2b-saas-compliance-automation-platform)
- [Negative example - the welcome dump](#negative-example---the-welcome-dump)

## Example 1 - B2C ecommerce (skincare DTC brand)

Context from the interview:

- B2C, activated = first tracked sale within 30 days.
- 12% one-time commission, 60-day cookie, $50 payout threshold, net-30 hold.
- Intake is coupon-deal sites, social creators, and email-list owners.
- Links + codes auto-provisioned, postback live.
- Tax form gates payout.
- Automated sender, no AM capacity.
- ~150 approvals/month.

```
SEQUENCE PLAN - GlowKit Affiliates, 2026-08
Activation       : first tracked sale within 30 days of approval
Routing          : coupon-deal -> self-serve | social creator -> self-serve
                   email-list owner -> self-serve; 150/mo, no AM => no high-touch tier
Touches          :
  T1 Trigger : approval (within minutes)
     Channel : email  Sender : automated program address
     Action  : copy your link/code (both in the first two lines)
     Content : approval confirmed; commission, cookie, payout terms; disclosure
               attestation; what arrives next; under 300 words
     Asset   : pre-coded link + unique coupon code
  T2 Trigger : day 2
     Channel : email  Sender : automated
     Action  : grab the one asset for your archetype
     Content : coupon sites -> code page template; creators -> caption pack +
               disclosure line; list owners -> ready-cut email
     Asset   : asset hub, deep-linked per archetype
  T3 Trigger : first click (behavior; overrides calendar)
     Channel : email  Sender : automated
     Action  : place the code where that traffic came from
     Content : celebration + top-converting page for their traffic source
     Asset   : best-converting-pages one-sheet
  T4 Trigger : no click by day 10 (behavior)
     Channel : email  Sender : automated
     Action  : send one ready-cut email / publish one post (30-minute job)
     Content : lowest-effort promo, pre-written, link pre-coded
     Asset   : swipe copy
  T5 Trigger : day 21, only if not activated
     Channel : email  Sender : automated
     Action  : make one sale by day 30
     Content : activation bonus - extra $25 on the first sale before the window
               closes (deadline-bound; rate design per the commission skill)
     Asset   : reminder of code + link
  T6 Trigger : first sale (behavior)
     Channel : email  Sender : automated
     Action  : none - celebrate; state when commission approves (after the
               30-day return window) and pays (net-30)
Activation nudge : $25 first-sale bonus, expires day 30, carried by T5
Never-activated  : entry = no click and no sale by day 30; re-engagement at day
                   45 (one incentive + one asset); final notice at month 6 -
                   "account archived unless you reply"; then suppress, not delete
KPIs             : 30-day activation rate | median time-to-first-sale |
                   never-activated share at day 30 | open/click per touch
```

Why it holds:

- Every touch asks one action.
- The link leads touch 1.
- T3/T4 are behavior-driven, so stuck and gone get different messages.
- Payout terms and gating are stated on day 0, not at first commission.

## Example 2 - B2B SaaS (compliance-automation platform)

Context from the interview:

