// execute

Creator Partnerships

Use when the user wants to work with creators or influencers - choosing who to pay, agreeing terms, or judging whether a sponsorship worked - on YouTube, LinkedIn, podcasts, newsletters, TikTok or Instagram. Also use when the user mentions influencer marketing, creator marketing, sponsorships, paid posts, B2B creators, podcast ads, newsletter sponsorships, or "is this creator worth it". Compares creators on the buyers they actually reach rather than follower counts, writes the terms both sides sign - including disclosure - and reads the result on what it sold, tracked and self-reported.

ExecuteengineIntermediate
bash
$npx skills add sarojkjha/aaj-marketing-skills --skill creator-partnerships
Or run it in Claude or ChatGPT through the AAJ engines connector

Claude Code

/plugin marketplace add sarojkjha/aaj-marketing-skills
/plugin install aaj-marketing-skills@aaj

OpenAI Codex

codex plugin marketplace add sarojkjha/aaj-marketing-skills
codex plugin add aaj-marketing-skills@aaj

Claude or ChatGPT — no install

Add this as a custom connector:

https://aaj-engines-mcp.sarojkjha10.workers.dev/mcp
Recipes: vendor data in, a verdict out

// when to use it

Use when the user wants to work with creators or influencers - choosing who to pay, agreeing terms, or judging whether a sponsorship worked - on YouTube, LinkedIn, podcasts, newsletters, TikTok or Instagram. Also use when the user mentions influencer marketing, creator marketing, sponsorships, paid posts, B2B creators, podcast ads, newsletter sponsorships, or "is this creator worth it". Compares creators on the buyers they actually reach rather than follower counts, writes the terms both sides sign - including disclosure - and reads the result on what it sold, tracked and self-reported.

// what you give

The creators on your shortlist with their fee, median views, engagements, the share of their audience that fits your buyer and where that share came from, how many recent posts were sponsored and whether they disclose; the terms agreed; afterwards the tracked conversions and survey mentions per creator and your target cost per acquisition

// what you get

The shortlist ranked by cost per thousand people who fit your buyer, with every gap in the data named; the contract terms still missing; and a keep, renegotiate, drop or too-early call per creator

Engine spec

// decision it supports

Which creators reach our buyers at the lowest cost, what terms are still missing, and which sponsorships should we keep or drop?

// inputs

The creators on your shortlist with their fee, median views, engagements, the share of their audience that fits your buyer and where that share came from, how many recent posts were sponsored and whether they disclose; the terms agreed; afterwards the tracked conversions and survey mentions per creator and your target cost per acquisition

// output

The shortlist ranked by cost per thousand people who fit your buyer, with every gap in the data named; the contract terms still missing; and a keep, renegotiate, drop or too-early call per creator

// acts or analyzes

Analyzes only

Makes no network calls, writes no files and changes nothing in your tools.

// where your information goes

How you run itWhere your inputs go
Skill installed in your agent, or the local MCP serverThe engine runs on your machine and AAJ receives nothing. Your AI agent still sends what you give it to its model provider (Anthropic, OpenAI and so on), under your own account's terms.
Hosted MCP serverYour agent's model provider, as above, plus AAJ's server on Cloudflare, which receives the engine inputs over HTTPS, runs them in memory and keeps no copy. Cloudflare keeps error logs (no inputs) for up to 7 days.

No engine needs customer names, emails or other personal data; leave them out.

// limitations

  • Cost per thousand matched people rests on the creator's median views and audience-match share. If those come from a media kit, not platform analytics, the ranking is the creator's claim.
  • Creators are ranked only against each other on your shortlist. No outside benchmark is used, so the cheapest creator on the list can still be poor value.
  • Tracked conversions undercount and survey mentions can overlap them, so the true cost sits between the two figures. Under five conversions in total the call is too early.
  • The contract check confirms the eight terms are marked agreed. It does not review their wording and is not a legal check of FTC or local disclosure rules.

// review before acting

The marketing owner checks each creator's views and audience share against a platform analytics screenshot or export, and reviews disclosure wording and placement in every post before it goes live.

// setup

install
$npx skills add sarojkjha/aaj-marketing-skills --skill creator-partnerships
hosted mcp
$https://aaj-engines-mcp.sarojkjha10.workers.dev/mcp

Agents listed: Claude Code, Cursor, OpenAI Codex, Windsurf, Cline. Tested and recorded: Claude Code.

Full spec on GitHub

Creator Partnerships

A creator's follower count is the number most sponsorships are priced on and the one that matters least. What you are buying is a recommendation, heard by people who could become customers, from someone they trust. So the questions are: how many people see a post, how many of them fit your buyer, what does each of those cost, and afterwards, what did it sell.

This skill answers those four questions and writes down the terms in between.

When to use

The user is choosing creators to sponsor, negotiating with one, or looking back at a round of sponsorships and deciding what to renew.

