influencer-outreach

{
  "skill_name": "influencer-outreach",
  "evals": [
    {
      "id": 1,
      "prompt": "Havli here - enamel cast-iron cookware, average order around $68. We want @noorcooksslow (92k on Instagram, slow-cooking recipes). Offer is 15% commission, 45-day cookie, paid on the 15th of the month after the sale clears. Her bio says \"mgmt: [email protected]\". I'd rather just DM her directly - going through an agent adds three weeks and I've had agents kill deals on me before. Heads up, our warehouse is mid-move so we can't ship samples for at least a month, and it's just me here so I don't have time to run intro calls. Write it.",
      "expected_output": "A first-touch pitch addressed to the listed manager, with the commission mechanics front-loaded, a single CTA that the brand can actually honor, and a follow-up plan capped at three total messages.",
      "files": [],
      "expectations": [
        "Recommends pitching the listed manager or agent rather than sending a DM to the creator",
        "States that the contact the creator publishes for business decides the channel, and that the ranking only breaks ties between published routes",
        "Warns against opening a parallel direct thread to the creator while the manager thread is live",
        "Does not present the DM as a lower-ranked but still available option once a manager is published",
        "The drafted message states the 15% commission rate",
        "The drafted message states the 45-day attribution window",
        "The drafted message states the monthly payout timing",
        "The drafted message includes an approximate per-sale earning derived from the $68 average order value (roughly $10)",
        "The drafted message contains no rate card and no multi-deliverable campaign plan",
        "The drafted message contains exactly one ask",
        "Does not offer a product sample as the CTA, because no stock can ship",
        "Does not offer a call as the CTA, because nobody is available to run one",
        "The follow-up plan stops at three messages in total, counting the first touch"
      ]
    },
    {
      "id": 2,
      "prompt": "I run partnerships at Rivenwork, a B2B compliance SaaS. I'm putting about 40 individual creator pitches into next quarter's plan and my VP won't sign off without a reply-rate target in the doc. What's the industry benchmark for creator outreach, and what number should I commit to? Also tell me what I should be tracking week to week so I can show progress.",
      "expected_output": "A refusal to supply a published reply-rate benchmark, an explanation that none is trustworthy, and an own-log measurement plan with named columns and core KPIs.",
      "files": [],
      "expectations": [
        "Declines to supply an industry benchmark reply rate for creator outreach",
        "States plainly that no trustworthy published reply-rate benchmark exists for creator outreach",
        "Does not quote an \"8.5% reply rate\" figure",
        "Does not quote a \"99% ignored\" claim about generic salutations",
        "Does not quote a \"90% response rate\" practitioner claim",
        "Does not present a \"two personalized details raise replies by 56%\" (5.6% vs 3.6%) figure as fact",
        "Recommends building the team's own pitch log as the only valid benchmark",
        "Names the log fields to capture: creator, channel, angle, sent date, follow-ups used, outcome, and days to reply",
        "Says the log becomes a usable benchmark after roughly a dozen pitches",
        "Lists reply rate, positive-reply rate, and days to first reply among the core KPIs",
        "Includes which angles produced pitches that survived the downstream negotiation stage, recorded after that stage closes"
      ]
    },
    {
      "id": 3,
      "prompt": "Mossline, DTC clean skincare. I'm pitching Ines Vardar (@inesvardar, 71k on Instagram, ingredient-breakdown videos) - offer is a $1,800 flat fee for one reel plus two stories. Walk me through the standard framework everyone uses for creator pitches, then write the email. Legal wants the FTC disclosure requirement and our 90-day category exclusivity spelled out in that first email so there are no surprises down the line.",
      "expected_output": "A pitch that declines to invent a standard framework and keeps disclosure, exclusivity, and contract terms out of the cold message, routing the disclosure obligation elsewhere.",
      "files": [],
      "expectations": [
        "Does not present any framework as the industry standard for creator pitching",
        "Does not invent a branded acronym or numbered \"standard\" method for creator outreach",
        "States that no named industry-standard framework exists for creator pitching",
        "Keeps FTC or disclosure language out of the first-touch message",
        "Keeps the 90-day exclusivity demand out of the first-touch message",
        "Keeps contract terms out of the first-touch message",
        "Explains that disclosure, contract, and exclusivity topics belong after the creator agrees to talk, not in a cold pitch",
        "Routes the substantive disclosure obligation to a separate disclosure-compliance skill rather than drafting label copy here",
        "Does not claim the disclosure obligation is triggered by the contract or agreement date",
        "The drafted message names the flat-fee model and the $1,800 figure without pasting a rate card",
        "The drafted message uses no generic salutation such as \"Dear influencer\"",
        "The drafted message is signed by a named human rather than a team alias"
      ]
    },
    {
      "id": 4,
      "prompt": "Nullstack, observability SaaS. I want to sponsor Cold Start, the podcast Tobias Renn hosts for platform engineers - roughly 11k listeners an episode. Budget is confirmed at $4,000 for one mid-roll read plus a show-notes link, and he lists [email protected] on the site. Our 2.0 launch is in 12 days and the slot has to be locked before then. Quick sanity check: is $4k right for 11k listeners? Then write the email and tell me how to chase it.",
      "expected_output": "A deadline-adjusted pitch: near-zero-effort angle, reply-for-details CTA, single follow-up cadence, the flat fee named without a rate card, and no CPM figure presented as a market rate.",
      "files": [],
      "expectations": [
        "Re-ranks the default option order against the 12-day deadline and names which answer moved which option",
        "Promotes a near-zero-effort personalization angle over the hour-of-grounding \"specific content\" angle because of the hard date",
        "Recommends the reply-for-details CTA",
        "Recommends a single follow-up at roughly five to seven business days instead of the two-follow-up default",
        "Notes the cost of the single-follow-up cadence: a host who was simply travelling gets missed",
        "The drafted message names the $4,000 flat fee",
        "The drafted message contains no rate card and no attached campaign proposal",
        "The deliverable stays a one-line sketch in the first message",
        "The drafted message speaks to a business need the host's audience has, not only a product feature",
        "Does not present podcast CPM ranges as market rates for an 11k-listener show",
        "Declines to redesign or adjust the confirmed $4,000 offer, noting this work communicates the offer rather than setting it",
        "The subject line names something specific to the show rather than a generic \"sponsorship opportunity with Nullstack\" pattern",
        "The subject line contains no emoji"
      ]
    },
    {
      "id": 5,