- B2B, activated = first qualified referral (demo booked and accepted) within 90 days.
- 25% recurring commission for 12 months, 90-day cookie, deal registration in place.
- Intake is consultants, agencies, and two tech partners.
- Named partner manager with call capacity.
- ~15 approvals/month.

```
SEQUENCE PLAN - CompliCo Partners, 2026-08
Activation       : first qualified referral (accepted demo) within 90 days
Routing          : consultant -> hybrid | agency -> high-touch
                   tech partner -> high-touch; 15/mo with named PM => calls viable
Touches          :
  T1 Trigger : approval (within minutes)
     Channel : email  Sender : named partner manager
     Action  : book your kickoff call (high-touch) / log in and register a
               deal (hybrid)
     Content : approval, terms + disclosure attestation, commission and
               clawback terms, tax/payout setup link (gates payout)
     Asset   : portal credentials + deal-registration link
  T2 Trigger : day 1
     Channel : in-portal + email  Sender : automated
     Action  : activate your sandbox account
     Content : sandbox/NFR access, integration docs for tech partners
     Asset   : sandbox credentials
  T3 Trigger : kickoff call booked (behavior) - else day 7 reminder
     Channel : call  Sender : partner manager
     Action  : agree the activation plan - first-deal date, next-step owners
     Content : terms recap, targets, product walkthrough, promo-calendar slot
     Asset   : sales deck + comparison/buyer-question material
  T4 Trigger : certification completed (behavior) - certification exists as a
               touch; its curriculum belongs to the enablement skill
     Channel : email  Sender : automated
     Action  : register your first deal
     Content : celebration + deal-registration walkthrough
  T5 Trigger : no deal registered by day 45 (behavior)
     Channel : email  Sender : partner manager
     Action  : reply with the one account you'd pitch first
     Content : personal check-in; offer co-pitch help; one concrete next step
  T6 Trigger : day 75, only if not activated
     Channel : email  Sender : partner manager
     Action  : register a qualified referral by day 90
     Content : time-boxed ramp incentive (economics per the commission skill)
  T7 Trigger : first qualified referral (behavior)
     Channel : email  Sender : partner manager
     Action  : none - celebrate; explain pending -> approved commission flow
               and the 12-month recurring window
Activation nudge : ramp incentive, expires day 90, carried by T6
Never-activated  : entry = no registered deal by day 90; one re-engagement at
                   day 120 (new use-case angle + offer to co-sell); final
                   notice at month 6; archive, keep door open
KPIs             : 90-day activation rate | median time-to-first-referral |
                   never-activated share at day 90 | call-booked rate | per-touch engagement
```

Why it holds:

- The window (90 days) matches B2B deal physics instead of importing the B2C 30-day cutoff.
- Sandbox and deal registration are provisioned day 0.
- Calendar touches only fire when behavior hasn't.
- The clawback-aware payout explanation arrives with the first success, when the partner actually cares.

## Negative example - the welcome dump

One email at approval containing: login, link, code, 14 banners, the full terms PDF, tax form, brand guidelines, a certification invite, and "let us know if you have questions" - then silence. No single action, the link below the fold, no follow-up branch, no deadline. This is the most common real-world shape and the one the plan above exists to replace.

# Touch Plan Patterns

The skeleton, the trigger layer, the routing, and the never-activated branch. Vendor and practitioner names appear here only as sourced evidence - never treat any of them as a required tool.

## The day-offset skeleton

Documented practitioner consensus (Matt McWilliams, Nicole Pyzyk, Hunter.io's Antonio Gabric, all independently) lands on 4-6 emails over 10-14 days, one action each. Affiliate-platform vendors converge on the same shape from a different angle: GoAffPro, PromoteKit, and ReferralRocket each publish a near-identical Day 1/3/7/10/14 cadence (welcome and quick start, product deep dive, promotion playbook with a "first link by day 7" micro-goal, social proof, personal check-in), independently landing on the same touch count and window. No single canonical industry calendar exists, but the convergence across six independent practitioners and platforms is real. Present this as a defensible starting point rather than a fixed industry calendar - each of those practitioners layers behavior triggers on top of the offsets.

| Touch               | Offset                 | One action asked                                                    | Content                                                                                  |
| ------------------- | ---------------------- | ------------------------------------------------------------------- | ---------------------------------------------------------------------------------------- |
| T1 Welcome          | Minutes after approval | Place your link (link/code in the first lines)                      | Approval confirmed, terms + disclosure attestation, what's coming next; under ~300 words |
| T2 Tools            | Day 2-3                | Grab one asset and use it                                           | Creatives, best-converting pages, coupon code, asset-hub location                        |
| T3 Connection       | Day 5-7                | Reply: how do you plan to promote?                                  | Personal note from the AM (or founder), offer of help and custom creative                |
| T4 Nudge            | Day ~14                | One low-effort promo ("here's an email you can send in 30 minutes") | Ready-cut copy attached; friction near zero                                              |
| T5 Activation offer | Day ~30                | Hit the time-boxed goal                                             | Deadline-bound bonus; last calendar touch before the never-activated branch              |

Prussakov's approval-email format (AM Navigator, 2009) compresses the same jobs into touch 1:

- Why the program is worth promoting.
- A fast-action incentive.
- A deadline-bound bonus.
- Pre-coded links pasted directly into the email ("a simple cut-n-paste job").
- A personal support offer.

