mbfinotti/partnerships-skillsチェック済み
SKILL DETAIL
partner-channel-program
mbfinotti/partnerships-skills/partner-channel-program
Design the overall structure of a B2B partner or channel program from scratch - a readiness gate that can conclude not yet, partner value proposition, which motions to run (referral, reseller, services, technology), tier count, benefit stack, program-wide economics envelope, governance, and success horizon. Covers B2C analogs (dealer, franchise, ambassador). Use whenever the user mentions building a partner program, channel program design, launching a channel, a partner ecosystem program, or asks whether they are ready for partners, even if they never say program. Do NOT use for reworking tiers inside a program that already exists - use mbfinotti/partnerships-skills@partner-tiering instead.
インストール · 510出典を見る
Installation
npx skills add https://github.com/mbfinotti/partnerships-skills --skill partner-channel-program
スキルファイル
SKILL.md
最終同期 · 2026/10/23
evals/evals.json›
{
"skill_name": "partner-channel-program",
"evals": [
{
"id": 1,
"prompt": "I run growth at Tellurian Ops, we sell a workflow tool to facilities teams. ACV is $4,200. Direct sales have been flat at about $90k new ARR per quarter for four quarters straight and we've missed plan twice. Our CRO thinks a reseller channel is the answer - he wants four tiers (Bronze through Platinum), a 30% margin table copied off a competitor's partner page, a partner portal, and 100 signed resellers in year one. I've been asked to write the program plan. Where do I start on the tier criteria and the margin table?",
"expected_output": "A refusal to design the requested tiers and margin table, replaced by a readiness-gate verdict of 'not yet' that names the disqualifiers, states what must change, names referral-only as the lighter motion that still fits, and stops there.",
"files": [],
"expectations": [
"Concludes the honest answer is 'not yet' and declines to design the tier criteria or the margin table as asked",
"Names at least three separate disqualifiers rather than resting on one",
"Identifies the flat, non-repeatable direct motion as a disqualifier, stating that partners amplify a working motion rather than fix a broken one",
"Identifies expecting partners to solve a growth problem as a disqualifier in its own right",
"Identifies the $4,200 ACV as too low to fund a reseller motion, explaining that the partner's own cost of sale has to clear before the vendor's does",
"Flags the absence of a named executive sponsor as a disqualifier, or asks who the named executive sponsor is before proceeding any further",
"States that any single disqualifier is enough to return 'not yet' - the gate is not scored, averaged, or traded off",
"Names referral-only as the lighter motion that still fits at this ACV",
"Flags copying the competitor's 30% margin table as a structural mistake rather than a number to tune",
"States what must change before the gate can be re-run, covering direct-motion repeatability and the ACV or margin headroom",
"Does not produce a four-tier Bronze/Silver/Gold/Platinum ladder",
"Does not set a recruitment target of 100 partners or otherwise treat signed-partner count as the success metric"
]
},
{
"id": 2,
"prompt": "Bellweather Analytics here - B2B data quality product, $38k ACV, gross margin 79%, direct sales are repeatable (61 closed deals in 18 months with three reps running the same playbook). Our COO Priya Raman is the named sponsor. Four boutique data consultancies already implement us for their own clients with no agreement in place. We can't fund a new hire this year; I'd be running this alone alongside my current job, maybe six hours a week. The board wants a partner-sourced number on the books by the end of Q2, five months out. Which partner motions should we run? I'm assuming we do referral, reseller, tech integrations and an SI program and just phase them.",
"expected_output": "A motion selection that states the efficiency ordering explicitly, deletes motions the constraints rule out with reopen triggers attached, re-ranks against the consultancies and the team-of-one constraint, and lands on a minimum viable set rather than a phased rollout of all four.",
"files": [],
"expectations": [
"Refuses to run all four motions and picks a minimum viable set instead of phasing every motion",
"States the efficiency ordering of the motion menu out loud rather than leaving it implied by list order",
"Gives the default efficiency order as referral > technology > services-SI > reseller",
"Presents at least one further axis - effort, value ceiling, time-to-effect, or compliance cost - whose order disagrees with the efficiency order",
"Deletes reseller outright rather than scheduling it for a later phase, citing the absence of fundable standing headcount",
"Deletes services-SI on the same headcount constraint, or explicitly resolves the conflict between that deletion and the consultancy-driven promotion",
"Records a reopen trigger alongside each deleted motion rather than leaving it deleted with no stated condition",
"Uses the five-month board deadline to promote referral, saying that a hard date inside two quarters is the answer that moved it",
"Notes that the four consultancies already implementing the product promote services-SI, because the recruitment effort is already spent",
"States that a team of one collapses the menu to referral plus at most one integration",
"Names the condition that would promote reseller anyway: reaching a named segment or geography is the stated goal of the program and standing headcount is funded",
"Names the condition that would promote services-SI anyway: implementation appears as a named reason in lost or stalled deals, not merely that an SI relationship exists",
"Chooses motions before any Ideal Partner Profile or formal partner-scoring work rather than starting from a partner profile"
]
},
{
"id": 3,
"prompt": "Kestrel Freight Systems. We're standing up our first partner program - referral only for now, maybe eight partners by next summer. I've drafted the ladder: Bronze at 0 referrals, Silver at 3, Gold at 10, Platinum at 25, with a nicer logo badge and priority support at each step. Can you tighten up the criteria thresholds and tell me what benefits belong at each tier?",
"expected_output": "A refusal to set the count or design per-tier criteria and benefits here, limited instead to whether to tier and what the differentiation is for, with the count, criteria and benefit design handed to the tiering skill and the four-tier ladder challenged.",
"files": [],
"expectations": [
"Declines to set the tier count or design the per-tier criteria and benefits in this answer",
"Hands tier count, criteria, and benefit design to mbfinotti/partnerships-skills@partner-tiering",
"Explains that fixing a tier count here risks re-papering signed partner agreements once the tiering skill re-derives the count from partner-base size",
"Limits its own decision to whether to tier at all and what the differentiation is for",
"States that a tier exists only when something measurable changes at its boundary",
"States that criteria with no benefits delta is a leaderboard nobody plays",
"States that a benefits delta with no criteria is favoritism wearing a logo",
"Challenges the four-tier ladder for a program this size rather than accepting it, noting four tiers is usually vanity",
"Flags a logo badge and priority support as cosmetic differentiation with no economic change at the boundary",
"Does not present an efficiency ranking over tier counts"
]
},
{
"id": 4,
"prompt": "Orlando at Vantablack Logistics. We sell a $26k ACV routing platform, 68% gross margin. I pulled our closest competitor's partner page and they pay resellers 30%, so I've put 30% in our model too and rounded our referral fee to 15% of year one. Finance has signed off on the total spend. What else do I need in the economics section before we launch?",
"expected_output": "A rejection of the copied margin, replaced by the envelope method: derive margin from the task the partner takes over, model the partner's own side on the stated deal size, compare channel cost-of-sale to direct CAC, fit total cost inside gross margin, and set direct-rep comp neutrality now.",
"files": [],
"expectations": [
"Rejects the copied 30% competitor margin as a structural mistake rather than validating it against benchmark bands",
"States that partner margin must come out of a cost the partner takes off the vendor's hands",
"States that paying margin for work the vendor would have done anyway buys nothing from the partner",
"Asks or specifies which operational task the partner is actually taking over before any number is set",
"Models the partner's own side, computing what 30% leaves the partner on a $26k deal (roughly $7,800) against their acquisition and delivery costs",
"Compares the resulting channel cost-of-sale against direct CAC, naming a channel costlier than direct as a strategy tax rather than leverage",
"Requires total program cost to fit inside gross margin",
"Sets direct-rep compensation neutrality on partner-touched deals now, at architecture time, rather than leaving it as a later fix",
"Names misaligned direct-rep compensation as a root cause of channel conflict rather than a symptom of it",
"Requires cost-to-serve per partner to be tracked from day one, not just partner-generated revenue",
"Hands per-partner modeling to mbfinotti/partnerships-skills@partner-economics rather than doing it here"
]
},
{
"id": 5,
"prompt": "We launched our channel program in February. It's now September and our CRO wants it shut down - 14 partners signed, two registered deals, one closed for $31k. He's comparing it to our direct team, which turns a lead into closed business in about 70 days. I have one meeting on Thursday to save it. What do I tell him?",
"expected_output": "A diagnosis naming the month-seven judgment against a 70-day direct cycle as a horizon error, with the horizon reset placed first in the fix order and justified on cost, plus the year-1/2/3 ladder to replace the current expectation.",
"files": [],
"expectations": [
"Identifies the month-seven judgment against a 70-day direct cycle as a horizon error, a structural problem rather than a performance problem",