      "prompt": "Palta, small-batch hot sauce, we're four people. We want to send a 6-bottle box to @tinyfirekitchen (24k on TikTok, one-pan chili videos). No money in this at all - the box is the whole offer. Her bio says \"DMs open, no email pls\". We do have an application page at palta.co/creators I'd love her to fill out. Write the DM.",
      "expected_output": "A 2-4 sentence DM with no links, stating plainly that the offer is product rather than payment, with one low-friction ask and no mass-funnel application link.",
      "files": [],
      "expectations": [
        "The drafted DM is two to four sentences long",
        "The drafted DM contains no links",
        "Does not paste the palta.co/creators application URL into the message",
        "States plainly that the offer is product, not payment",
        "Does not imply a fee, commission, or future payment",
        "Does not offer exposure or audience growth as the compensation",
        "Does not open on an asserted audience overlap as the angle",
        "Opens on one verifiable creator-specific fact framed as audience fit rather than flattery",
        "The drafted DM contains exactly one ask",
        "Deletes the application-link CTA outright rather than ranking it last, because it routes an individually vetted creator into a mass funnel",
        "Does not fabricate familiarity such as \"I've been following you for ages\"",
        "Confirms the DM is the right channel here because the creator publishes no business email and no manager"
      ]
    },
    {
      "id": 6,
      "prompt": "I emailed Rowan (@rowanbuildsdesks, 60k on YouTube, hand-tool woodworking) nine days ago about our chisel set - 10% commission, 30-day cookie, net-30 payout. Nothing back, not even an open as far as I can tell. Write me the follow-ups so I can queue them up, and tell me when to send each one.",
      "expected_output": "One or two short follow-ups, each one or two sentences carrying a new angle, on the same channel, capped at three total messages, with an explicit close-the-file rule.",
      "files": [],
      "expectations": [
        "Each drafted follow-up is one or two sentences",
        "No follow-up restates the commission rate, cookie window, or payout terms from the first touch",
        "No follow-up restates the content of the first touch generally",
        "Each follow-up carries one new small angle: a held-back proof point, an easier version of the ask, or a genuine timing hook",
        "No follow-up uses \"just checking in\", \"bumping this\", or \"did you see my email?\"",
        "No follow-up invents a deadline or manufactures scarcity",
        "No follow-up uses guilt-tripping or chasing language",
        "The sequence stops at three messages in total, counting the original email",
        "The second follow-up lands roughly two weeks after the first follow-up",
        "The follow-ups stay on the same channel as the first touch",
        "Scores the follow-ups against the follow-up quality items only, not the full first-touch checklist",
        "States that silence after the final message closes the file, reopened only with something genuinely new and never as a re-send",
        "Runs a humanizer pass over the follow-up copy before presenting it"
      ]
    },
    {
      "id": 7,
      "prompt": "Verdiga, contract analytics for legal ops teams. I pitched Amara Oyelaran's legal-ops newsletter at $2,500 flat for one dedicated issue. She just wrote back: \"Interested, though $2,500 is low for a dedicated send - I'm usually at $4,000 - and I'd want to keep usage rights.\" Draft my reply countering at $3,000 with a six-month usage grant, and tell me what exclusivity window I should push for.",
      "expected_output": "A hard stop at the positive reply: no counter drafted, remaining follow-ups cancelled, and a handoff package carrying the stated compensation facts unchanged into the negotiation stage.",
      "files": [],
      "expectations": [
        "Stops the outreach sequence at the positive reply",
        "Does not draft the counter-offer or the negotiation reply",
        "Hands off to a separate creator-negotiation skill rather than negotiating inside the pitch thread",
        "Names that handoff target explicitly as the place where rates and deal terms are worked",
        "Carries the $2,500 figure into the handoff unchanged",
        "Does not invent, round, or adjust any of the stated compensation facts",
        "Does not propose usage-rights terms",
        "Does not propose an exclusivity window",
        "Cancels the remaining planned follow-up touches",
        "Produces handoff notes covering the compensation as stated, the deliverable mentioned, and the CTA promised",
        "Logs the outcome and the days-to-reply against this pitch"
      ]
    },
    {
      "id": 8,
      "prompt": "Ostara Labs, electrolyte tablets. Creator is @dunelanepace - 38k on Instagram, ultra-trail runner. Three weeks ago she posted a training-week recap and listed our tablets in her kit, unprompted, nobody paid her a cent. She also launched a paid training plan ten days ago, and honestly half our team uses her free race-pacing spreadsheet. She lists [email protected]. Offer is 18% commission, 30-day window, paid monthly. We've got pallets of product, a fulfilment partner who ships same day, and my calendar is wide open. Just write the email, you know the drill.",
      "expected_output": "An angle brainstorm run before any drafting, landing on the she-already-uses-us angle, with the sample-on-reply CTA and its disclosure consequence flagged, delivered section by section.",
      "files": [],
      "expectations": [
        "Writes no pitch message until the user has explicitly picked an angle",
        "Proposes two to three candidate angles with trade-offs and a recommendation",
        "Recommends the angle that she already uses the brand, citing the unprompted kit-list mention",
        "Ranks that angle above both the recent-launch angle and the specific-content angle",
        "Does not offer an asserted audience overlap as a candidate angle",
        "Asks brainstorming questions one at a time rather than in a single block",
        "Recommends the sample-on-reply CTA over reply-for-details, because the product's value is obvious on contact and fulfilment ships without a second approval",
        "Flags that a shipped sample creates a material connection the moment it lands, attaching a disclosure obligation to anything she posts about it whether or not a deal follows",
        "Notes that a shipped sample cannot be un-shipped, unlike the other CTAs",
        "Applies the swap test: substituting another creator's name would force a rewrite of the message",
        "Presents the package section by section for validation rather than as one finished block",
        "Delivers the artifact with labelled sections for the chosen angle, first touch, follow-up plan, stop rule, and handoff notes"
      ]
    }
  ],
  "trigger_queries": [
    { "query": "write a first-touch email to an influencer we want to work with", "should_trigger": true },
    { "query": "draft a collab pitch for a creator on instagram", "should_trigger": true },
    { "query": "how should I word a DM to an influencer about a paid partnership", "should_trigger": true },
    { "query": "we vetted a creator, now what do I actually send her", "should_trigger": true },
    { "query": "pitch message for a newsletter writer we want to sponsor", "should_trigger": true },
    { "query": "I need outreach copy for a youtuber, 12% commission offer", "should_trigger": true },
    { "query": "her bio only lists a talent agency - who do I email, and what do I say", "should_trigger": true },
    { "query": "how many follow-ups before I give up on a creator who hasn't replied", "should_trigger": true },
    { "query": "she didn't reply to my collab email, what do I send next", "should_trigger": true },