## The behavior-trigger layer

Every sourced practitioner favors milestone triggers over the fixed calendar; one platform's published automation pattern (PartnerStack) fires messages on milestones - training completed, first signup, first sale - after the first two calendar sends. Layer these on the skeleton; when a trigger fires, it overrides the next calendar touch:

- **First click** → celebration + "here's what converts best next".
- **First sale/referral** → distinct celebration touch; reinforce the milestone, state when the commission approves and pays.
- **No click by day N** (N ≈ half the activation window) → the T4-style low-effort nudge, early.
- **Training/certification completed** (B2B) → unlock the next enablement step or the kickoff call.
- **Tax/payout setup incomplete by day N** → reminder that payout is gated until done.

## Tier routing

Ranked by activation bought per manager hour, best ratio first - not by activation rate, which runs the other way:

| Rung          | What it is                                                         | Manager hours per affiliate                     | Activation bought                             | Reach for it when                                                           |
| ------------- | ------------------------------------------------------------------ | ----------------------------------------------- | --------------------------------------------- | --------------------------------------------------------------------------- |
| 1. Self-serve | Portal plus automated drip                                         | near-zero once built; the build is about a week | Lowest per affiliate, but applied to everyone | Always - it is the floor the other rungs sit on, not an alternative to them |
| 2. Hybrid     | The same drip plus one personal check-in in week 1-2 (the T3 slot) | minutes per affiliate, recurring                | Middle                                        | A named AM exists and some archetypes' expected volume repays the check-in  |
| 3. High-touch | AM-run personal welcome and kickoff call                           | an hour or more per affiliate, forever          | Highest per affiliate                         | A partner's expected revenue on its own exceeds the hours the calls consume |

- efficiency: `self-serve > hybrid > high-touch`
- value: `high-touch > hybrid > self-serve`
- recurring effort: `high-touch > hybrid > self-serve`
- one-off build: `self-serve == hybrid > high-touch`

Hybrid ties self-serve on build because hybrid _is_ the self-serve drip with a calendar slot attached - the same week of work stands both up. High-touch inverts the shape entirely: almost nothing to build, and a standing job thereafter.

High-touch is also what this order starves: highest activation bought per affiliate, highest recurring hours per affiliate, and therefore last on every ratio. Promote it deliberately on either condition - one partner whose expected revenue on its own exceeds the hours the calls consume, or a roster small enough that the drip's build never pays back - rather than waiting for a round it cannot win.

Size against interview questions 8-9 and 12, and delete rather than demote what the answers rule out.

- No named AM, or more than roughly a hundred approvals a month, removes high-touch from the menu outright: a tier with nobody to staff it is not a slower option, it is an absent one, and leaving it listed invites a plan that quietly never runs.
- Running the two survivors in parallel - self-serve long tail, high-touch strategic - is the pattern PartnerStack's team describes publicly.

Re-rank against the program's own advantages before accepting the default order. A platform with milestone automation already live has paid self-serve's build cost already, widening its lead. A roster of a dozen partners each worth a large deal flips the order: at that volume the per-affiliate hour is cheap and the drip's build never pays back, so high-touch wins on efficiency too.

## Archetype routing

Route the first action and the leading asset by archetype; the skeleton stays the same. This table is deliberately unranked, and ranking it would be false precision: the rows are not competing options to pick between, they are a lookup keyed by the archetypes the intake actually contains. Every approved archetype gets served, so "which first" has no meaning here - only "which of these do we have".

| Archetype             | First action to ask                              | Leading asset                                               |
| --------------------- | ------------------------------------------------ | ----------------------------------------------------------- |
| Content/review site   | Place link in an existing high-traffic post      | Best-converting pages, product one-sheet, comparison angles |
| Coupon-deal           | Publish the exclusive code                       | Unique per-affiliate code (leak-resistant if available)     |
| Loyalty-cashback      | Confirm catalog/feed integration                 | Product feed access, rate table                             |
| Social creator        | One post/story with link or code                 | Short talking points, visuals, disclosure reminder          |
| Email-list owner      | Send one ready-cut email                         | Swipe copy with pre-coded link                              |
| Sub-network           | Distribute offer to their publishers             | Feed/API access, program one-pager                          |
| B2B agency/consultant | Register the first deal or book the kickoff call | Deal-registration link, sales deck, sandbox                 |
| B2B tech partner      | Stand up the integration/sandbox                 | Sandbox or NFR access, integration docs                     |