"Names the horizon reset as the first fix to make, ahead of any other remediation",
"Justifies the horizon reset leading on cost: one leadership conversation, and it stops the program being killed before any other fix can land",
"States the ranking of remediations out loud rather than implying it by list order",
"Gives year 1 as leading indicators only, naming at least two of partners activated, deals registered, time-to-first-deal, or enablement completion",
"Gives year 2 as meaningful partner-sourced pipeline",
"Gives year 3 as material revenue contribution",
"States that partner relationship maturity takes roughly two to three years, so little net revenue in year 1 is expected rather than a failure signal",
"Does not recommend recruiting more partners as the answer to the low deal count",
"States that the horizon decision belongs in the architecture at design time, not as a rescue argued after the fact"
]
},
{
"id": 6,
"prompt": "Inherited a two-year-old channel program at Northgate Instruments. 214 signed partners, 19 did anything at all last quarter. Our direct AEs quietly steer deals away from partners because their quota relief drops 40% when a partner is on the deal. Partners have stopped submitting registrations since we took three accounts back as house accounts in May. We allocated $400k of market development funds and $37k got claimed. We run a single tier, no metals. Exec team already knows this is a three-year build and isn't threatening to kill it. Give me the fix list in the order I should work it.",
"expected_output": "An explicit remediation ordering with the rows that do not apply to this user deleted rather than demoted, conflict and registration policy first among what remains, comp neutrality named highest-value but placed later on cost, and each symptom traced to its structural cause.",
"files": [],
"expectations": [
"States the fix ranking out loud as an explicit order rather than leaving it implied by list order",
"Deletes the horizon reset from this user's list because leadership already holds a three-year expectation, rather than demoting it to last",
"Deletes the tier rebuild because the program runs a single tier",
"Places conflict and registration policy first among the fixes that still apply",
"Places comp neutrality below activation-over-recruitment and funds mechanics in the working order, despite naming it the highest-value fix",
"Justifies comp neutrality's placement by its cost: a mid-year compensation change, slow and politically expensive",
"Traces the stalled registrations to the house-account carve-outs and prescribes honoring registration without carve-outs",
"Traces the 214-signed / 19-active gap to recruitment being tracked instead of activation, and stops recruiting rather than adding partners",
"Traces the unclaimed funds to market-development funds designed apart from deal registration, and puts both under one owner",
"Proposes automatic registration approval for funded leads as part of the funds fix",
"Names the 40% quota-relief cut as an architecture decision rather than a behavior problem with the AEs",
"Prescribes a mutual action plan per partner as part of the activation fix"
]
},
{
"id": 7,
"prompt": "Here's our partner one-pager draft for Saltmarsh Identity: 'Partners commit a dedicated named rep, complete our two-day certification within 60 days, register every opportunity within 5 business days, and maintain $250k annual influenced revenue to stay in good standing.' We're taking it to about 40 agencies next month. Can you sharpen the copy and suggest the outreach sequence?",
"expected_output": "A stop on the outreach, with the one-pager identified as all ask and no give, the give-before-ask order supplied, and a written Partner Hypothesis required and validated against 2-3 real prospective partners against a stated proceed threshold before any 40-agency push.",
"files": [],
"expectations": [
"Declines to sharpen the outreach copy or build the 40-agency sequence as the next step",
"Identifies the one-pager as entirely vendor asks with no partner give",
"States the give-before-ask rule: the program must give before it may ask",
"Lists the gives in order, covering a clear way to make money, training to sell well, tools to track activity, and vendor air cover or conflict resolution, before any ask",
"Requires a one-sentence partner value proposition answering what is in this for the partner",
"Requires a written Partner Hypothesis of roughly one to two pages naming the target partner type, why they would engage, and the expected economics",
"Names the Partner Hypothesis as a practitioner framework with attribution rather than presenting it as a universal standard",
"Requires validation against 2-3 real prospective partners before anything is built on it, rather than going to 40 at once",
"States the proceed threshold: at least two prospective partners confirm the joint value proposition and a shared customer overlap exists",
"States that the validation record is not optional and must name who it was tested on and what each confirmed or contradicted",
"Defers formal scored partner profiling until recruiting volume justifies it, using the validated hypothesis for the first cohort instead"
]
},
{
"id": 8,
"prompt": "Copperline Bikes - we make e-cargo bikes, sold direct online, average order $3,900. We want a dealer network: independent bike shops that stock, sell and service us. I want exclusive postcode territories so shops actually invest, a required shop layout and staff script, and I'd like to reuse the SaaS partner program structure my last company ran (registered/silver/gold, deal registration, MDF). I figured we'd also ship dealers the full 180-page brand and operations manual on day one. Sound right?",
"expected_output": "A B2C fork: the territory-plus-operational-control combination flagged as franchising needing counsel, deal registration correctly distinguished from a territory, the opposite documentation defaults named, consumer-specific requirements raised, and what genuinely stays identical named.",
"files": [],
"expectations": [
"Flags that exclusive territories combined with a mandated shop layout and staff scripts drift toward franchising, and recommends franchise counsel rather than a partner program template",
"Names franchising as a separate legal discipline with a mandated pre-sale disclosure document, citing the FTC Franchise Rule where the US applies",
"States that B2B deal registration protects a named account for a time window and is not a substitute for a geographic territory held in perpetuity",
"States that documentation depth is the opposite default on the two sides, so the 180-page manual fits a franchise-shaped model and would be a named failure mode in B2B SaaS",
"Names the deliberately minimal B2B launch kit it contrasts against, covering pitch deck, demo script, pricing sheet, one or two battle cards, and a case study",
"Distinguishes point-of-sale consumer attribution from B2B's multi-month pipeline attribution, and notes that reward triggers and fraud exposure change with it",
"Raises front-line mindshare: winning the individual floor rep's point-of-sale recommendation, often worth more than any rebate structure",
"Raises dealer-locator findability on the vendor's own site as a consumer-side requirement with no meaningful B2B equivalent",
"Names what stays identical across both sides, including deep-qualifying a small first cohort before scaling recruitment and modeling economics from the partner's side first",
"States that a compelling economic offer, not marketing, is what recruits on both sides",
"Poses the consumer-side fork questions, including whether the reward fires at signup or only at verified sale and the sharper fraud trade-off that comes with instant conversion"
]
},
{
"id": 9,
"prompt": "Prepping the kickoff for our brand-new partner program at Halcyon Grid. My plan: (1) convene a partner advisory board of 20 people to shape it, (2) run a two-day journey-mapping workshop across the twelve partner personas we brainstormed, (3) use the numbers to get budget - I've got '70% of partnerships fail within two years' and 'partner-involved deals close 46% faster' in the deck already. Which workshop do I book first?",
"expected_output": "The advisory board deleted for a program with no partners yet, journey mapping placed last with its invalidation risk named, partner-hypothesis validation put first as the only format that can stop the program, and both quoted statistics refused as established fact.",
"files": [],
"expectations": [
"Deletes the partner advisory board from a from-scratch program because there are no partners with activity to convene yet, and names the condition that returns it to the menu",
"States the collaboration-format ranking out loud rather than leaving it implied",
"Gives partner-hypothesis validation as the leading format, ahead of any facilitated workshop",
"Justifies that lead by it being the only format whose output can stop the program",
"Places partner journey mapping last, because it only pays once motions and partner types are fixed",
"Warns that a journey map run before the architecture maps a journey the architecture then invalidates",
"Requires a named journey owner appointed before any mapping workshop, and names the failure of unowned sessions",
"Challenges the twelve personas, requiring the persona count be disciplined down before anything is mapped",
"Refuses to present '70% of partnerships fail within two years' as established fact, identifying it as anecdotal or secondhand",
"Refuses to present the 46%-faster deal-velocity claim as established fact, identifying it as a single vendor's survey of its own ecosystem",
"Flags the shared-customer discovery sprint's compliance cost where it is raised: it exchanges customer lists between two companies, triggers mutual data-sharing terms, and a list that has crossed cannot be un-shared",
"Does not book the two-day journey-mapping workshop as the first action"
]
},
{
"id": 10,
"prompt": "We've been round and round on this for three weeks at Ferrous Labs - $55k ACV data platform, 74% gross margin, direct motion is solid with two years of consistent quota attainment, and our CFO Dana Whitlock is sponsoring it. I need one document I can put in front of finance, legal and sales on Monday so they stop asking me questions. What goes in it and can you just write it?",