    { "query": "what subject line gets a creator to open a partnership email", "should_trigger": true },
    { "query": "proposal message for a podcast host we want to sponsor", "should_trigger": true },
    { "query": "we're offering free product only - how do I say that without sounding cheap", "should_trigger": true },
    { "query": "help me write to @slowloomtextiles about our spring drop", "should_trigger": true },
    { "query": "first message to a tiktoker for a brand deal", "should_trigger": true },
    { "query": "I found the perfect person to talk about our launch, what do I write to her", "should_trigger": true },
    { "query": "creator outreach message plus a plan for when to nudge", "should_trigger": true },
    { "query": "how do I personalize this pitch so it doesn't read like a mail merge", "should_trigger": true },
    { "query": "b2b creator program - drafting the first email to an industry expert", "should_trigger": true },
    { "query": "should I email or DM this influencer", "should_trigger": true },
    { "query": "write the pitch and tell me when to chase it", "should_trigger": true },
    { "query": "our affiliate offer is 20% recurring - message for one specific youtuber", "should_trigger": true },
    { "query": "what do I put in the opening line so it doesn't come across as flattery", "should_trigger": true },
    { "query": "sponsorship pitch to an SRE newsletter author, $3k flat", "should_trigger": true },
    { "query": "cold message to a creator who has never heard of us", "should_trigger": true },
    { "query": "she replied and said yes - do I keep sending the sequence", "should_trigger": true },
    { "query": "draft outreach to a creator's manager", "should_trigger": true },
    { "query": "how do I pitch one influencer without it sounding like a template", "should_trigger": true },
    { "query": "we want this reviewer to try our product - what's the first message", "should_trigger": true },
    { "query": "write a short instagram DM proposing a partnership", "should_trigger": true },
    { "query": "our brand voice is peer-casual, rewrite this creator pitch", "should_trigger": true },
    { "query": "pitch email for a gifted collab", "should_trigger": true },
    { "query": "timing for follow ups after a creator pitch", "should_trigger": true },
    { "query": "what should the ask be in a first message to an influencer", "should_trigger": true },
    { "query": "help me reach out to a twitch streamer about sponsoring a stream", "should_trigger": true },
    { "query": "she's on our shortlist and lists a partnerships@ address - draft it", "should_trigger": true },
    { "query": "message a substack writer about a paid sponsorship", "should_trigger": true },
    { "query": "collab proposal wording", "should_trigger": true },
    { "query": "I want to offer 10% commission to a creator - how much detail goes in email one", "should_trigger": true },
    { "query": "first contact with a creator who already posted about us organically", "should_trigger": true },
    { "query": "we have 10 days before launch - fastest way to get a creator to reply", "should_trigger": true },
    { "query": "outreach copy for one vetted creator, not a whole list", "should_trigger": true },
    { "query": "how do I not sound desperate when I nudge an influencer again", "should_trigger": true },
    { "query": "should I include our rate card in the first email to a creator", "should_trigger": true },
    { "query": "what goes in the pitch when the creator already uses our product", "should_trigger": true },
    { "query": "write to this creator and plan the nudges", "should_trigger": true },
    { "query": "which of these 20 creators should we approach first", "should_trigger": false },
    { "query": "score creators on audience fit and brand safety", "should_trigger": false },
    { "query": "build a shortlist of fitness creators for our campaign", "should_trigger": false },
    { "query": "how do I check whether an influencer's followers are fake", "should_trigger": false },
    { "query": "she countered at $6,000 - what's my move", "should_trigger": false },
    { "query": "usage rights and exclusivity terms for a creator deal", "should_trigger": false },
    { "query": "what's my walk-away number with this creator", "should_trigger": false },
    { "query": "draft the creative brief for the creator we just signed", "should_trigger": false },
    { "query": "deliverable specs and key messages for a signed creator", "should_trigger": false },
    { "query": "how many revision rounds should the creator get", "should_trigger": false },
    { "query": "KPIs for our creator campaign", "should_trigger": false },
    { "query": "how do we attribute sales back to an influencer post", "should_trigger": false },
    { "query": "recruit 300 affiliates into our program", "should_trigger": false },
    { "query": "outreach sequence for coupon and cashback sites", "should_trigger": false },
    { "query": "email template to pitch review sites to join our affiliate program", "should_trigger": false },
    { "query": "welcome emails for newly approved affiliates", "should_trigger": false },
    { "query": "what commission rate should we pay affiliates", "should_trigger": false },
    { "query": "does this caption meet FTC disclosure rules", "should_trigger": false },
    { "query": "where should the #ad label go in a reel", "should_trigger": false },
    { "query": "draft our affiliate program terms and conditions", "should_trigger": false },
    { "query": "our affiliate's traffic looks fake, what do I check", "should_trigger": false },
    { "query": "audit this month's commission payout before we disburse", "should_trigger": false },
    { "query": "what metrics go on our affiliate dashboard", "should_trigger": false },
    { "query": "how big should our refer-a-friend reward be", "should_trigger": false },
    { "query": "customers are gaming our referral codes", "should_trigger": false },
    { "query": "rank these five alliance candidates for us", "should_trigger": false },
    { "query": "design our deal registration policy", "should_trigger": false },
    { "query": "our reseller and our AE are fighting over the same account", "should_trigger": false },
    { "query": "how many partner tiers should we have", "should_trigger": false },
    { "query": "is this reseller worth the margin we give them", "should_trigger": false },
    { "query": "which partner category should we launch next", "should_trigger": false },
    { "query": "plan a joint webinar with our integration partner", "should_trigger": false },
    { "query": "certification path for our channel partners", "should_trigger": false },
    { "query": "should we list on the AWS marketplace", "should_trigger": false },
    { "query": "write a job posting for a partner manager", "should_trigger": false },
    { "query": "how do I break into partnerships as a career", "should_trigger": false },
    { "query": "which partnerships skill do I need for this project", "should_trigger": false },
    { "query": "cold email sequence for outbound sales prospects", "should_trigger": false },
    { "query": "pitch myself as a guest on a tech podcast", "should_trigger": false },
    { "query": "submit a talk to a conference CFP", "should_trigger": false },
    { "query": "reach out to speakers for our conference lineup", "should_trigger": false },
    { "query": "sponsorship prospectus for our developer conference", "should_trigger": false },
    { "query": "linkedin connection request message for a prospect", "should_trigger": false },
    { "query": "PR pitch to a journalist about our funding round", "should_trigger": false },
    { "query": "influencer marketing strategy and budget for next year", "should_trigger": false }
  ]
}

