--- name: creator_media_kit description: Build a creator media kit that brands actually read: audience proof with real numbers, past work matched to the buyer, a clear format menu, and contact and rate pointers, kept to a short document that stays current. Use when a creator asks for a media kit, a one sheet, a rate pointer sheet, or a refresh of an old kit, or when past brand work needs to become a short case study. --- # Creator Media Kit ## Purpose Build a short media kit a brand reader can scan in under a minute and verify in under five. The kit answers four questions in order: who is this creator, who follows them and how do we know, what have they made for brands like ours and what happened, and what can we buy, at what starting price, from whom. A kit is a proof document, not a portfolio and not a bio page. Every number in it traces to a platform analytics screen or export, dated, with the formula and the time window stated. If a number cannot be traced, it leaves the kit. ## Workflow 1. Pick the form before writing. Default to two files that share one source of numbers. - A one page kit for first contact. Identity, audience proof, one matched case study, format menu in short form, starting rates, contact. - An extended kit of two to three pages for a reader who asked for more. Full demographics, two or three case studies, full format menu, process, terms pointers. - Never send a long kit as a first contact document. Never build a separate deck per brand. Tailor by swapping the featured case study, nothing else. 2. Pull proof from the native analytics, in this order, and record the pull. - Instagram, professional account: Professional Dashboard, Insights, Total followers for age range, gender split, top cities and top countries. Content insights for reach and engagements per post. - TikTok, Creator Center analytics: video views, profile views, likes, comments, shares, followers, and per video average watch time, full video watches, reached audience, and viewer region where the account shows it. - YouTube, YouTube Studio Analytics: views, watch time in hours, average view duration, average percentage viewed, unique viewers, and audience age, gender, and top geography. - Newsletter or blog, from the sending or analytics tool: subscriber or reader count, open rate and click rate over a stated window, with the tool named. - Write each pull into the proof log in `references/proof-standards.md`: metric, value, platform screen or export, date pulled, window covered. Screenshot or export the screen the same day. A kit number with no log entry does not ship. 3. Compute performance with one stated method, and show the method. - Engagement rate, follower based: total engagements divided by followers at the time of posting, times 100. Engagements means likes plus comments plus shares plus saves where the platform counts them. State which parts were counted. - Report the median post in the window, not the best post. Use the last 12 to 20 organic posts or the last 30 days, whichever holds more posts, and state the window and the post count. - Reach based engagement is a different number. If you use it, label it reach based and give reach as the denominator. Never place a follower based rate and a reach based rate side by side without labels. - For video, report average views per video, average view duration, and average percentage viewed where the platform gives it. A view count alone proves distribution, not attention. - Run the checker in step 7 on every rate you print. Hand math in a kit is where errors live. 4. Write audience proof as fit, not size. - Lead with the split that decides fit for most buyers: top age groups, gender split if the platform reports it, top three countries, top three cities if the creator sells locally or regionally, and two or three audience interests supported by content and comment evidence. - Percentages for one split must sum to 100. If you list only the top groups, add an Other line for the remainder. See the sum check in the checker script. - Order platforms by engaged audience, not by follower count. A smaller platform with the right split and a stronger median rate goes first. - Cut vanity lines: total impressions with no window, lifetime views, press mentions with no link, and follower totals repeated on every page. 5. Build case studies matched to the buyer, including gifted work. - Use the block in `references/proof-standards.md`: brand and category, brief in one line, deliverables, live link, window, results with denominators, and a one line quote only with written permission from the person quoted. - Label each study Paid, Gifted, Affiliate, or Self published spec work. Gifted and trade work prove execution. They do not prove paid conversion, and the kit must not imply that they do. - Keep three to five studies on file and feature one to three per kit, chosen by category match to the reader. A study for a direct competitor is shown only if the contract allows it and no exclusivity term blocks it. - Past results are history, not a promise. Write the window and the conditions, for example paid support behind the post or a giveaway mechanic, next to the result. Omit a result you cannot source from the brand report or the platform screen. 