"expected_output": "A section-by-section build with candidate approaches and approval gates rather than a single unvalidated document, producing the Program Architecture Spec with every named line, a reason per line, and the three-condition pass threshold applied.",
"files": [],
"expectations": [
"Presents the architecture section by section with an approval gate, rather than emitting the whole document unvalidated in one block",
"Offers 2-3 candidate approaches with trade-offs and an explicit recommendation per section, not a single option",
"Names the deliverable as a Program Architecture Spec and lists the lines it must carry",
"Includes a Readiness line covering direct-motion evidence, ACV band, gross-margin headroom, and the named executive sponsor",
"Includes a Partner value line carrying both the one-sentence proposition and a validation record naming 2-3 prospective partners and what each confirmed or contradicted",
"Includes a Motions line naming which motions were chosen, the order they were chosen in after re-ranking, and every motion a constraint deleted with the trigger that would reopen it",
"Includes a First cohort line with a target size around ten, a qualification bar, and an activation milestone per partner",
"Names the qualification bar as Capability, Capacity, Coverage, Commitment",
"Includes Tiers, Economics, Governance and Horizon lines, with Governance carrying an owner plus reporting line and the deal-registration/conflict policy status",
"States that every line of the spec carries a reason, so a bare decision such as an unexplained tier count is not acceptable",
"Applies the three-condition pass threshold - economics inside gross margin, partner value proposition survived contact with 2-3 real prospective partners, a named executive sponsor exists - and recommends not launching if no architecture passes"
]
}
],
"trigger_queries": [
{ "query": "Design a B2B partner program from scratch for our SaaS company", "should_trigger": true },
{ "query": "How should we structure a channel program?", "should_trigger": true },
{ "query": "We want to build a partner ecosystem program - where do we start?", "should_trigger": true },
{ "query": "What should the architecture of our reseller channel look like?", "should_trigger": true },
{ "query": "Are we even ready for partners?", "should_trigger": true },
{ "query": "Help me decide whether to launch a partner program at all", "should_trigger": true },
{ "query": "We need a channel strategy for our $40k ACV product", "should_trigger": true },
{ "query": "Which partner motions should we run - referral, reseller, tech, or SI?", "should_trigger": true },
{ "query": "Write the program architecture for our new partner program", "should_trigger": true },
{ "query": "How many partner motions should a first program have?", "should_trigger": true },
{ "query": "boss wants a reseller channel and i have no idea where to begin", "should_trigger": true },
{ "query": "we keep getting asked to partner and we have nothing set up", "should_trigger": true },
{ "query": "is it too early for us to do channel?", "should_trigger": true },
{ "query": "how do i stand up a partner program with basically no budget", "should_trigger": true },
{ "query": "our CEO read about ecosystem-led growth and now wants a partner program, help", "should_trigger": true },
{ "query": "what does a partner program even need to have in it", "should_trigger": true },
{ "query": "i've been handed 'build the partner program' and a blank doc", "should_trigger": true },
{ "query": "Three agencies already install our product for their clients and none of them have an agreement with us. What should we do with that?", "should_trigger": true },
{ "query": "We have inbound requests from consultancies wanting to resell us and no way to say yes", "should_trigger": true },
{ "query": "Our direct sales are flat and leadership thinks other companies selling for us will fix it", "should_trigger": true },
{ "query": "Should we let other companies sell our product, and if so how do we set that up?", "should_trigger": true },
{ "query": "Somebody asked whether we have a program for integrators. We don't. Should we?", "should_trigger": true },
{ "query": "What's the minimum structure we need before we sign our first company that resells for us?", "should_trigger": true },
{ "query": "We want other firms bringing us deals but we don't want to build a whole bureaucracy around it", "should_trigger": true },
{ "query": "How do I know if this is worth building a whole function for or if we should just pay finder's fees?", "should_trigger": true },
{ "query": "We have 200 partners and almost no revenue from them. What's structurally wrong?", "should_trigger": true },
{ "query": "Our partner program is being called a failure at month six - is that fair?", "should_trigger": true },
{ "query": "Partners sign up and then go quiet. Why does that keep happening to us?", "should_trigger": true },
{ "query": "Rebuild our partner program structure from the readiness gate up", "should_trigger": true },
{ "query": "Our program has requirements nobody meets and benefits nobody wants", "should_trigger": true },
{ "query": "Direct reps are blocking partner deals - is that a people problem or a design problem?", "should_trigger": true },
{ "query": "Nobody claims our market development funds. Where did we go wrong in the design?", "should_trigger": true },
{ "query": "Partners stopped registering deals with us after we reserved some accounts for ourselves", "should_trigger": true },
{ "query": "Can our gross margin even support a channel?", "should_trigger": true },
{ "query": "How much of our margin should the whole partner program be allowed to cost?", "should_trigger": true },
{ "query": "Our competitor pays resellers 30% - should we just match that?", "should_trigger": true },
{ "query": "Who should the partner program report to at an early-stage company?", "should_trigger": true },
{ "query": "What should we promise leadership the partner program delivers in year one?", "should_trigger": true },
{ "query": "How long before a channel program produces real revenue?", "should_trigger": true },
{ "query": "We want to build a dealer network for our consumer hardware", "should_trigger": true },
{ "query": "How do I design a brand ambassador program structure from zero?", "should_trigger": true },
{ "query": "Is what we're describing a dealer program or actually a franchise?", "should_trigger": true },
{ "query": "What's in it for the partner? I genuinely can't answer that about our own program", "should_trigger": true },
{ "query": "Before we recruit anyone, how do we test whether partners would even want this?", "should_trigger": true },
{ "query": "Referral versus reseller for a product that's hard to implement - which one first?", "should_trigger": true },
{ "query": "We can't hire anyone for this. Does that rule out certain kinds of partners?", "should_trigger": true },
{ "query": "Board wants a partner-sourced number by end of Q2. Which partner types can actually deliver that fast?", "should_trigger": true },
{ "query": "Should we do technology integrations or a reseller motion first?", "should_trigger": true },
{ "query": "Do we even need tiers in our first partner program?", "should_trigger": true },
{ "query": "How many partners should we start with, and how do we qualify them?", "should_trigger": true },
{ "query": "What has to be true before partners are worth doing at all?", "should_trigger": true },
{ "query": "Redesign the qualification criteria for our existing silver and gold partner tiers", "should_trigger": false },
{ "query": "What benefits belong in our gold tier versus platinum?", "should_trigger": false },
{ "query": "Write promotion and demotion rules for our current partner tiers", "should_trigger": false },
{ "query": "How do we migrate our existing 80 partners into the new tier structure?", "should_trigger": false },
{ "query": "Model the P&L for our largest reseller and tell me whether to keep them", "should_trigger": false },
{ "query": "What's the payback period on this one partner relationship?", "should_trigger": false },
{ "query": "Compare partner CAC to direct CAC for this specific integrator", "should_trigger": false },
{ "query": "Write our deal registration policy: protection window, approval SLA, rejection grounds", "should_trigger": false },
{ "query": "How do we split deal credit between partner-sourced and partner-influenced?", "should_trigger": false },
{ "query": "Draft the rules of engagement between our direct AEs and partner sellers", "should_trigger": false },
{ "query": "Two partners are claiming the same account. How do we adjudicate it?", "should_trigger": false },
{ "query": "Write our named-account carve-out policy for channel conflict", "should_trigger": false },
{ "query": "Sequence the certification tracks our resellers need before their first deal", "should_trigger": false },
{ "query": "Build the training roadmap for our partner sales engineers", "should_trigger": false },
{ "query": "Design the per-partner scorecard with weights and red-flag thresholds", "should_trigger": false },
{ "query": "What's the QBR structure for reviewing an underperforming partner?", "should_trigger": false },
{ "query": "Build the underperformance ladder from trigger to de-tier or exit", "should_trigger": false },
{ "query": "Map our existing partner ecosystem and find the coverage gaps", "should_trigger": false },
{ "query": "Build a weighted partner scorecard and prioritization 2x2 for our current partners", "should_trigger": false },
{ "query": "Which partner category should we add next, MSPs or agencies?", "should_trigger": false },
{ "query": "Sequence our partner category expansion over the next six quarters", "should_trigger": false },
{ "query": "Should we list on the AWS marketplace or the Azure one first?", "should_trigger": false },
{ "query": "When should we delist from an app store that isn't producing?", "should_trigger": false },
{ "query": "Rank these six named alliance candidates by expected value and effort", "should_trigger": false },