# Channel choice and manager etiquette

## Which channel carries the pitch

Pitch the contact the creator publishes for business: check the bio link, media kit, and about page first, before applying any ranking below.

- A DM blast to a creator who lists a partnerships email signals the sender never looked.
- A cold email scraped for a DM-only creator signals the same.

The ranking below decides only the residual cases: the creator publishes more than one route, or publishes none.

Default order, highest value per unit of effort first (re-rank against the Interview answers):

1. **The listed manager or agent.** Effort: an hour, the same grounding and personalization a direct pitch needs, plus relationship capital you keep spending afterwards. Value: a decision from someone who evaluates offers for a living and whose yes carries the creator - and a relationship that pays again across their whole roster.
2. **The listed business email.** Effort: an hour, grounding plus the full message structure. Value: the channel the creator opened for business, where a complete offer fits and the thread stays followable.
3. **A DM.** Effort: near-zero. Value: lowest - it lands beside fan mail, and the 2-4 sentence no-links limit means a commission offer's mechanics barely fit.
   - Counter-signal, single-practitioner: brand-side practitioner Agita Matule (marketing team at a delivery platform) still prefers DMs - "less formal, and a lot of influencers don't list their email." This marks DM-first as a legitimate style, not a mistake.

Where the axes disagree:

- efficiency: manager > business email > DM
- value: manager > business email > DM
- effort, highest first: manager == business email > DM - a genuine tie: both need the same hour of grounding before a word is written, and the manager's extra cost is relationship capital rather than drafting time.
- compliance cost, highest first: DM > business email > manager
  - DM: violates a platform's unsolicited-commercial-message rule, an account-level action appealable only through the platform.
  - Cold email: triggers the unsolicited-contact regime (identify the sender, honor an opt-out, hold a lawful basis for the address in some jurisdictions), reversible by suppressing it.
  - Manager: a published listing is an invitation to pitch, so it triggers nothing.

Delete, never demote, the option ruled out: drop the DM row entirely once the creator publishes an email or a manager - a DM parked at the bottom reappears as a second parallel touch, which is the end-run failure below. Drop the manager row only when the creator has personally asked for direct contact.

What this order starves: the small creator who publishes neither an email nor a manager. The efficiency order pushes every hour toward the two top rows, so those creators get skipped, and they are often the ones whose reply rate is highest and whose rates are lowest.

Promote the DM whenever no business contact is published, or the creator's own bio invites DMs. Vendor survey direction supports the default: a "47.7% email-first vs. 6.8% DM-first" split, vendor-published with undisclosed methodology - use the direction, not the numbers.

**B2B**: channel choice at the pitch stage has no B2B-specific convention to inherit, so the same gate and the same order apply.

## When a manager or agent is listed

The ranking above already puts the manager first; this is the etiquette that makes that row work.

- **Treat the manager relationship as a relationship, not a gate to slip past.** Address the manager, name the creator, state the fit evidence and the offer plainly; managers filter on economics fast.
- **Never go around a listed manager unprompted.** An end-run reads as disrespect to both manager and creator and can end the conversation before it starts.
- **A counter-signal exists, and it is now corroborated**: consultant Jason Falls reports (2026) that about half a dozen creators asked him to contact them directly rather than through their talent manager, and that an informal poll of his peers found the same pattern - his own account is still single-caseload and an undisclosed-size peer poll. A named survey now backs the same direction: HypeAuditor's Influencer Outreach Survey found 67% of influencers who work with agents or talent agencies still prefer to be contacted directly by brands, and only 15% of influencers use an agent at all.
  - Read it as: if a creator _tells you_ to contact them directly, honor that preference without hesitation.
  - Never read it as license to bypass a manager on your own initiative.
- If the manager goes silent, follow up with the manager per the standard cadence. Do not open a parallel direct thread to the creator while the manager thread is live.

## Recording the choice

Log the channel used and whether a manager was in the loop alongside each pitch in the outreach log - channel is one of the variables the manager's own reply-rate history will eventually rank.

# Examples

The positive examples are original compositions following the documented first-touch structure, validated against a published commission-pitch example from a creator-platform vendor. The negative example is an original composition assembled from the sourced negative signals (bad subject lines, documented mistakes) and annotated line by line. In a real engagement every fact comes from the interview; never invent numbers, posts, or familiarity.

## Positive - B2C, commission offer, direct email

Scenario (illustrative):

- Brand: a home-coffee-gear brand pitching a vetted creator whose channel reviews budget espresso setups.
- Offer: already decided at 12% commission, 30-day attribution, monthly payout.
- Contact: creator lists a business email.
- CTA the brand can honor: sample on reply.

> Subject: your $300 espresso setup video - and our grinder
>
> Hi Marisol,
>
> Your "$300 espresso setup" video is the most honest budget build I've seen - and people setting up their first machine at home are exactly who buys our hand grinder.
>
> I run partnerships at [brand]. We just opened our creator program: 12% commission on every sale through your link, tracked for 30 days, paid monthly. On our average order that's about $11 per sale.
>
> If you're curious, just reply and I'll send a grinder over so you can try it before deciding anything.
>
> Dana Okafor, Partnerships, [brand]

Why it passes the gate:

- Subject names her actual work.
- Opener is a verifiable specific framed as audience fit, not flattery.
- Commission mechanics front-loaded exactly as the model requires: rate, window, payout, per-sale estimate. Vague "great commission" language is the documented failure.
- One low-friction sample-on-reply CTA.
- No rate card, no deliverable list, no contract talk.
- Fails the swap test in the right direction: no other creator fits this message.

## Positive - B2B, flat fee, via listed contact

Scenario (illustrative):

- Brand: an incident-management SaaS pitching a vetted SRE-newsletter writer (18k subscribers, mostly on-call engineers).
- Offer: flat-fee sponsored deep-dive, budget confirmed at $2,500.
- Contact: she lists a sponsorship email.
- CTA: reply for the details doc.

> Subject: sponsored deep-dive for On-Call Weekly
>
> Hi Priya,
>
> Your postmortem-culture series reads like it was written for our users - engineers who own the pager and hate their current alerting stack. That overlap is why I'm writing.
>
> We'd like to sponsor one deep-dive issue: your voice, your angle, our tool as the worked example. Paid flat fee - we've budgeted $2,500 for this - and we know a sponsored issue only works if it's genuinely useful to your readers.
>
> Want the one-page details? Just reply and I'll send it over.
>
> Sam Reyes, Partnerships, [product]

B2B notes: the pitch speaks to a business need of her audience (on-call pain), not just a product feature - the one consistently documented B2B-specific rule.

- The fee is named (budget signal, no rate-card dump).
- The deliverable stays a one-line sketch with the detail doc gated behind the reply.

## Follow-up plan (fits either example)

| Touch | Day | Channel                | Angle           | Summary                                                                                                                                            |
| ----- | --- | ---------------------- | --------------- | -------------------------------------------------------------------------------------------------------------------------------------------------- |
| 1     | 0   | Email (listed contact) | Fit + offer     | First touch above                                                                                                                                  |
| 2     | 7   | Same thread            | Held-back proof | Two sentences: one conversion or audience proof point not in touch 1, re-offer the CTA                                                             |
| 3     | 21  | Same thread            | Easier ask      | Two sentences: shrink the ask (sample only / details doc only), then close: "if the timing's off, no pressure - I'll leave it here." Then honor it |

Positive reply at any touch: stop, hand off to negotiation with the stated numbers unchanged. Silence after touch 3: close the file; reopen only with something genuinely new.