6. Write the format menu and the rate pointer so a buyer can budget. - Menu rows: format, what the buyer gets, posting or delivery only, typical timeline, starting rate. Cover at least sponsored post or video, short video, story or ephemeral set, UGC delivered without posting, and live or event appearance if offered. - Price the add ons as separate lines, never buried in the base rate: usage rights by duration and channel, whitelisting or paid amplification rights, exclusivity by category and duration, extra revision rounds, rush delivery, travel. - If full rates stay off the kit, the pointer must still carry a starting rate per core format, what moves the price, and the reply commitment, for example starting rates current as of a date, full schedule shared after a short fit call. Rates upon request with no number and no reason is a dead end, so do not print it. - State what is not offered as plainly as what is: categories refused, no posting without disclosure, no script approval beyond factual and legal points, or your own equivalent lines. 7. Assemble, check, and date the kit. - Follow the page order in `templates/kit-one-pager.md` for the one pager and add demographics detail and extra studies after page one for the extended kit. - Put business contact on page one and again on the last page: business email, manager or agent contact if there is one, base location and time zone, and preferred contact method. Social handles are not a contact method for a deal. - Stamp the kit: As of YYYY-MM-DD on page one, next to the headline numbers. No undated kit leaves the creator. - Run the checker and fix every error before export: ```bash python3 scripts/mediakit_check.py templates/kit-one-pager.md ``` Replace the template path with the finished kit draft in Markdown. Warnings need a written reason in the proof log, not silent acceptance. - Export the one pager as a single page PDF and check it on a phone at full width: contact tap targets work, no text needs a zoom, no chart label is cut off. 8. Keep the kit current on a fixed cadence. - Refresh the numbers monthly, and before any pitch after a viral post, a platform change, or a rate change. Update the As of date every time, even if only one number moved. - Retire a case study when its links die, its brand asks for removal, or its results window is older than the work the creator wants to sell now. Keep the retired file in an archive folder, out of the live kit. - Keep one master numbers file. Both the one pager and the extended kit read from it. Two kits with two different follower counts for the same date is the fastest trust loss in this process. ## Output Contract Deliver: - A one page media kit in Markdown and PDF, stamped with an As of date, with contact on the page. - An extended kit of two to three pages when the reader asked for depth, built from the same master numbers. - A proof log covering every printed number: metric, value, source screen or export, date pulled, window. - A format menu with starting rates or a rate pointer that carries a starting number, price movers, and a reply commitment. - A checker report: errors fixed, warnings left, with a reason per warning. - Three to five case study blocks on file, with the featured set noted per recipient category. ## Operating Rules - Verify every factual claim against a primary source or the platform screen it came from. No invented numbers, dates, quotes, or results. Unverifiable claims are cut, not softened with words like impressive or huge. - Do not quote a universal good engagement rate. Published benchmark ranges conflict by method and by source, so the kit states its own rate, formula, window, and median basis, and lets fit do the selling. - Disclose the nature of past work. Under the FTC Endorsement Guides, a material connection includes payment, free or discounted products, and personal, family, or employment ties, and the endorser carries a disclosure duty of their own. Label Paid, Gifted, and Affiliate work in the kit, and state in the format menu that sponsored content runs with a clear disclosure up front, in the same format as the content. - Use brand logos and quotes only with permission. A logo wall built from gifted samples or one off purchases misstates the relationship. - Show demographics as the platform reports them. Do not round a split into a cleaner story, do not merge windows from different pulls, and do not present an engaged audience split as a follower split. - Keep the kit short. If a section does not help a buyer judge fit, proof, formats, price, or contact, it belongs on a website, not in the kit. - Protect the creator in the menu: usage, exclusivity, and whitelisting are separate priced rights, revision rounds are capped in writing, and payment terms are stated in the extended kit or the rate schedule that follows it.