## B2B kickoff call agenda (high-touch tier)

- Partnership terms and commission recap.
- Targets and the concrete activation plan with next-step owners.
- Product walkthrough with sandbox access.
- First-content or first-deal date.
- Promo-calendar slot.
- Custom landing-page needs.

Setting the activation plan and expectations upfront - instead of signing and letting it sit - is the practice partner managers at PartnerStack-published companies (Secureframe, Trainual) credit for activation.

## Single-hub alternative

One documented pattern (monday.com, via PartnerStack) replaced a 7-step email drip + 5-step PDF with a single onboarding landing page, a recorded dashboard walkthrough, weekly kickoff calls, certification courses, and a sandbox - automating account setup specifically to reduce dropout.

Against the drip it replaces: `activation bought per week of build: trim the drip > swap T2-T4 for a hub`. Trimming costs an hour and is reversible in an hour; the hub costs a quarter to build and a standing job to keep current, and reverting it means rebuilding the drip you deleted.

Reach for the hub only when the per-touch engagement data shows opens decaying across the whole sequence - touch volume itself being the friction - never because one touch underperforms, which is a copy problem the hub does not solve. When you do swap, keep T1 and the trigger layer and collapse only T2-T4.

## The never-activated branch

- **Entry**: no qualifying action by the activation window's end (30 days B2C, 90 days B2B as defaults).
- **Diagnose before re-engaging**: distinguish never-activated (no click ever) from dormant (was active, stopped) - they get different messages; only never-activated belongs to this branch.
- **Re-engagement touches (1-2)**: re-excite (why the program is worth promoting), one concrete incentive, one ready-to-use asset - Prussakov's "3 Es" (Excite, Encourage, Equip), usable if labelled as one practitioner's coinage, not an industry method.
- **Final notice**: one email stating the account will be archived unless they respond - a documented list-hygiene practice (McWilliams), typically after ~6 months of zero clicks and zero replies.
- **Archive, don't delete**: suppression keeps the active-rate metric honest and leaves the door open.
- **Never threats**: motivation research cited in the affiliate literature (Alexander Hiam, 1999, via Prussakov) holds that "threats are generally counterproductive" - frame the final notice as housekeeping, not punishment.

---
name: affiliate-onboarding-sequence
description: Design the onboarding sequence that carries a newly approved affiliate from approval to first tracked conversion - touch plan (timing, channel, one action per touch), asset kit, time-boxed activation nudge, and never-activated follow-up branch. Covers B2B SaaS partner programs and B2C ecommerce affiliate programs. Use whenever the user mentions affiliate onboarding, an affiliate welcome email series, affiliate activation, a new-affiliate ramp, partner activation emails, or approved affiliates who never send a click or a sale, even if they never say onboarding. Starts at approval - to pitch affiliates who have not joined yet use mbfinotti/partnerships-skills@affiliate-recruitment-outreach.
license: MIT
metadata:
  author: Maya-Beth Finotti
  version: "1.0.4"
---

# Affiliate Onboarding Sequence

Design the sequence that moves a newly approved affiliate to their first tracked conversion. Most programs activate only a small minority of approved affiliates - practitioner estimates cluster around 5-20%, a convergent range rather than a controlled measurement - so the sequence's whole job is to compress time-to-first-conversion while intent is still warm.

Two commitments run through everything below:

- **Day offsets are a starting skeleton, not a standard.** Documented practitioner consensus is 4-6 touches over 10-14 days, but every sourced practitioner favors behavior triggers - first click, first sale, no-click-by-day-N - layered on top of, and overriding, the calendar.
- **An affiliate who cannot get a working link cannot activate.** Provisioning and tracking readiness is the pre-condition of the sequence, never a touch inside it.