## Negative - the pitch that gets deleted

Original composite of the sourced negative signals, annotated line by line.

> **Subject: Collab opportunity with GlowCore!! 👀**

Both documented bad-subject patterns at once - the generic "collab opportunity" template and the emoji-bait "quick question" energy. Brand-first, creator-absent.

> **Dear influencer,**

The documented generic salutation. Two words in, the mail-merge is visible.

> **I hope this message finds you well! I've been a huge fan of your amazing content for a long time.**

Filler opener plus fabricated familiarity - no post named, because none was watched. Unverifiable flattery where the audience-fit reason should be.

> **GlowCore is a revolutionary clean-beauty platform with 4 product lines, AI-matched shades, dermatologist partnerships, and over 40,000 five-star reviews.**

Feature dump. Nothing answers "why my audience?" - the personalization principle every source agrees on is missing at the exact spot it belongs.

> **We offer very competitive compensation. Our full rate card and 12-week campaign plan with 9 deliverables across 3 platforms are attached, and rates are non-negotiable.**

Three documented leave-outs stacked:

- Vague compensation where a real number belongs.
- The rate-card/campaign-proposal dump.
- Aggressive negotiation language ("non-negotiable") in message one.

> **Please also note all content will require our standard exclusivity terms and FTC disclosure language, per the attached contract.**

Contract, exclusivity, and disclosure terms in a cold pitch - every documented workflow places these after agreement. Here they read as handcuffs before hello.

> **Sign the contract here, follow us on Instagram, and book an onboarding call this week!**

Three asks, none low-friction. The one-ask rule is unanimous across outreach guidance.

> **Best regards,
> The GlowCore Growth Team**

No named human. A team alias forecloses the reply-to-a-person CTA that the entire pitch should funnel into.

Net effect: every documented negative signal in under 130 words - swap-test-proof that it was written for no one. The two positive examples fix each line structurally.

# Follow-up cadence

Cadence is the best-sourced area of creator outreach: one vendor survey's named practitioners and an unrelated commerce-platform guide independently converge on the same shape.

## Three cadences, ranked

Best value per unit of effort first. This is a default; the Interview's deadline and one-off-versus-compounding answers re-rank it.

1. **Two follow-ups, hard stop at three messages.** First follow-up about one week after the pitch, second about two weeks after that. Effort: an hour to plan, then two short messages. Value: the full realistic yield from a busy creator, without crossing into aggravation.
   - Dmitri Cherner (former head of influencer marketing at two consumer brands, in a creator-platform vendor's outreach survey of 51 marketers, self-reported, 2026): "The first a week later, and the second two weeks after that. Anything beyond three messages just becomes aggravating" and "I plan for two follow-ups, max."
   - Sarah Saffari (agency founder, same survey): "We space them out and stop at 3."
2. **One follow-up, five to seven business days out, then move on.** Effort near-zero beyond the first message. Value: most of the reply yield lands in that window anyway, but a creator who was simply travelling gets missed. A different vendor's guide (Michael Keenan, 2024) corroborates the shape in direction, not identically - it implies a shorter total window than row 1.
3. **The long tail for a must-have creator.** After the standard follow-ups, one light touch at about a month, then three months, then six. Effort: a standing job where someone holds the file across quarters and owes a genuinely new reason to send each touch. Value: the only cadence that can land a creator whose no was really a "not this quarter" - single-vendor guidance, uncorroborated.

Where the axes disagree:

- efficiency: two follow-ups > one follow-up > long tail
- value: long tail > two follow-ups > one follow-up
- effort, highest first: long tail > two follow-ups > one follow-up

What this order starves: the long tail - highest value on the rare creator the brand genuinely must have, highest effort, so it loses every round on ratio. Promote it when the manager names that creator as a priority and signs off on holding the file open. Delete it rather than demote it when nobody owns the log across quarters: parked at the bottom, it becomes an unowned reminder that fires six months later with nothing new to say.

## Follow-up copy rules

- Short: one or two sentences. Nothing from the first touch restated.
- Each follow-up carries a new small angle - a proof point held back from message one, an easier version of the ask, a genuine timing hook. Never "just checking in", "bumping this", or "did you see my email?".
- Low pressure by design: silence is busy, not rejection - the consistent practitioner framing across sources. A guilt-tripping follow-up converts a busy creator into a lost one.
- Same channel as the first touch by default; switch channels only if the first channel was a guess and a better-listed one surfaced since.

## No reply-rate benchmark exists

- The same vendor survey explicitly declines to give one ("I don't have a good answer for you here"), citing too many confounds.
- Circulating figures are unusable, never quote any of these as fact or set one as a target:
  - an "8.5% reply rate" (citation resolves to the wrong domain)
  - a "99% ignored" claim for generic salutations (citation resolves to a homepage, not a study)
  - one practitioner's self-reported "90% response rate" (single self-report, not a benchmark)
- The only valid benchmark is the manager's own pitch log - see the KPIs section of the skill.

## Stop rules

- Positive reply at any point: stop the sequence immediately and hand off to negotiation with the compensation facts unchanged.
- Silence after the final message: close the file. Reopen only with something genuinely new - a product launch relevant to their audience, a changed offer - never a re-send.
- A "no" is a no: log it with the reason if given, and do not re-pitch the same offer.

## B2B and B2C

Every figure above is published unsegmented by business model. Apply the same plan to both - and know that "it applies to both" is an inference from that unsegmented guidance, not a documented comparison.

# Pitch anatomy - the first touch, element by element

## Subject line (email only)

- Specific and personalized beats generic - vendor-published guidance, consistent with the corroborated personalization principle below.
- Three working patterns, best value per unit of effort first: **name a specific piece of their content** ("Loved your [specific post] - thought of you for our partner program") > **name the concrete offer** > **name the audience benefit**.
  - The content subject leads on both axes at once: once the opener below is grounded, it costs nothing extra to write, and it is the only one of the three that another creator could not receive unchanged.
  - The offer subject is near-zero effort and earns an open from anyone already selling their audience.
  - The benefit subject is the one every brand sends.
- Documented bad subjects: "Collab opportunity with [Brand]!" and "Quick question 👀" - flagged by a creator-platform vendor's outreach guide (Lizzie Davey, 2026).
- Keep it short enough to survive mobile truncation; no emoji bait.

## Opener

- One verifiable, creator-specific fact, framed as audience fit rather than flattery.
  - Vendor-published formulation: "lead with audience fit rather than flattery."
  - Practitioner consensus (Kalin Anastasov, trade-publication writer): "generic outreach is one of the biggest reasons creators get ignored."
- The fit line states the reasoning, not a compliment: their audience is X, and X is exactly who buys or uses this product.
- A vendor claims two personalized details raise reply rate 56% (5.6% vs. 3.6%) - vendor-published with undisclosed methodology; use the principle, never the number.