{ "query": "Write the go/no-go memo on a strategic alliance with a specific platform vendor", "should_trigger": false },
{ "query": "Plan the co-marketing campaign with the partner we just signed", "should_trigger": false },
{ "query": "How should we split leads and attribution on this joint launch?", "should_trigger": false },
{ "query": "Set the commission rate and cookie window for our affiliate program", "should_trigger": false },
{ "query": "Draft the terms and conditions affiliates accept when they join", "should_trigger": false },
{ "query": "Audit this month's affiliate payout run before we send the money", "should_trigger": false },
{ "query": "Build the detection rules for cookie stuffing and self-referral rings", "should_trigger": false },
{ "query": "Design the welcome sequence from affiliate approval to first conversion", "should_trigger": false },
{ "query": "What metrics belong on our affiliate performance dashboard?", "should_trigger": false },
{ "query": "Does this influencer caption meet FTC disclosure rules?", "should_trigger": false },
{ "query": "Write the recruitment emails we send to prospective affiliates", "should_trigger": false },
{ "query": "Draft the campaign brief for the creator we just signed", "should_trigger": false },
{ "query": "Shortlist creators for our next campaign and score them", "should_trigger": false },
{ "query": "Prepare my negotiation positions on usage rights and exclusivity with a creator", "should_trigger": false },
{ "query": "Write the first-touch pitch DM to this YouTuber", "should_trigger": false },
{ "query": "Set the KPIs and attribution method for one influencer campaign", "should_trigger": false },
{ "query": "Design the reward for our refer-a-friend program - cash or account credit?", "should_trigger": false },
{ "query": "Set velocity limits and reward holds to stop referral abuse", "should_trigger": false },
{ "query": "Install a referral program on our Shopify store", "should_trigger": false },
{ "query": "How do I break into partnerships from a sales role?", "should_trigger": false },
{ "query": "Write the job description and interview loop for a channel manager", "should_trigger": false },
{ "query": "Which partnerships skill should I use for the task I'm about to start?", "should_trigger": false },
{ "query": "Which PRM platform should we buy to run our partner program?", "should_trigger": false },
{ "query": "Set up our partner portal and configure single sign-on for it", "should_trigger": false },
{ "query": "Configure partner offers on our AgentExchange platform listing", "should_trigger": false },
{ "query": "How do I join the Shopify Partner program as an agency?", "should_trigger": false },
{ "query": "Build a Slack channel structure for our internal partnerships team", "should_trigger": false }
]
}
references/b2b-b2c-contrast.md›
# B2B vs B2C Partner Program Design
Consumer-side analogs exist in dealer/distributor networks, franchising, and brand ambassador programs. The economic machinery of B2B channel (margin structures, deal registration, market-development funds, tiering, partner segmentation) largely descends from consumer practice: co-op advertising funds are the direct ancestor of MDF.
What follows separates what is genuinely shared from what genuinely diverges. The side-by-side comparison itself is a research synthesis across independently sourced material, not a claim from any single citable source: treat it as well-grounded inference.
## Identical across B2B and B2C (sourced convergence)
- **Deep-qualify a small first cohort before scaling recruitment.** Explicit in B2B ("recruit the first ~10 with deep qualification") and in franchising ("the wrong franchisee costs far more than an empty territory").
- **Model economics from the partner's side first.** The partner's/franchisee's/dealer's own cost structure and required profitability come before the vendor sets any number - present in the B2B envelope method, franchise unit-economics requirements, and dealer procurement rationality alike.
- **A compelling economic offer, not marketing, is what recruits.** Stated near-verbatim on both sides: the B2B partner-value-proposition-first principle, and the franchise finding that "what ultimately sells franchises is not advertising - it is strong unit economics and franchisee validation."
- **The management architecture is structurally the same everywhere**: recruitment, onboarding, performance tracking, ongoing relationship management appear in every format.
## Genuinely different
- **Legal weight and territory (franchising).** Franchising is a separate legal discipline: a federally mandated disclosure document (in the US, under the FTC Franchise Rule) filed before any sale, and legally exclusive geographic territories negotiated as a core deal term. B2B SaaS has no equivalent instrument - deal registration is the closest analog, and it deliberately protects a _named account_ for a _time window_, never a geography in perpetuity. Exclusive territories trade early partner investment against the vendor's expansion speed. That trade-off simply does not appear in B2B SaaS program design.
- **Documentation depth at launch - opposite defaults.** Franchising requires an exhaustive, "constitution"-level operations manual before the first franchisee signs, detailed enough for a stranger to execute correctly. B2B SaaS explicitly warns against the identical behavior: an over-built day-one enablement package is a named failure mode, and the recommended launch kit is deliberately minimal (pitch deck, demo script, pricing sheet, a couple of battle cards, one case study). Same instinct, opposite correct answers, because a franchisee replicates an entire operating system while a B2B partner grafts one motion onto an existing business.
- **Attribution and conversion tooling.** Consumer ambassador and dealer-to-consumer motions convert instantly at point of sale and need e-commerce-style tracking. B2B referrals run through a multi-month, multi-stakeholder CRM pipeline from introduction to closed-won. This is a genuinely different tooling and attribution requirement, not a difference of tone: reward triggers, fraud exposure, and reporting all change with it.
- **Front-line mindshare and findability (consumer-facing dealers).** Dealer program design must win the individual floor rep's point-of-sale recommendation, called out as often worth more than any rebate structure, and must invest in dealer-locator findability on the vendor's own site. Neither requirement has a meaningful B2B SaaS equivalent. Conversely, the distributor/procurement side of the same dealer network behaves like B2B: rational, financial, program-economics-driven.
- **Persuasion register.** Consumer programs lean emotional, visual, user-generated content. Ambassadors are often everyday enthusiasts compensated with non-cash rewards. B2B programs lean expertise-led (credentialed practitioners, webinars, professional networks) because the target is a considered, multi-stakeholder purchase surrounded by trusted advisors, not impulse advocacy.
## Design-time fork questions for a consumer-side user
When the interview reveals B2C, settle these before reusing any B2B structure:
1. Individual consumers or a buying committee on the receiving end?
2. Conversion at signup/purchase, or over a CRM-tracked multi-month cycle?
3. Reward at signup or only at verified sale - the fraud-risk trade-off is far sharper with instant conversion?
4. Everyday enthusiast or credentialed expert as the partner?
5. Does the model drift toward operational control and territory - if yes, it is franchising, and it needs franchise counsel, not a partner program template?
references/collaboration-formats.md›
# How Channel Professionals Workshop a Program Design
Adapt the skill's brainstorming mode to these documented formats instead of generic ideation. Propose the format that fits the design step at hand. Never run one as theater.
The formats are ordered below by evidence bought per hour of coordination. The axes disagree, so read all of them:
- efficiency: `partner-hypothesis validation > joint value session > shared-customer sprint > discovery by immersion > journey mapping`
- effort: `shared-customer sprint > discovery by immersion > journey mapping > partner-hypothesis validation > joint value session`
- value: `partner-hypothesis validation > shared-customer sprint > discovery by immersion > joint value session > journey mapping`
- compliance cost: `shared-customer sprint > all others` (only the sprint exchanges customer lists between two companies; it triggers mutual data-sharing terms, and a list that has crossed cannot be un-shared)
This order is a default, not a law: it shifts with who is running it. A founder who already knows the direct motion cold deletes discovery by immersion rather than demoting it.
A user who arrives with a written hypothesis already tested on real partners starts at the joint value session. Where a technology motion is the candidate, the shared-customer sprint moves up past the joint value session, because customer-verified overlap is the evidence that motion actually turns on.
## Partner Hypothesis validation (the default for a new program)
Leads the ranking because it is the only format whose output can stop the program: a 1-2 page written hypothesis (target partner type, why they would engage, expected economics) tested in direct conversations with **2-3 real prospective partners**. Effort is a written page plus a handful of calls spread over a few weeks, with no facilitation and no standing commitment.
Threshold to proceed: at least two prospective partners confirm the joint value proposition and a shared customer overlap exists. This is a documentation-and-conversation exercise, not a facilitated multi-day workshop: resist upgrading it into one before there is anything to facilitate.
## Joint value proposition working session
Cheapest format on the menu - an hour plus a pre-read - and it buys language a partner rep can repeat without you in the room.
- Underlying tool: the Value Proposition Canvas - partner-customer jobs, pains, gains first, then the value map.
- Test every value statement against the customer "so-what" question. A neutral third party asking it works better than the vendor grading its own homework.