Two entry paths, both starting with the Interview:

- Build from scratch - "write the welcome emails for affiliates we just approved", "design our affiliate onboarding sequence", "what should new partners get in their first 30 days?" - run the full Workflow.
- Diagnose a running sequence - "our approved affiliates never send a click", "nobody activates after the welcome email" - run the Interview, then jump to Diagnosing an existing sequence before touching the plan.

## Interview

Ask one question per message; offer the multiple-choice options where given. Skip any question the user already answered.

1. B2B SaaS or B2C ecommerce? Both means the plan needs two branches.
2. What event counts as "activated": first click, first lead/trial, first paid sale, or first closed deal? Inside what window (days from approval)?
3. Program mechanics: commission model (recurring %, one-time %, flat CPA), cookie window, payout threshold, payout hold period?
4. Which archetypes does the intake actually contain: content/review site, coupon-deal, loyalty-cashback, social creator, email-list owner, sub-network - and in B2B: agency, consultant, tech partner?
5. Which assets exist today vs must be created: tracking links, banners, swipe copy, product one-sheets - and for B2B: sandbox/demo access, comparison and buyer-question material?
6. Tracking setup: are links or coupon codes auto-provisioned at approval? Is server-side (postback) conversion tracking live?
7. Tax and payout collection: which steps (tax form, payout method, terms attestation) - and do they gate payout eligibility?
8. Who sends the touches: an automated program address, or a named affiliate manager? Does capacity for calls and 1:1 exist?
9. How many new approvals arrive per month?
10. By what date must the first activation numbers land? A date inside the next cohort's window deletes every option that needs a build and promotes the near-zero fixes; no date leaves the default ordering intact.
11. A one-off win on this cohort, or a compounding asset? Compounding promotes the automated drip and the persistent asset hub; one-off promotes manual outreach to the affiliates already approved.
12. Effort ceiling: manager hours per week, whether a named AM exists at all, and whether changing a commission needs finance or legal sign-off. Zero AM hours deletes the high-touch and hybrid tiers outright rather than demoting them.

Questions 10-12 set the ordering of every menu in this skill; the answers are what the rankings below get re-sorted against.

## Workflow

1. Run the Interview.
2. Fix the activation definition and its measurement window before designing anything. Touch count, nudge deadline, and every KPI depend on it; a plan without it cannot be judged.
3. Verify day-0 readiness: link/code provisioning at approval, tracking that records the activation event, and payout/tax collection wired as a payout gate - checklist in [references/asset-kit-and-provisioning.md](references/asset-kit-and-provisioning.md). Stop and fix gaps before mapping a single touch.
4. Segment the intake twice: by archetype (question 4) and by touch tier. Archetype is a lookup, not a ranking: route the archetypes the intake actually contains. Tier is a ranking, and it decides where the manager's hours go:

   - efficiency (activation bought per manager hour): `self-serve > hybrid > high-touch`
   - value (activation rate per affiliate): `high-touch > hybrid > self-serve`
   - recurring effort (manager hours per affiliate): `high-touch > hybrid > self-serve`
   - one-off build before the first send: `self-serve == hybrid > high-touch`

   Self-serve leads because its marginal cost per affiliate is near-zero once built, not because it converts best - it converts worst. Build it as the floor for everyone, then buy activation upward for the segment that repays it. The build tie is real: hybrid _is_ the self-serve drip plus one calendar slot, so it costs the same week to stand up; high-touch needs almost no build and instead becomes a standing job.

   High-touch is what this order starves. It tops the value line and tops the recurring-effort line at once, so the ratio buries it every round and a program that only computes efficiency ends up serving its largest partners the same automated drip as its long tail. Promote it explicitly - never wait for it to win - when a single partner's expected revenue exceeds the hours the calls consume, or when the roster is small enough that the drip's build never pays back.

   Default: self-serve for the whole intake, hybrid layered on the archetypes whose expected volume repays a check-in.

   Delete high-touch from the menu, do not demote it, when no named AM exists or approvals run past roughly a hundred a month. A tier nobody can staff is absent, not slow, and one parked at the bottom quietly reappears as scope the plan cannot execute.