### Angle menu, ranked

The documented personalization signals, ordered by value per unit of effort - this is the candidate pool the brainstorming step draws from, and this order is its default, not its verdict.

1. **They already use or recommend the brand.** Effort near-zero when true: you already hold the receipt. Value highest of all: their own behavior has proved the fit, so the pitch is only "here's the paid version of that" - a documented second pitch type, and unbeatable when verifiable.
2. **Something they launched in the last few weeks.** Effort near-zero - it is pinned to their feed. Value good but perishable: it reads as attention today and as staleness in a month.
3. **A specific piece of their content.** Effort an hour: watch or read two or three recent pieces before writing. Value: the strongest cold angle, and it passes the swap test by construction.
4. **A product of theirs the brand team actually uses.** Effort near-zero when true, and rarely true. Value strong - the fit runs in both directions and both sides can verify it.
5. **A natural fit with their existing content format.** Effort an hour, since the format must be known cold. Value moderate: it answers "what would this even look like", which matters most when the deliverable is unusual.
6. **An asserted audience overlap.** Effort near-zero. Value lowest as an opener, because every brand claims it and the claim survives no scrutiny.

Where the axes disagree:

- efficiency: already-recommending > recent launch > specific content > their-product-we-use > format fit > asserted overlap
- value: already-recommending > specific content > their-product-we-use > format fit > recent launch > asserted overlap
- effort, highest first: specific content == format fit > recent launch == their-product-we-use == already-recommending == asserted overlap
  - specific content and format fit tie because both demand the same thing before a word is written: sitting through their actual work.
  - the other four tie at near-zero because each is a fact you either already hold or simply assert.

Delete, never demote, an angle that is not true. A creator who does not already use the brand deletes row 1 outright; parked at the bottom, "they love us" comes back as the fabricated familiarity in the leave-out list. Same for row 2 with no live launch and row 4 with no product the team actually uses.

What this order starves: the audience-overlap argument backed by real data - the creator's media kit, analytics they shared, or your own customer list matched against their audience. Backed that way it is the second-most valuable angle and the most expensive one, so it loses every round on ratio and stays stuck at row 6 in its asserted form. Promote it when the data is already in hand, or when the offer is large enough that a data-backed fit argument is the thing that unlocks it.

## The offer - compensation transparency by model

The documented split is by compensation model, not by preference:

- **Commission / affiliate offer**: front-load the exact mechanics in message one - commission rate, attribution window, payout schedule, and an approximate per-sale earning. Documented warning against "great commission" vagueness: give a real number or an anonymized example payout. Rationale (Lizzie Davey, vendor writer):
  - a flat-fee collab puts the risk on the brand, which gets paid content either way
  - a commission offer flips the risk onto the creator, so vague mechanics read as risk-dumping
- **Flat fee**: name the model and signal that a real budget exists. Do not paste a full rate card or campaign proposal into message one, directional guidance since the source is written creator-to-brand: the restraint principle plausibly transfers, but the specifics are unverified for this direction. Know your number before reaching out either way - vendor-published, but uncontested anywhere.
- **Gifted-only**: say plainly that the offer is product, not payment, and never imply otherwise. No verified creator-side sentiment on gifted-only or "exposure" pitches is available; assume such pitches are unwelcome to professionalized creators and make gifted offers only where the discovery work supports it.

## Deliverable sketch, timing, and CTA

- Deliverables: one line, kept light. One commerce-platform guide (Michael Keenan, 2024) frames the first message as pre-answering four questions, a useful single-outlet checklist, not a standard:
  - who you want
  - what you want them to do and where
  - when you need a response
  - why the brand is a match

### CTA menu, ranked

Exactly one CTA, low-friction. Two documented phrasings exist for any of them, both presented in the sources as legitimate, with no data showing one converts better:

- the soft CTA ("would you be open to collaborating?")
- the specific ask (deliverables and terms up front)

Pick the phrasing by brand voice, and the ask itself by the order below.

1. **Reply for details.** Effort near-zero per pitch. Value: a reply you can act on, and nothing spent if it never comes. The default rung.
2. **Sample on reply.** Effort: a unit of stock, plus picking, packing and a fulfilment loop you run again for every pitch. Value highest - the product ends up in their hands, which is the strongest step toward a yes that a first touch can buy.
3. **A short call.** Effort: calendar coordination, a booked half hour, and a person to run it. Value: it surfaces the real objection faster than a thread will, which is what a multi-stakeholder B2B offer needs.
4. **An application link.** Effort near-zero, the asset already exists. Value lowest here: it routes an individually vetted creator into a mass funnel and contradicts the hour of personalization the rest of the pitch just spent.

Where the axes disagree:

- efficiency: reply for details > sample on reply > short call > application link
- value: sample on reply > short call > reply for details > application link
- effort, highest first: sample == call > reply == link
  - sample and call tie because each spends something unrecoverable per pitch: a unit of stock, or a half hour of someone's calendar.
  - reply and link tie at near-zero because both assets already exist and cost nothing more per pitch.
- compliance cost, highest first: sample > call == reply == link. A shipped sample creates a material connection the moment it lands, attaching a disclosure obligation to anything the creator posts about it whether or not a deal follows - and it cannot be un-shipped. The other three create no obligation until the creator agrees to something.

Default rung: reply for details. Move up one to sample on reply when the product's value is obvious on contact and fulfilment ships without a second approval. Move to the call when the offer is complex or the buying side is multi-stakeholder.

Delete, never demote, a CTA the brand cannot honor this week - the interview establishes which. A "short call" parked at the bottom with nobody to run it reappears as a promise broken in the first reply. Delete the application link for any individually vetted creator; it belongs to list-scale recruitment, which is a different skill.

## Leave out of message one

Sourced negative signals - each of these is documented as a first-message mistake:

- Full rate card or pricing sheet (directional).
- Complicated deliverable lists (directional).
- Aggressive negotiation language (directional).
- Long campaign proposals (directional).
- Generic salutations - "Dear influencer", "To whom it may concern" (single-practitioner, widely repeated).
- Fabricated familiarity - "I've been following you for a while" with nothing verifiable behind it.
- Disclosure language, contract terms, exclusivity demands: documented campaign workflows draft disclosure copy and contract terms only after the creator agrees to talk, never in a cold pitch - an operational sequencing point, not a claim about the legal deadline. The substantive obligation itself (16 CFR 255.5: any material connection disclosed clearly and conspicuously when the endorsement is published; UK rules require an obvious ad label) is real, and the FTC's own guidance ties the trigger to whether the connection exists and is unexpected at the time the endorsement runs, not to when it was contractually agreed - route the substance to the disclosure-compliance skill.