- Output: short narratives at fixed lengths (25 / 50 / 100 words) that survive being read by someone who attended no meeting.
- One documented discovery-call variant: a 30-60 minute working session with a joint pre-read sent 24 hours ahead (overlap account list, draft joint value frame, proposed motion, three open questions), each segment ending in a captured decision.
## Rapid shared-customer discovery sprint
The strongest evidence any format produces, and the heaviest: a documented ~4-week method with one prospective partner, two-sided coordination throughout.
- Agree ground rules: customer count, selection criteria, learning goals.
- Identify overlapping customers from both sides' data.
- Split the list.
- Interview shared customers about how they use both products together, the value received, and their pain points.
Strong fit for validating a technology-partner motion before committing program structure to it. Settle the data-sharing terms before either side sends a customer list.
## Discovery by immersion
Three weeks of the architect's own calendar and near-zero coordination with anyone external: sit inside the direct sales motion, attending sales meetings and standups as an observer before contributing. The architecture depends on what the direct motion actually is, not what the deck says it is - which is also why this format is worthless to someone who already runs that motion daily. A published first-90-days variant adds concrete goals: ten prospective-partner conversations by day 90 and a specific partner-sourced pipeline target for year one.
## Partner journey mapping
Last on the menu because it only pays once motions and partner types are fixed; run before the architecture, it maps a journey that the architecture then invalidates.
- Appoint a **named journey owner before the workshop** - the documented failure of unowned sessions is "a flurry of sticky notes and little else."
- One map per partner type and critical persona, stages typically recruitment → onboarding → growth → advocacy.
- Discipline the persona count first: a real case narrowed 300+ candidate personas to 4 before mapping anything.
## Not on the menu: Partner Advisory Board
Deleted for a from-scratch program rather than ranked last - there are no strategic partners to convene yet, so the format has no input to work with. It returns to the menu once the program has partners with a year of activity behind them and genuinely undecided questions to put to them.
When it does return, both sides of the evidence apply.
**Favorable case:** 10-25 attendees, one _senior_ person per partner organization (not day-to-day account managers), membership rotated on staggered ~2-year terms, and day-to-day vendor channel staff kept out of sessions so conversation stays strategic. Output is a published summary of what the vendor will, and explicitly will not, act on.
**Skeptical case**, from a practitioner source and worth taking seriously: the "Partner Advisory Board Myth", a standing forum that produces no actionable outcomes, rewards the same partners yearly, and creates a false impression of consultation. Bring only genuinely undecided topics: "if every agenda item is a fait accompli, partners will stop investing their time."
## Running the interview itself
- One question per message; multiple-choice options whenever the answer space is enumerable.
- Skip anything already answered; never re-ask to fill a template.
- For each architecture section, present 2-3 candidate approaches with trade-offs, an explicit recommendation, and the order you rank them in - a menu without a recommendation outsources the architect's job to the user, and an unranked one outsources it to whichever option you listed first.
- Validate section by section; do not present a finished architecture in one block.
- Explicit approval gate before emitting the final spec.
references/decision-sequence-and-frameworks.md›
# Decision Sequence and Named Frameworks
The convergent finding across analyst and practitioner sources: strategy and architecture before program mechanics. Every ordering rule below exists because the reverse order is a documented failure pattern.
## The sequence, with why
1. **Readiness before anything.** The program is close to the _last_ piece of the partnering puzzle: business and channel strategy must be settled first. Skipping the gate is itself a named structural mistake.
2. **Partner value proposition before the vendor's own asks.** Vendor-centric framing ("we need resellers to hit pipeline targets") fails. Partner-centric framing works because partners partner so their own business thrives.
Give-before-ask order:
1. A clear way to make money: compensation, funds, deal protection.
2. Training to sell well.
3. Tools to track activity.
4. Vendor air cover and conflict resolution.
5. Only then, asks: dedicated reps, qualification, registration discipline, a performance bar.
3. **Partner types/motions before the Ideal Partner Profile.** The rigorous sequence (analyst-corroborated) is business objective → partner types → profile built _within_ a type. A further practitioner wrinkle: defer formal scored profiling until the recruiting pool is large enough to need repeatable scoring - for the first ~10 partners, a validated Partner Hypothesis stands in for a formal profile.
4. **Tier purpose before tier structure, and criteria and benefit delta as a pair.** A tier exists when something measurable changes at the boundary. "Criteria with no benefits delta is a leaderboard nobody plays, and a benefits delta with no criteria is favoritism wearing a logo." Practical method: reverse-engineer the top tier from your existing best partners' actual behavior rather than inventing tiers abstractly.
5. **Deal-registration and conflict policy before the first selling partner signs.** The policy "needs to exist before the conflict, not after it." Registration should trigger a concrete support bundle (engineering hours, proof-of-concept funding, executive sponsorship): the support, not the margin point, is what partners remember a year later.
6. **Activation designed before recruitment scales.** Recruit a deliberate first cohort of ~10, deep-qualified on Capability, Capacity, Coverage, Commitment (the 4C qualification, a practitioner framework). Give each partner a mutual action plan targeting a registered deal by ~week 8 and a closed deal or qualified pipeline by ~week 12.
## ACV as a deletion gate on the motion menu
This table does not rank motions - SKILL.md step 4 owns the ranking, and a second order here would contradict it. ACV only deletes rows from that menu, and a deleted motion is excluded outright rather than demoted to "later".
One practitioner's published bands (directional, not a law - a single operator's guide, internally consistent but not independently validated):
| ACV band | Motions that survive the gate | Deleted at this band |
| ------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------- |
| Under ~$3-5k | Referral only | Reseller, services-SI, anything human-heavier - it cannot be funded |
| ~$5-20k | Referral plus light co-selling | Reseller, services-SI |
| ~$20-50k | Referral, technology, VAR/SI on the partner's side - their margin starts covering their own acquisition and delivery - with the vendor still funding most of the motion | Full reseller motion |
| Above ~$50k | Nothing deleted: resellers, systems integrators, service partners are fundable on both sides | (none) |
The logic generalizes even if the thresholds shift by market: a reseller earning 25% of a $20k deal nets $5k, which must cover their own acquisition and delivery costs - at low ACV the partner's side of the math fails before yours does. Record each deletion with the ACV trigger that would reopen it.
## Tier count
This is evidence for the tier-count decision, not the decision itself - `mbfinotti/partnerships-skills@partner-tiering` sets the count against partner-base size, and its default is 3 collapsing to 2 below roughly 50 partners. Hand it what follows rather than fixing a number here. Deliberately unranked: an efficiency order over tier counts would be false precision, since the ratio depends entirely on the partner-base size that skill measures and this one does not know.
The evidence favours **few tiers** (an entry tier and an earned tier) over the classic four-metal ladder. Multi-tier ladders create demotivating middle tiers, and most programs' extra tiers are "mostly vanity."
The metals model remains defensible for one case: a high-volume reseller motion, where analyst opinion holds it still drives performance. Real-world direction of travel supports the default: Salesforce retired a four-tier consulting hierarchy for a two-tier Select/Summit model in 2026, and Microsoft retired its Silver/Gold tiers for competency-based designations, both sourced case examples, not endorsements of any vendor.
## Named frameworks, with honest status flags
Use a framework only where it genuinely fits the user's situation. Never force-fit one.
| Framework | Status | Use for |
| --------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------- |
| Build / Sell / Service partner typology (popularized by Jay McBain, Forrester-era) | Widely adopted analyst-popularized lens; note the exact three-word phrase is industry shorthand without one canonical primary text | First-pass sorting of candidate motions |
| Trifurcated channel: influence / transact / retain (McBain, Canalys) | Genuine analyst framework | Reminding the design that most partner value sits before and after the transaction |
| Ideal Partner Profile | De facto common practice, analogous to the ideal customer profile; most published templates come from vendors | Formal partner scoring, once volume justifies it |
| Partner Value Proposition | Common practice; a one-sentence answer to "what is in this for the partner" | Step 3 of the workflow |
| Partner journey mapping | Common practice; a formal paid analyst framework also exists | Operationalizing engagement per partner type, post-architecture |
| Value Proposition Canvas | Established general-strategy tool, commonly used in joint-value workshops | Joint value proposition work with a specific partner |
| Partner Hypothesis (PartnerStandard, Bernhard Friedrichs) | Practitioner framework, worth naming with attribution | The pre-program business case validated on 2-3 real partners |
| Minimum Viable Ecosystem (same author) | Practitioner framework | Choosing the smallest set of motions that proves the thesis |
| 4C qualification: Capability, Capacity, Coverage, Commitment | Practitioner framework | Scoring the first cohort |
| ROCKER: Relegate, Offset, Coverage, Knowledge, Expertise, Relationships (Channelnomics) | Practitioner framework | Defining the partner's _role_ in the go-to-market before designing anything downstream |
**Vendor constructs - useful vocabulary, not neutral standards.** Flag these as marketing-adjacent when the user brings them up:
- **Ecosystem-Led Growth**: coined and promoted by the CEO of the company selling the associated software, partly an extended case study of that company.