   This ordering is a default, not a law, and it shifts with who executes it. Re-rank it against what the Interview already told you:

   - An affiliate platform with milestone automation already live collapses the self-serve build to near-zero and widens its lead.
   - A manager with no capacity (question 12) deletes hybrid and high-touch both.
   - A roster of a dozen partners each worth a large deal inverts the order outright, because near-zero volume makes the per-affiliate hour cheap and the self-serve build never pays back.

   Tier detail in [references/touch-plan-patterns.md](references/touch-plan-patterns.md).

5. Map the touch plan per segment: day-offset skeleton plus behavior-trigger layer, patterns in [references/touch-plan-patterns.md](references/touch-plan-patterns.md). Every touch asks exactly one action.
6. Place terms/disclosure attestation in the approval touch, and tax/payout collection before the first payout is possible - never after the first commission is earned. Decide placement only; drafting the clauses belongs to the terms skill.
7. Attach a time-boxed activation nudge inside the first window, when intent peaks. Decide its slot, deadline, and framing here; its rate economics belong to the commission skill.
8. Design the never-activated branch: diagnostic checkpoint at the window's end, re-engagement touches, one final-notice email, then archive. Never use termination threats as the lever - motivation research documents threats as counterproductive.
9. Draft copy for every touch - one action, one CTA, link or code in the first lines of touch 1 - and set the exit, suppression, and re-entry rules, using [references/copy-and-sequence-logic.md](references/copy-and-sequence-logic.md).
10. Run every drafted email through your preferred humanizer skill; reject raw first-draft model output as final copy. The pass rewrites tone only - the one-action structure and money mechanics stay verbatim.
11. Validate the plan with the user section by section - activation definition, routing, touch plan, nudge, never-activated branch, copy - revising on pushback before moving to the next section.
12. Emit the Sequence Plan artifact below and set the KPI review cadence. If your harness has persistent memory, memorize the activation definition and routing decisions so a later diagnosis run starts from them.

## The Sequence Plan

Deliver every engagement as this artifact - it is the deliverable contract, not an example:

```
SEQUENCE PLAN - <program>, <date>
Activation       : event + window (e.g. first paid sale within 30 days of approval)
Routing          : archetype -> tier (high-touch | self-serve | hybrid), volume note
Touches          : one block per touch
  T<n> Trigger : day offset OR behavior event (behavior overrides calendar)
       Channel : email | call | in-portal | chat
       Sender  : automated program address | named AM
       Action  : the ONE action asked
       Content : one-line summary
       Asset   : asset attached or linked
Activation nudge : incentive + deadline + which touch carries it
Never-activated  : entry condition, touches, final-notice rule, archive rule
Exits            : on activation | on branch entry | suppression + re-entry rules
KPIs             : activation rate @ window | time-to-first-conversion |
                   never-activated share at cutoff | per-touch engagement
```

Every touch line carries a reason the reader can review - "T3 fires on no-click-by-day-7 because the calendar alone can't tell stuck from gone" is reviewable; a bare day number is not. Two fully worked plans (one B2C, one B2B) live in [references/sequence-plan-examples.md](references/sequence-plan-examples.md).

## B2B SaaS vs B2C ecommerce

Where the two diverge, the plan must branch:

| Dimension          | B2C ecommerce                                   | B2B SaaS                                                            |
| ------------------ | ----------------------------------------------- | ------------------------------------------------------------------- |
| First conversion   | Tracked sale                                    | Lead, trial, demo, or closed deal - fix which                       |
| Time horizon       | Days to weeks; 30-day window fits               | 30-180 days; 90-day window fits                                     |
| Touch model        | Automated self-serve for nearly all             | Hybrid/high-touch; kickoff call for strategic partners              |
| Asset kit          | Links, codes, banners, swipe copy, product feed | Adds sandbox/demo, comparison pages, buyer-question material        |
| Day-0 provisioning | Link + coupon code                              | Adds deal registration and sandbox/NFR access                       |
| Payout risk        | Clawback ends with the return window            | Clawback rides every recurring invoice; longer holds                |
| Sequence span      | 10-14 day skeleton, then triggers               | Same skeleton, but trigger layer stretches across the 90-day window |

Identical for both - stated so it is not left implicit:

- The activation-definition-first rule.
- The provisioning pre-condition.
- One action per touch.
- Tax/payout gating at onboarding.
- The never-activated branch structure.
- The humanizer pass.
- The Sequence Plan format itself.