## Channel fitting

- **Email**: the full structure above - subject, opener, fit line, offer, sketch, one CTA. Short; every sentence earns its place.
- **DM**: compress to 2-4 sentences - fit fact, offer with the number, one CTA. No links in the first message.
- **Via manager**: same personalization effort as a direct pitch; address the manager, name the creator and the fit evidence, state the offer plainly - managers filter on economics fast.

## B2B adaptation

- Make the pitch speak to a business need the creator's audience has, not just a product benefit - the one consistently documented B2B-specific rule (trade-publication sourced).
- Documented B2B deliverables: webinars, white papers, articles, podcast sponsorship reads. Podcast CPM ranges circulate (pre-roll $10-20, mid-roll $20-40, post-roll $10-15) - single-outlet figures with no disclosed methodology; treat as a rough anchor for rate expectations, never as market rates.
- Gifting is documented B2B practice (ContentGrip's dedicated B2B gifting guide), just reshaped for the audience: platform access, beta invites, conference passes, and hosted events replace the physical-product seeding common in B2C. Zapier's creator-partner glamping retreat (~$136K in earned media, per the report) is a real named example, and TopRank Marketing's 2025 B2B Influencer Marketing Report ranks in-person events as the second most effective B2B creator content type (39%), after social posts. Free access still counts as material consideration for FTC purposes regardless of form. Default to fee or commission only where none of these gifting shapes fit the creator or the budget.
- Everything else on this page is published unsegmented by B2B/B2C; treating it as common to both is an inference, not a documented rule.

---
name: influencer-outreach
description: Draft the personalized first-touch pitch message and follow-up plan used to propose a partnership to one already-vetted influencer or creator - personalization angle, the compensation stated on offer, a message fitted to the channel the creator actually lists (email, DM, or their manager), and a bounded follow-up cadence. Covers B2C brand collabs and B2B creator programs (newsletter writers, podcasters, industry experts). Use whenever the user mentions pitching a creator, DMing an influencer, a collab proposal, or a creator outreach message, even if they never say outreach. Stops at the first positive reply. Do NOT use for list-scale affiliate recruitment - use mbfinotti/partnerships-skills@affiliate-recruitment-outreach instead.
license: MIT
metadata:
  author: Maya-Beth Finotti
  version: "1.0.10"
---

# Influencer Outreach

Draft the first-touch pitch and follow-up plan a partnership manager uses to propose a collaboration to one individually vetted influencer or creator, then stop at the first positive reply.

No named industry-standard framework exists for creator pitching. Published guidance is a scatter of single-outlet formulations, none branded or adopted cross-vendor:

- one vendor's soft-CTA vs. specific-ask split
- one commerce guide's four-question first-message checklist
- one consultant's unnamed six-step process

Never present any of them as a standard.

## Interview

Ask before drafting anything. One question per message; offer multiple-choice answers when possible. Skip any question the user already answered.

- Is this a B2C brand collab or a B2B creator program? (Product sold to consumers vs. to businesses - drives message framing and deliverable types.)
- What does the brand make, and what is the one-line reason this creator's audience would care?
- Who is the creator, on which platform(s), and what evidence says they fit? (Specific posts, audience data, a discovery brief - fit evidence is required input, not something to invent here.)
- What compensation model is on offer, with the actual number? (Flat fee $X / commission % + attribution window + payout schedule / gifted product only / hybrid.) This skill communicates the offer; it never designs or adjusts it.
- What campaign shape do you have in mind - rough deliverables and timing? (Stays a one-line sketch in message one.)
- What contact does the creator actually list: business email, DM only, or a manager/agent?
- What is the brand voice? (Formal, peer-casual, playful - or point to an example message.)
- Who signs, from which address or account?
- Which CTA can you actually honor: reply-for-details, product sample on reply, a short call, or an application link? (Anything you cannot honor this week is deleted from the menu, not ranked low.)
- By when does a reply have to land - is there a launch or campaign date this pitch must beat? (A hard date promotes the near-zero-effort angles, the reply-for-details CTA, and the single-follow-up cadence; no date lets an hour of grounding and the longer cadence pay off.)
- One-off win, or a relationship you intend to compound? (Compounding promotes the manager route and the long-tail cadence; one-off promotes the creator's own listed contact and a hard stop at three messages.)
- What is your effort ceiling per pitch - hours for grounding, product you can ship, calendar you can offer? (An hour of grounding buys the content-specific angle; no product deletes the sample CTA; no calendar deletes the call CTA.)

Every ordering this skill gives 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 proposing anything, then say which answer moved which option:

- an existing relationship with this creator or their manager
- a brand the creator already recognizes
- audience data already in hand
- no product to gift at all

## Workflow

1. Run the Interview; collect every answer before writing a word.
2. Ground the personalization: if you can browse the web, load the creator's two or three most recent posts and confirm the fit evidence yourself; otherwise ask the user for the specific post, series, or audience trait the pitch will cite. Never fabricate familiarity.
3. Confirm the recipient against [references/channel-and-manager.md](references/channel-and-manager.md). The creator's published business contact decides it; where more than one is published, the default order is `listed manager > listed business email > DM`. Going around a listed manager is a failure mode, not a shortcut.
4. Enter brainstorming mode for the personalization angle: propose 2-3 candidate angles with trade-offs and a recommendation, phrased in the manager's own vocabulary (angle, hook, fit, deliverable, CTA). Draw them from the ranked angle menu in [references/pitch-anatomy.md](references/pitch-anatomy.md) - default order `already-recommending > recent launch > specific content > their-product-we-use > format fit > asserted overlap` - and delete any angle that is not true of this creator instead of offering it low.
5. Ask brainstorming questions one at a time. Write no message until the manager explicitly picks an angle.
6. Draft the first touch per [references/pitch-anatomy.md](references/pitch-anatomy.md), fitted to the confirmed recipient. Pick the one CTA from that file's ranked menu - default order `reply for details > sample on reply > short call > application link` - and drop any the brand cannot honor. Compensation transparency follows the model:
   - commission offers front-load exact mechanics
   - flat-fee offers name the model without dumping a rate card
7. Draft the follow-up plan per [references/follow-up-cadence.md](references/follow-up-cadence.md): default order `two follow-ups > one follow-up > long tail`, then close the file.
8. Run your preferred humanizer skill over the pitch and every follow-up. Never ship raw first-draft model output as final - AI-tell phrasing is exactly the mass-template smell that gets a pitch deleted on sight.
9. Self-check against the Quality gate below; redraft and re-run the humanizer until every item passes.
10. Present the package section by section - angle, first touch, follow-up plan, stop rule - and validate each with the user before finalizing.
11. Hand off the full artifact (next section). On a positive reply, this skill's job is done: carry the stated compensation facts into negotiation unchanged.
12. If your harness has persistent memory, memorize the approved angle, brand voice, compensation facts, and the outcome once known - later pitches for the same brand then skip straight past most of the interview.