- **Ecosystem multiplier models** ("$X of partner value per $1 of product"): vendor-commissioned analyst studies. Attribute, never treat as a neutral benchmark.
- **Partner-program maturity models**: multiple vendors each publish a competing 4-5 stage version. Useful as a diagnostic vocabulary, not a scorecard to race up.
- Several formally named analyst program-design and journey-map models are real but paywalled. Their public abstracts are directionally useful, not verbatim guidance.
references/economics-and-benchmarks.md›
# Program Economics Envelope and Benchmarks
This file covers the _program-wide_ economic shape - whether gross margin can fund a channel at all, and what compensation bands exist per motion. Modeling one partner relationship belongs to `mbfinotti/partnerships-skills@partner-economics`.
## The envelope method
Partner margin is a purchase, not a gift. Derive it from four questions, in order:
1. **What operational task is the partner taking over?** Sales motion, delivery, support, local-language coverage, vertical expertise, implementation, ongoing success work.
2. **What would that task cost in-house?** The partner margin comes out of this saving - "if you're paying margin for work you would have done anyway, you're not buying anything from the partner."
3. **What is the partner's own cost structure?** The offer must leave the _partner_ a business: a 25% margin on a $20k deal is $5k, which must cover their acquisition and delivery costs. Model their side before setting your number.
4. **How does the resulting cost-of-sale compare to direct CAC?** A channel whose all-in cost exceeds the direct motion's is a strategy tax, not leverage.
Two envelope-level rules that belong in the architecture, not in later tuning:
- **Direct-rep compensation neutrality.** Reps must never earn less because a partner touched the deal. Misaligned comp is a root cause of channel conflict, not a symptom.
- **Track cost-to-serve per partner from day one.** Most programs track partner revenue but not the discounts, funds, headcount, and training behind it. A partner "generating" $2M can net half that after true costs. The program P&L is the number finance will eventually ask for.
Copying a competitor's margin table is a named structural mistake - their economics bought different work.
## Compensation bands by motion
These are per-motion bands to price a motion already chosen, not a ranking of motions - SKILL.md step 4 ranks the menu, and richer bands here do not mean a better motion.
Almost all bands below are vendor-blog or practitioner consensus - remarkably convergent, but not audited analyst data. Treat as starting points for first-principles math, never as answers. The two analyst-grade exceptions are marked.
| Motion | Band | Sourcing quality |
| ------------------------------------- | -------------------------------------------------------- | ------------------------------------------------- |
| Referral commission | 10-20% of first-year value (some cite 5-20, mode ~10-15) | Vendor/practitioner consensus |
| Reseller margin | 20-40% | Vendor-blog consensus |
| Value-added reseller | 25-35% | Vendor-blog example |
| Managed service provider | 20-40%, usually recurring | Vendor-blog consensus |
| Distributor | 8-15% (wider ranges cited) | Vendor-blog, wide variance |
| Hyperscaler cloud resale | 2-10% | **Analyst-sourced (Canalys) - higher confidence** |
| Deal-registration extra discount | +5-15% on top of base | Vendor-blog consensus |
| Market-development funds | 2-5% of channel revenue | Vendor-blog |
| Legacy on-premise enterprise discount | ~40% off list (≈20% cost coverage + 20% profit) | Long-standing practitioner rule of thumb |
Program-shape reference points, same caveats:
- Deal-registration protection windows commonly 90-180 days (account-scoped and time-bound, never territorial).
- ~1 partner manager per 25-50 partners as a staffing rule of thumb, with 10-20 _active_ resellers per manager the sustainable engagement range per one practitioner.
- Indirect share of B2B IT spend is large (analyst-sourced ~two-thirds and, notably, recently _declining_ as AI infrastructure spend bypasses the channel), so "everyone does channel" is true and still not an argument that this company should.
## The success horizon
Set at design time, because the most common structural killer is judging a channel on the 30-90 day cycle of direct sales:
- **Year 1: leading indicators only.** Partners recruited _and activated_, first deals registered, time-to-first-deal (a sub-90-day target per partner is one practitioner's benchmark), enablement completion. Expect little net revenue, since relationship maturity takes 2-3 years.
- **Year 2: meaningful partner-sourced pipeline** and the beginnings of a repeatable co-sell play.
- **Year 3: material revenue contribution.** The frequently cited band for maturing programs is 20-30% of revenue, with ~28% cited for mature ones - vendor-survey numbers, directionally useful, not guarantees.
Course-correction triggers worth writing into the spec:
- No partner-sourced pipeline building by end of year 1 → revisit partner fit and enablement, not headcount.
- Most recruited partners dormant → stop recruiting, fix activation.
- Executive sponsor not defending the program by year 2 → existential risk, address before requesting budget.
Widely recirculated deal-velocity claims ("deals close 46% faster / 53% more likely with a partner") come from a single vendor's survey of its own ecosystem. Attribute them if cited, but never present as established fact.
references/readiness-and-failure-modes.md›
# Readiness Gate and Failure Modes
## The full readiness gate
Ask these before any design work. The gate exists because the single best-documented failure cause is launching a partner program to compensate for a direct motion that is not yet repeatable - "partners do not solve a broken direct motion; they amplify a working one."
1. Can you clearly explain your ideal customer?
2. Have you proven direct sales closure capability - not one heroic founder deal, a repeatable motion?
3. Is your go-to-market consistent and repeatable enough for a third party to copy? (Sales, onboarding, support processes a partner could actually replicate.)
4. Does your ACV support human-heavy partner work? Very low ACV funds referral-only at most; reseller and systems-integrator motions need materially more, because the partner's own cost of sale has to clear before yours does.
5. Are you expecting partners to solve a growth problem? **A "yes" here is a disqualifier by itself.**
6. Is there organic momentum - partners already servicing your customers, inbound requests to resell or integrate? No external interest at all suggests there may not be enough value for both parties yet.
7. Is there internal support: a named executive sponsor who will defend the program in cross-functional meetings, and access to sales, product, and success resources? A practitioner heuristic holds that the single most predictive failure indicator is whether the revenue chief actively defends the partner program.
8. Can gross margin fund partner margin and still leave the vendor a business?
9. Opportunity cost: what else could the company invest in right now instead?
No fixed quantitative launch checklist exists: readiness varies with market, product, and go-to-market maturity, so the call is a judgment made against the questions above. What is fixed: any disqualifier above means "not yet." Say so, name what must change, suggest the lighter motion that still fits (often referral-only or plain integration partnerships without a formal program), and stop.
## Failure catalog for existing programs
Structural causes first; tooling is rarely the reason a program dies - "the software is rarely the reason a partner program ends; the program is." This catalog is grouped by mechanism and its row order carries no priority: when several failures apply at once, fix them in the order ranked in SKILL.md's Diagnosing an Existing Program section.
| Failure mode | Mechanism | Fix |
| ------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------- |
| Paper partnerships | Agreement signed without either side knowing how they benefit each other; dies 60-90 days into onboarding | Validate a partner value proposition per partner before signing; no validation, no signature |
| Recruitment over activation | Signed-partner count tracked instead of active-after-90-days; the vanity metric masks a dead program | Stop recruiting; make activation the core metric; mutual action plan per partner |
| Cosmetic tiers | Badges with no economic differentiation at the boundary; partners notice fast | Rebuild criteria + benefit delta as a pair, or drop to a single tier honestly |
| Unclaimed program funds | Roughly half of allocated market-development funds go unredeemed (weakly sourced but widely repeated): claims friction, activities that don't match how partners operate | Design funds and deal registration together under one owner; funded leads get automatic registration approval |
| Reversed enablement sequencing | Shipping dozens of decks and a 200-page reference on day one | Minimal launch kit: pitch deck, demo script, pricing sheet, 1-2 battle cards, one case study |
| House-account conflict | Vendor keeps reserved accounts and competes on registered deals; partners stop registering and the program dies from inside | Honor registration without carve-outs; write the conflict policy before the first selling partner |
| Overloaded partner managers | Portfolios beyond the sustainable ~10-20 active resellers per manager; engagement collapses into triage | Cap portfolios in the architecture; fund headcount before recruiting past the cap |
| Annual-operating-plan trap | A 50-page plan written in Q4, ignored after Q1 | One-page quarterly mutual action plan with named owners on both sides |
| Advisory-board theater | A standing forum with no undecided topics on the agenda; false consultation without obligation to act | Convene only with genuinely open questions; publish what will and won't be acted on |
| Misaligned ownership | Channel under the sales chief inherits quarterly KPI horizons; a function that matures over 2-3 years gets killed at its first budget review | Reset the reporting line or explicitly contract a multi-year horizon with leadership |
| Penalized direct reps | Reps compensated less on partner-involved deals block those deals | Compensation neutrality on partner deals - an architecture decision, set before launch |
Widely cited failure-rate statistics ("70% of partnerships fail within two years") are anecdotal or secondhand. Do not present them as established fact. The "roughly 80 of every 100 recruited partners go inactive" figure is one practitioner's estimate: credible, not rigorous.