## Failure modes

Fix in this order when several of these are true at once - activation bought per hour of fixing: `link placement > welcome dump > approval-as-finish-line > threat framing > activation nudge > archetype routing == never-activated branch`. The tie is argued: routing and the branch each cost about a week to build, but routing pays inside the window you are already in while the branch only pays on the next cohort - equal ratio, different clock.

| Failure                                                                | Fix                                                                                                  | Fix effort                                       |
| ---------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------- | ------------------------------------------------ |
| Tracking link buried, or not yet provisioned at send time              | Link/code in the first lines of touch 1; provisioning verified before the sequence launches          | near-zero                                        |
| The welcome dump - everything crammed into email one, no single action | One action per touch; move the rest to later touches or a persistent asset hub                       | an hour                                          |
| Approval treated as the finish line                                    | Set the concrete activation plan and expectations in touch 1                                         | an hour                                          |
| Termination threats used to force activation                           | Replace with opportunity-framed incentives; motivation research finds threats counterproductive      | near-zero, when threats are present at all       |
| No activation nudge, no time-boxed first-window incentive              | Deadline-bound bonus inside the first window, carried by a named touch                               | a week - the rate needs sign-off before it ships |
| One-size-fits-all sequence across archetypes and tiers                 | Route by archetype and tier before drafting anything                                                 | a week - one leading asset per archetype present |
| No branch for the never-activated                                      | Diagnostic at window end, re-engagement, final notice, archive - designed upfront                    | a week to design, then a standing job            |
| Payout/tax collection deferred until after the first commission        | Gate payout eligibility at onboarding; unclaimed-setup commissions have documented expiry precedents | a week plus finance coordination                 |

The payout/tax row sits outside that ordering deliberately. It buys avoided payout disputes and reduced legal exposure, not activation, so it cannot be ranked on the same ratio as the rows above it - fix it on the compliance clock, before the first commission is earned, whatever the activation work is doing.

## Diagnosing an existing sequence

When the user arrives with a running program ("our affiliates never send a click"), run checks in this order instead of redesigning from scratch. The order is information bought per unit of effort - not cheapest-first, which would put the conversion trace last and let you redesign a sequence that was working all along:

- information bought per hour spent: `conversion trace > approval-email check > per-touch engagement > cohort split > nudge and branch audit > rewriting touch 1`
- effort: `rewriting touch 1 > conversion trace == cohort split > approval-email check == per-touch engagement == nudge and branch audit`

Both ties are real, not hedges: the trace and the cohort split each cost about an hour of query and click work, and the last three each cost near-zero because they only read what already exists.

| #   | Check                                                                                                            | Effort        | What it settles                                                                                                            | Symptom pointing here                                                                       |
| --- | ---------------------------------------------------------------------------------------------------------------- | ------------- | -------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------- |
| 1   | Trace one conversion end to end: click → recorded conversion → pending commission → reversal state               | an hour, once | Whether the reported activation number is real. A dark hop makes every row below measure fiction                           | Activity visible in-platform but activation reported low; clicks tracked, conversions never |
| 2   | Open the approval email as a newly approved affiliate: is a working link or code in it, provisioned at approval? | near-zero     | Whether the fault is the pre-condition rather than the sequence at all                                                     | Approved affiliates never click                                                             |
| 3   | Read the per-touch open and click rates the platform already records                                             | near-zero     | Where the sequence loses people - localizes the fault, never names it                                                      | Opens collapse after touch 1                                                                |
| 4   | Split activation by archetype and by tier                                                                        | an hour       | Whether one sequence is being sent to archetypes that need different first actions                                         | One archetype activates, the rest never                                                     |
| 5   | Audit the plan for a deadline-bound nudge inside the window and a designed never-activated branch                | near-zero     | Why the cohort stalls at the deadline - explains a late-window stall only, nothing earlier                                 | Cohort stalls before the window closes; long tail piles up untouched                        |
| 6   | Rewrite touch 1 or the asset kit                                                                                 | a week        | Nothing, until 1-5 have run - it is a change, not a check, and it destroys the baseline that would have told you it worked | (none - never the first move)                                                               |

Re-rank against the program in front of you:

- A program already running server-side tracking has bought row 1's information in advance: start at row 2.
- A program with a single archetype in its intake drops row 4 from the list entirely, rather than ranking it last.