## Invocation and expected output

Typical invocations:

- "Pitch @plantsbypriya for our ceramic planter launch - she's vetted, 48k on Instagram, offer is gifted product plus 15% commission."
- "Write the first-touch email to the DevOps newsletter writer from our shortlist - $2,000 sponsored deep-dive, she lists a partnerships email."

Deliver one structured artifact:

```
INFLUENCER PITCH PACKAGE - <creator>, <date>
Chosen angle   : approved personalization angle + why it beat the alternatives
First touch    : subject (email only) + message body, fitted to the confirmed channel
Follow-up plan : table - touch / day / channel / angle / one-line summary
Stop rule      : positive reply -> stop and hand off; silence after final touch -> close the file
Handoff notes  : compensation facts as stated, deliverables mentioned, CTA promised - for the negotiation stage
```

## B2B and B2C

- **B2B**
  - Speak to a business need the creator's audience has, not just a product benefit - the one consistently documented B2B-specific rule.
  - Deliverables skew to webinars, white papers, articles, and podcast reads.
  - Gifting is documented B2B practice, but shaped differently than in B2C: extended platform access, a conference pass, a beta invite, or event hosting, rather than a physical product. Zapier's creator-partner retreat is a real named example, and a 2025 industry report ranks in-person gifted events as the second most effective B2B creator content type after social posts. Where none of those forms fit, default to a fee or a commission.
  - Expect longer, multi-stakeholder buying cycles to shape what "conversion" means in the pitch.
- **B2C**
  - Deliverables skew to posts, videos, and stories.
  - Gifting and commission offers are common and well documented.
- Honest limit: the pitch-anatomy and cadence guidance here is published unsegmented by business model, which weakly suggests practitioners treat it as universal. Apply it to both, knowing that universality is an inference, not a documented rule.

## Quality gate

The gate splits by message type: a follow-up is capped at one or two sentences with nothing from the first touch restated (see the cadence reference), so it cannot carry the compensation terms or the full email structure.

Scoring a follow-up against the first-touch list would make the gate unpassable and the loop below never terminate. Iterate, redraft, and re-run the humanizer until nothing fails.

**First touch - all nine items, 9/9.**

1. Opens on one verifiable, creator-specific fact (a named post, series, or audience trait) framed as audience fit, not flattery.
2. The swap test: replacing the creator's name with another creator's would force a rewrite. If the message survives the swap, it is a template.
3. Compensation model and the real number are present and match the interview exactly.
   - Commission pitches state rate, attribution window, and payout schedule.
   - Flat-fee pitches name the model without a rate card.
   - Gifted-only pitches say so plainly and never imply payment.
4. Exactly one ask, low-friction, matching a CTA the manager confirmed they can honor.
5. Fitted to the confirmed channel: full structure for email, 2-4 sentences and no links for a DM, addressed to the manager when one is listed.
6. None of the documented negative signals: generic salutation, emoji-bait subject, fabricated familiarity, feature dump, rate-card dump, aggressive negotiation language, long campaign proposal.
7. B2B pitches state the business need; B2C pitches state the audience-purchase fit.
8. No disclosure language, contract terms, or exclusivity demands - documented practitioner workflows place all of these after agreement, never in the cold pitch.
9. Passes the humanizer and reads human when spoken aloud.

**Every follow-up - items 4, 6, 8 and 9 above, plus these three, 7/7.**

10. One or two sentences. Nothing from the first touch restated.
11. Carries one new small angle - a proof point held back from message one, an easier version of the ask, or a genuine timing hook. Never "just checking in", "bumping this", or "did you see my email?".
12. Low pressure: no guilt-tripping, no invented deadline, nothing that reads as chasing. Silence is busy, not rejection.

Items 1, 2, 3, 5 and 7 are first-touch content by construction and are not scored on follow-ups. Item 11 does the work item 2 does on the first touch: a follow-up that could be sent to any creator unchanged fails it.

## KPIs

- No trustworthy reply-rate benchmark exists for creator outreach. The only survey on the topic explicitly declines to publish one. Never quote any of the circulating figures as fact or as a target - each traces to broken citations or a single-source self-report:
  - an "8.5% reply rate"
  - a "99% ignored" claim for generic salutations
  - one practitioner's self-reported "90% response rate"
- Benchmark against your own log instead: track every pitch in a flat table (creator, channel, angle, sent date, follow-ups used, outcome, days-to-reply). After a dozen pitches, your own reply rate per angle and per channel is the benchmark that matters.
- Core KPIs:
  - reply rate (any response)
  - positive-reply rate (open to talking)
  - days to first reply
  - which angles produced pitches that survived negotiation, recorded after the downstream stage closes
- If your harness has persistent memory, keep the log there across sessions; otherwise hand the table to the user to maintain.

## Common failure modes

| Failure                                                   | Fix                                                                                                                    |
| --------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------- |
| Template smell - personalization swappable to any creator | Rerun the angle brainstorm; cite one verifiable specific; apply the swap test                                          |
| Compensation question buried or omitted                   | State model + real number per the anatomy rules; vagueness reads as a lowball                                          |
| Copying a paid-collab template for a commission offer     | Commission flips the risk onto the creator - front-load rate, window, payout, per-sale estimate                        |
| Pitching the wrong channel                                | Re-run the recipient gate in step 3 before anything else - the published contact decides, the ranking only breaks ties |
| Ignoring or going around a listed manager                 | Pitch the manager with full personalization; go direct only when the creator has asked for it                          |
| Fabricated familiarity ("longtime fan of your work")      | One verifiable specific, or say honestly how you found them                                                            |
| Rate card or full campaign proposal in message one        | One-line deliverable sketch; details belong after the reply                                                            |
| Disclosure, contract, or exclusivity terms in the pitch   | Post-agreement topics; route disclosure to the compliance skill                                                        |
| Negotiating in the pitch thread after a positive reply    | Stop - the handoff notes carry the facts into the negotiation stage                                                    |

## Reference

- See [references/examples.md](references/examples.md) before writing copy - a positive B2C and B2B example with follow-up plan, and a line-by-line annotated negative example.
- See `mbfinotti/partnerships-skills@influencer-discovery-brief` for choosing and vetting which creator to pitch - this skill assumes that work is done.
- See `mbfinotti/partnerships-skills@influencer-negotiation-playbook` for everything after the creator replies with interest.
- See `mbfinotti/partnerships-skills@influencer-campaign-brief` for the creative brief once the deal is signed.
- See `mbfinotti/partnerships-skills@affiliate-recruitment-outreach` for recruiting affiliates across a prospect list rather than pitching one vetted creator.
- See `mbfinotti/partnerships-skills@affiliate-disclosure-compliance` for the disclosure obligations that attach once a material connection exists.