## Negative example - what not to do
A $4k-ACV SaaS vendor with flat direct sales launches a reseller program: four metal tiers, a 30% margin table copied from a competitor, a partner portal, and a recruitment target of 100 signings in year one.
Every architecture decision here is a documented failure:
- Partnerships deployed to fix a broken motion.
- An ACV that cannot fund a reseller's own cost of sale.
- Margin copied instead of derived from work the partner takes over.
- Tiers before any validated partner value proposition.
- Recruitment volume as the success metric.
- A program judged, and killed, at month 9 on direct-sales timelines.
The correct output for this user is the gate's "not yet": fix the direct motion, consider referral-only in the interim, and revisit when ACV and repeatability support more.
SKILL.md›
---
name: partner-channel-program
description: Design the overall structure of a B2B partner or channel program from scratch - a readiness gate that can conclude not yet, partner value proposition, which motions to run (referral, reseller, services, technology), tier count, benefit stack, program-wide economics envelope, governance, and success horizon. Covers B2C analogs (dealer, franchise, ambassador). Use whenever the user mentions building a partner program, channel program design, launching a channel, a partner ecosystem program, or asks whether they are ready for partners, even if they never say program. Do NOT use for reworking tiers inside a program that already exists - use mbfinotti/partnerships-skills@partner-tiering instead.
license: MIT
metadata:
author: Maya-Beth Finotti
version: "1.0.6"
---
# Partner Channel Program
Design the overall architecture of a B2B partner/channel program from scratch. This skill is the architect: it decides whether a program should exist at all, which structural components it needs - motions, tiers, economics envelope, governance, horizon - and how they fit together, then hands the detailed design of each component to a specialist skill (see Reference).
The field's best-documented failure cause is structural, not tactical: launching a program to fix a direct motion that is not working. Partners amplify a working motion; they do not fix a broken one. The sequence below runs strategy before structure, and its gate is allowed to answer "not yet."
## Interview
Ask before designing anything. Ask one question per message, offering multiple-choice answers when possible. Skip a question only when the user already gave the answer.
- Is the business B2B or B2C? For consumer businesses (dealer, franchise, ambassador models) part of the design genuinely diverges - see [references/b2b-b2c-contrast.md](references/b2b-b2c-contrast.md).
- What is the ACV or price point of what partners would sell or influence?
- Is the direct sales motion already repeatable, with product-market fit proven? What is the evidence?
- What is the gross margin, and how much of it could fund partner compensation?
- Do informal partners already exist - agencies, integrators, consultants servicing your customers - or has anyone approached you about partnering?
- Who will own the program day to day, and who do they report to?
- What resourcing is committed: dedicated headcount, budget, and an executive sponsor by name?
- How are direct reps compensated today on a deal a partner touches - neutral, penalized, or undefined?
- What does success mean, and by what date must the first partner-sourced result land? Name the metric (revenue share, pipeline, coverage), the date, and who judges it then.
- Do you want a one-off win this year or a compounding asset? Both are legitimate answers; they select different motions.
- Given that resourcing, what is the effort ceiling this program must live inside - hours per week for the owner, whether standing headcount is fundable at all, and how much of the design must stay reversible?
Those last three answers re-rank step 4's motion menu - carry them there instead of re-asking:
- A hard date inside two quarters deletes reseller and services-SI: neither produces revenue that fast.
- A compounding mandate promotes technology partnerships and reseller coverage.
- A number needed this year promotes referral.
- No fundable standing headcount deletes reseller and services-SI again, and a team of one leaves referral plus at most one integration.
## Workflow
1. Run the Interview; collect every answer the gate and the architecture depend on.
2. Apply the readiness gate. Disqualifiers, any one of which means the honest answer is "not yet":
- The direct motion is not repeatable, or partnerships are expected to fix a growth problem.
- ACV too low to fund human-heavy partner work (very low ACV supports referral-only at most - reseller/SI motions need far more).
- Gross margin that cannot fund partner margin.
- No named executive sponsor.
Say so, name what must change and the lighter motion (if any) that still fits, and stop. Full gate: [references/readiness-and-failure-modes.md](references/readiness-and-failure-modes.md).
3. Write the partner value proposition before any vendor ask - a one-sentence answer to "what is in this for the partner", plus the economic shape behind it. A program must give (a clear way to make money, training, deal protection, vendor air cover) before it may ask (dedicated reps, certifications, registration discipline). Capture it as a Partner Hypothesis - 1-2 pages naming the target partner type, why they would engage, and the expected economics - to be validated with 2-3 real prospective partners before anything is built on it.
4. Choose partner motions before any partner-profile work, from this menu, ranked by value returned per unit of effort - not by ceiling, and not cheapest-first. Say the order out loud to the user, then pick the minimum viable set, not every motion. The axes disagree, so read all of them:
- efficiency: `referral > technology > services-SI > reseller`
- effort: `reseller > services-SI > technology > referral`
- value ceiling: `reseller > services-SI > technology > referral`
- time-to-effect: `reseller > technology > services-SI > referral`
- compliance cost: `reseller > services-SI > technology == referral` (tie argued: neither a technology nor a referral motion creates a contractual margin commitment or third-party delivery liability, and both are terminable in a paragraph)
Each motion, with its trade-off:
- **Referral** - near-zero standing effort: a commission rule, a link or a form, one tracking field. Buys warm introductions into accounts you already know how to sell. Low ceiling, adds no delivery capacity. Default first rung unless a constraint below deletes it.
- **Technology** - a quarter of engineering, then a standing maintenance obligation, and no margin paid out. Buys durable surface: an integration keeps producing joint accounts and retention after the launch push stops. Slowest of the light motions to first revenue, so it needs the compounding mandate from the Interview to justify itself.
- **Services-SI** - enablement content, a delivery-quality bar, and standing partner-manager time. Buys delivery capacity you did not hire and unblocks deals that stall on implementation. Pays only where implementation is a real buying obstacle; where it is not, this is coordination bought for nothing.
- **Reseller** - a standing job: agreements, a margin table, order flow, deal registration, conflict policy, direct-rep comp neutrality, legal sign-off on resale terms. Buys coverage - segments and geographies you cannot reach directly - at the highest ceiling and the longest lead time. Hardest to reverse: unwinding it means re-papering signed agreements.
What this order starves: reseller and services-SI. Both top the value-ceiling line and top the effort line at once, so the ratio puts them last in every round. A program that only ever obeys the efficiency line stays a referral program permanently: the coverage direct cannot reach and the delivery capacity that unblocks stalled implementations both stay unbought.
Promote reseller above the light motions when reaching a named segment or geography is the stated goal of the program and standing headcount is funded. Promote services-SI when implementation shows up as a named reason in lost or stalled deals, not when an SI relationship merely exists. Neither is promoted by ambition, and neither wins a ratio it is built to lose.
Constraints delete rows rather than demoting them. A deleted motion goes on the spec's Motions line with the trigger that would reopen it:
- An ACV that cannot fund the partner's own cost of sale deletes reseller and services-SI outright.
- A product that cannot be resold for contractual or technical reasons deletes reseller.
- No fundable standing headcount deletes both.
This order is a default, not a law: it shifts with context and with who executes it. Re-rank it against what you already know about this user before presenting it:
- An agency or SI already servicing their customers promotes services-SI to the top, because the recruitment effort is already spent.
- An existing platform-marketplace relationship promotes technology for the same reason.
- A team of one collapses the menu to referral plus at most one integration.