Copy-level vs structure-level separation lives in [references/copy-and-sequence-logic.md](references/copy-and-sequence-logic.md).

## KPIs and pass threshold

- Activation rate at the defined window - the sequence's primary score.
- Time-to-first-referral / time-to-first-conversion, tracked as a cohort median.
- Never-activated share of each monthly cohort at the cutoff.
- Per-touch engagement (open/click) to locate where the sequence loses people.

Measure by monthly approval cohort: each month's approvals form one cohort, judged only when its window closes. A blended all-time rate hides whether the sequence is improving.

Pass gates come in two parts, because the sequence ships before any activation number exists to measure.

**Design-time**, check before emitting the artifact, and iterate until every line holds:

- The activation definition and its window are fixed (step 2).
- Day-0 provisioning is verified, not assumed (step 3).
- Every touch asks exactly one action.
- The tracked link or code appears in the first lines of touch 1.
- A time-boxed nudge sits inside the activation window.
- The never-activated branch ends in archive, with no termination threat.
- Every drafted email has been through the humanizer.

**Post-launch**, scheduled as the first review gate: do not report a verdict you cannot compute.

- 30-day activation at or above ~18% for B2C, the TrackRev _median_, not its ~31% top quartile; 90-day at or above ~20% for B2B.
- Track360's "below 30% signals onboarding problems" is a diagnostic trigger for investigating, not a ship gate; used as one, it demands top-quartile performance on day one, above the entire 5-20% range this skill opens by citing.
- These are vendor and practitioner numbers, not audited studies: thresholds for iteration, never facts to quote.
- Below the gate, diagnose the sequence (provisioning, first touch, nudge), not recruitment.
- If the program has no data yet, say so in the deliverable and set the measurement date; the user may also document why their economics justify a different gate.
- Reliability framing for every number lives in [references/benchmarks-and-frameworks.md](references/benchmarks-and-frameworks.md) - read it before quoting any benchmark to the user.

## Invocation Examples

- "We approve about 40 affiliates a month and almost none ever send traffic. Design the onboarding sequence."
- "Map the first 30 days for a newly approved B2B SaaS affiliate - links auto-provision at approval, tax forms gate payout, one affiliate manager for the whole program."
- "Design the never-activated branch: what we send at day 30, how many times, and when we archive."

## Reference

- See [references/touch-plan-patterns.md](references/touch-plan-patterns.md) for the touch skeleton, trigger layer, tier and archetype routing, kickoff-call agenda, and the never-activated branch.
- See [references/asset-kit-and-provisioning.md](references/asset-kit-and-provisioning.md) for the day-0 checklist, tracking-readiness checks, kit contents per side, and payout/tax gating.
- See [references/copy-and-sequence-logic.md](references/copy-and-sequence-logic.md) for per-touch copy rules, subject lines, the humanizer pass, and exit/suppression/re-entry logic.
- See [references/sequence-plan-examples.md](references/sequence-plan-examples.md) for one fully worked B2C plan, one B2B plan, and a negative example.
- See [references/benchmarks-and-frameworks.md](references/benchmarks-and-frameworks.md) for every benchmark with its reliability flag, and the honest verdict on named frameworks.
- See `mbfinotti/partnerships-skills@affiliate-recruitment-outreach` for everything before approval.
- See `mbfinotti/partnerships-skills@affiliate-program-terms` for drafting the terms surfaced in touch 1.
- See `mbfinotti/partnerships-skills@affiliate-commission-structure` for the economics of the activation bonus and the commission ladder.
- See `mbfinotti/partnerships-skills@partner-enablement` for the training/certification curriculum a B2B touch may point to.
- See `mbfinotti/partnerships-skills@affiliate-performance-dashboard` for building the reporting behind the KPIs named here.