Do not use it to find creators from nothing. Build the shortlist from where your buyers already are: the podcasts, newsletters and accounts customers name when you ask how they heard about you. The Self-Reported Attribution Coder turns those answers into a list.

Before you start

  1. Know your buyer in one sentence (.agents/product-marketing.md), because every creator is judged on the share of their audience that fits it.
  2. Ask each creator for three things before talking price: median views on their last ten posts (not followers), the audience breakdown from their own platform analytics as a screenshot or export, and links to three past sponsored posts.
  3. Set a target cost per acquisition from your unit economics, so the result can be judged against something.
  4. Set up tracking before anything is posted: a unique discount code and a tagged link per creator, and an open "How did you hear about us?" field on your signup or checkout.

Method

Price on matched reach, not followers. The engine multiplies median views by the share of the audience that fits your buyer, then divides the fee by it: the cost per thousand people who could buy. Creators are compared with each other on your own shortlist, so no outside benchmark is needed and none is used.

Name every gap instead of guessing. A follower count with no views, an audience share with no source, or a missing audience breakdown makes a creator not comparable yet. The engine says what to ask for.

Check the feed and the disclosure. The engine flags a creator whose recent posts are mostly sponsored (above half, which is a rule of this engine) so you compare how their paid posts perform against their own before buying one. It also flags a creator who does not disclose.

Disclosure is not optional, and it is partly yours. In the US, the FTC expects an influencer to disclose any financial, employment, personal or family relationship with a brand. The disclosure should sit with the endorsement itself so it is hard to miss - not only in a profile, at the end, or behind "more" - and in a video it belongs in the video. The FTC also expects advertisers to have reasonable programs to train and monitor the people promoting them. Put the wording and placement in the contract and review posts before they go live. Outside the US, check the local rules.

Write eight terms before work starts:

  1. deliverables and dates;
  2. disclosure wording and placement;
  3. review before posting;
  4. usage rights;
  5. an exclusivity window;
  6. tracking;
  7. payment terms;
  8. what happens if a post is removed or underdelivers.

The engine lists the ones not yet agreed.

Judge on what it sold, counted two ways. Tracked conversions (code and link) undercount, because many people hear a recommendation and later search for you. Survey mentions recover some of that, but they can overlap with tracked ones. So the engine reports cost per tracked conversion (the ceiling) and cost counting every survey mention as extra (the floor). The call:

  • keep: tracked alone meets the target;
  • renegotiate: only the floor meets it;
  • drop: neither meets it;
  • too early: fewer than five conversions in total, which is a rule of this engine.

Workflow

Step 1 - Shortlist. Five to ten creators your buyers already follow. Collect views, audience breakdown, sponsored history and fee for each.

Step 2 - Compare. Run the engine's shortlist check. Fill the gaps it names; drop anyone who will not share analytics.

Step 3 - Agree the terms. Run the contract check until all eight are agreed, with disclosure wording written out.

Step 4 - Track. Code, tagged link and the survey question live before the first post.

Step 5 - Read and decide. After the posts have had time to work, give the engine each creator's fee, tracked conversions and survey mentions against your target. Renew, renegotiate or stop per creator, not for the programme as a whole.

Run the tool

node resources/creator-fit.js --demo

run({ shortlist, contract, results }) returns the ranked shortlist with each creator's cost per thousand matched people and findings, the contract terms still missing, and a verdict per creator with both cost figures.

Present the result

  1. The shortlist ranked by cost per thousand matched people, with anyone not yet comparable listed separately with what to ask them for.
  2. The terms still missing, especially disclosure.
  3. Per creator after the round: fee, tracked conversions, survey mentions, both costs, and the call.
  4. The next round: who to renew, what to renegotiate, and one new creator to test.

Guardrails & common mistakes

  • Never price on followers. Ask for views and audience data; walk away if a creator will not share them.
  • Never accept a media-kit claim as audience data. Ask for the platform analytics.
  • Never let a post go live without disclosure. It protects the creator's audience and it is the brand's responsibility too.
  • Do not judge on views or likes. Judge on conversions, counted both ways.
  • Do not renew the programme as a whole. Creators vary; decide one by one.

Sources

  • partnerships-and-co-marketing - the same discipline with a company as the partner
  • founder-led-content - building your own audience instead of renting one
  • paid-media-budget-allocation - where creator spend sits against other paid channels
  • content-repurposing - reusing creator content, once usage rights allow it

// faq

// install

$npx skills add sarojkjha/aaj-marketing-skills --skill creator-partnerships

// worked example

Get the worked example for Creator Partnerships

One email. The filled-in config, the output, and how to read it.

No spam. Unsubscribe anytime.

// human in the loop

Need a human in the loop?

AAJ builds the strategies, tools, and playbooks these skills are made of — with engagements for growth-stage and enterprise teams.