Sort borderline candidates with the build/sell/service typology and the influence/transact/retain lens. Ideal Partner Profile detail comes later, inside a chosen motion, and formal scored profiling waits until recruiting volume needs it. See [references/decision-sequence-and-frameworks.md](references/decision-sequence-and-frameworks.md).
5. Decide whether to tier, and what differentiation is for, before designing any tier. A tier exists only when something measurable changes at its boundary, and criteria and benefit delta are designed as a pair:
- Criteria with no benefits delta is a leaderboard nobody plays.
- A benefits delta with no criteria is favoritism wearing a logo.
Hand the tier count, criteria, and benefit design to `mbfinotti/partnerships-skills@partner-tiering`, which owns all three: decide only whether to tier and what the differentiation is for here. Two tiers is the right instinct for a small or new program, and four is usually vanity, but that skill sets the count against partner-base size. Do not fix a number here that it will re-derive, or signed partner agreements get re-papered.
Do not rank tier counts by efficiency here either, since a second order on this page would contradict the one that skill derives from a base size you do not yet know.
6. Set the program-wide economics envelope: partner margin must come out of a cost the partner takes off your hands ("if you're paying margin for work you would have done anyway, you're not buying anything"), and total program cost must fit inside gross margin. Neutralize direct-rep compensation on partner deals here, not later.
Hand per-partner modeling to `mbfinotti/partnerships-skills@partner-economics`. See [references/economics-and-benchmarks.md](references/economics-and-benchmarks.md).
7. Fix governance: a named owner and reporting line, and a deal-registration/channel-conflict policy written before the first selling partner signs, not after the first dispute. Practitioners argue for CEO-reporting at early stage, since ownership under the sales chief imposes quarterly horizons on a function that matures over 2-3 years. Hand rule detail to the co-selling and conflict skills (see Reference).
8. Set the success horizon at design time: year 1 judged on leading indicators only (partners activated, first deals registered, time-to-first-deal), year 2 on partner-sourced pipeline, year 3 on material revenue share. The most common structural error in this field is judging a channel on direct-sales 30-90 day cycles; the horizon decision is part of the architecture, made now. Hand the measurement system itself to `mbfinotti/partnerships-skills@partner-performance`.
9. Present the architecture section by section - for each, 2-3 candidate approaches with trade-offs and an explicit recommendation - and validate each section with the user before moving to the next. Where the profession's own collaboration formats fit, propose them in the order that file ranks them - evidence bought per hour of coordination, leading with partner-hypothesis validation: [references/collaboration-formats.md](references/collaboration-formats.md).
10. After explicit approval on every section, emit the Program Architecture Spec (next section) and check it against the pass threshold. Iterate until it passes; if no architecture passes, recommend not launching and record why.
11. If your harness has persistent memory, memorize the approved architecture's key decisions, economics envelope, and gate answers - later tactical runs (tiering, enablement, performance) start from them instead of re-interviewing.
## The Program Architecture Spec
Deliver every from-scratch engagement as this artifact - a decision record the user can hand to finance, legal, and sales:
```
PROGRAM ARCHITECTURE - <company>, <date>
Readiness : direct-motion evidence, ACV band, gross-margin headroom, named executive sponsor
Partner value: one-sentence partner value proposition + validation record (2-3 named
prospective partners, what each confirmed or contradicted)
Motions : which of referral / reseller / services-SI / technology, why each, and the
efficiency order they were chosen in after re-ranking for this user; motions
deleted by a constraint, and the trigger that would reopen each
First cohort : target size (~10), qualification bar (capability, capacity, coverage,
commitment), activation milestone per partner
Tiers : whether to tier + what differentiation is for; count set by partner-tiering,
or "no tiers - revisit
when <condition>"
Economics : compensation shape per motion, program cost envelope vs gross margin,
direct-rep comp neutrality statement
Governance : owner + reporting line, deal-registration/conflict policy status
(written before first selling partner)
Horizon : year-1 leading indicators, year-2 pipeline expectation, year-3 revenue
share target; review cadence and who judges
```
Three rules about the spec itself:
- **Every line carries a reason.** "Two tiers because the only motion is referral and nothing measurable changes at a third boundary" is reviewable; a bare tier count is not.
- **The validation record is not optional.** A partner value proposition that has never met a real prospective partner is a guess; the spec must name who it was tested on and what happened.
- **The spec has a pass threshold.** The economics must fit inside gross margin, the partner value proposition must have survived contact with 2-3 real prospective partners, and a named executive sponsor must exist. Iterate motions, tiers, or economics until all three hold - if no architecture passes, recommend not launching.
## Diagnosing an Existing Program
When the user arrives with a running program instead of a blank page ("we have 200 partners and no revenue"), diagnose structure before tactics. Several rows below usually apply at once; the table is a lookup, so the fix order is stated here rather than implied by row order:
- efficiency: `horizon reset > conflict and registration policy > activation over recruitment > funds mechanics > comp neutrality > tier rebuild`
- effort: `comp neutrality > tier rebuild > activation over recruitment > conflict and registration policy > funds mechanics > horizon reset`
- value: `comp neutrality > activation over recruitment > conflict and registration policy > tier rebuild > funds mechanics > horizon reset`
- compliance cost: `tier rebuild == comp neutrality > conflict and registration policy > all others` (tie argued: each reopens a signed instrument - partner agreements in one case, rep comp plans in the other - and neither rolls back without a second amendment)
Resetting the horizon leads because it costs one leadership conversation and stops the program being killed before any other fix can land. Comp neutrality is the highest-value fix and still not first: it is a mid-year comp change, slow and politically expensive. Re-rank against the user - a program with neutral comp already deletes that row, and a single-tier program deletes the tier rebuild.
| Symptom | Structural cause to check first |
| ---------------------------------- | --------------------------------------------------------------------------------------------- |
| Many signed partners, few active | Recruitment tracked instead of activation; no partner value proposition ever validated |
| Partners ignore tier progression | Cosmetic tiers - nothing measurable changes at the boundary |
| Program funds unclaimed | Market-development funds designed apart from deal registration, claims friction |
| Partners stopped registering deals | Vendor competes on registered deals (house accounts); conflict policy written after the fact |
| Direct reps block partner deals | Compensation penalizes partner involvement - an architecture decision, not a behavior problem |
| Program "failing" at month 6 | Horizon error: judged on direct-sales cycles; reset expectations to the year-1/2/3 ladder |
| Overloaded partner managers | Portfolio far beyond the sustainable range; program scaled recruitment before activation |
Full failure catalog with fixes and a worked negative example: [references/readiness-and-failure-modes.md](references/readiness-and-failure-modes.md).
## B2B and B2C
Identical in both:
- Deep-qualify a small first cohort before scaling recruitment.
- Model the economics from the partner's side before setting any number.
- A compelling economic offer, not marketing, is what recruits.
Genuinely different:
- Franchising's legal instruments and exclusive territories.
- Day-one documentation depth: exhaustive in franchising, deliberately minimal in B2B SaaS.
- Consumer programs' point-of-sale attribution versus B2B's multi-month pipeline attribution.
- Front-line mindshare/findability requirements unique to consumer-facing dealers.
Details and reliability flags: [references/b2b-b2c-contrast.md](references/b2b-b2c-contrast.md).
## Invocation Examples
- "We have inbound partner requests and no formal program. Are we even ready for partners, and if so what should the program look like?"
- "Design a referral program for a $12K ACV product - we tried resellers and the margin math never worked."
- "Our partner program has requirements nobody meets and benefits nobody wants. Rebuild the structure from the readiness gate up."
## Reference
- See [references/decision-sequence-and-frameworks.md](references/decision-sequence-and-frameworks.md) for the full decision sequence, ACV-banded motion selection, and named frameworks with honest status flags.
- See [references/readiness-and-failure-modes.md](references/readiness-and-failure-modes.md) for the complete readiness gate and the failure-mode catalog.
- See [references/collaboration-formats.md](references/collaboration-formats.md) for how channel professionals actually workshop a program design.
- See [references/economics-and-benchmarks.md](references/economics-and-benchmarks.md) for the economics-envelope method and benchmark bands with reliability flags.
- See [references/b2b-b2c-contrast.md](references/b2b-b2c-contrast.md) for the dealer/franchise/ambassador contrast.
- See `mbfinotti/partnerships-skills@co-selling-strategy` for deal registration rules, credit splits, and engagement rules.
- See `mbfinotti/partnerships-skills@partner-channel-conflict` for direct-versus-partner conflict resolution.