A podcast sponsorship media kit has six parts: a short show description, audience statistics from your own analytics, the ad formats you sell, a rate card, evidence from past partnerships, and a contact line. AI can assemble it from real download data you paste in, but every number must come from your hosting dashboard and be checked before it goes out.
Examples of media kits are everywhere, which is the problem: most show attractive pages full of figures with no source. Sponsors have seen a lot of these. What sets a kit apart is that each number can be traced to a dashboard and each price can be explained. The way to get there is to start from exports rather than a template, and to finish with a single filled page where every figure carries its source.
Pull the data before you open an AI tool
Start with exports, not summaries. A sponsor cares about five things, and each has a source you can open now.
- Downloads per episode from your podcast hosting dashboard, as an export: episode title, publish date, downloads at 7, 30 and 90 days if available.
- Audience location and device from the same dashboard, if it offers them, which tells sponsors where listeners are and what they listen on.
- Platform split from Apple Podcasts Connect, whose analytics report followers, listeners, engaged listeners (people who played at least 20 minutes or 40 per cent of an episode) and average consumption, and from Spotify for Creators, whose analytics include stream counts. Check which fields your own dashboards offer.
- Video views and watch time from YouTube Studio, if you publish video, as a separate line so audio and video are not double counted.
- Newsletter and social reach if you sell those: subscriber count, open rate and average clicks, from your email platform.
Download the data for a fixed period, such as the last 90 days, and note the dates. Hosting dashboards count downloads differently, for example in how they filter repeat requests and automated traffic, so ask your host whether its counts follow the industry measurement standard and how it treats bots. A sponsor's agency may ask, and a short, accurate answer is a good sign.
Work out the numbers sponsors actually use
Three calculations do most of the work. You can do them in a spreadsheet, and AI can check them, but understand them yourself first.
Average downloads per episode at 30 days. Use the median or the mean of your last ten episodes, and say which. A median is harder to inflate with one lucky episode.
Downloads per month. Add the 30-day downloads of episodes published in the period. Do not add figures for older episodes unless you label them as back-catalogue.
CPM, cost per thousand downloads. This is the price divided by downloads in thousands. If a 60-second host-read ad costs $400 and an episode averages 8,000 downloads at 30 days, the CPM is $400 divided by 8, which is $50. These are illustrative figures. The formula is what matters: sponsors compare quotes in CPM, so your kit should show how your price translates.
Do not copy market rates from other kits or ask an AI to tell you what podcasts "usually charge". Without a verified source, those figures are guesses, and the model will state them with confidence. Instead, base your rates on your own numbers, the format, and what the sponsor gets, and test with a small number of sponsors. If you want a range to anchor against, ask a professional association or an agency you work with, and cite them in your notes.
The Complete Excel Course with AI Integration
From formulas and pivot tables to Power Query, DAX, VBA and AI with Copilot, ChatGPT and Claude — hands-on from the first lesson.
By Sagnik Bhattacharya and Virginia Thorn, Primetime Emmy Award winner
Buy nowThe main example: building the kit for a weekly business show
Say a weekly interview show for small-business owners has 52 episodes in the back catalogue and 13 in the last quarter. All figures from here are illustrative; replace them with yours. From the hosting export, the last ten episodes average 6,200 downloads at 30 days, with a median of 5,900. Last quarter's 13 episodes produced 78,000 downloads within 30 of release. The audience skews towards desktop and mobile apps in roughly equal parts, and Apple Podcasts and Spotify together account for the majority of listens. A short YouTube version has 2,400 views a month and an average watch time of 11 minutes. The newsletter has 3,100 subscribers with a 46 per cent open rate. (Apple's Mail Privacy Protection stops senders seeing whether a message was opened, so open rates are unreliable; say how you measure it, or lead with clicks instead.)
Paste a cleaned table of those figures into the prompt, with a clear instruction.
You are helping a podcaster write a sponsorship media kit.
Use ONLY the figures in the data block. Do not add, round up or estimate any
number. If a number is missing, write "[to confirm]". Label every figure with
its source and period. Use British English. Plain, factual tone; no hype words.
Data (period: last 13 episodes, [start date] to [end date]):
- Hosting dashboard: median downloads at 30 days per episode: 5,900;
mean 6,200; total 78,000 for the period.
- YouTube Studio: 2,400 views a month; average view duration 11 minutes.
- Newsletter (email platform): 3,100 subscribers; open rate 46% (note:
unreliable because of mail privacy features).
- Audience: small-business owners and managers (from our listener survey,
62 responses; mark as self-reported).
- Ad formats we sell: 30-second pre-roll, 60-second host-read mid-roll,
newsletter mention, 15-minute sponsored segment.
Write: (1) a 60-word show description, (2) an "Audience at a glance" table,
(3) the ad formats with one line each on what the sponsor receives,
(4) a note explaining how downloads are counted, and (5) a list of the
numbers you would want a sponsor to ask about, so I can prepare answers.A realistic output gives you a neat table, a clear description and a list of likely questions such as "how many downloads come from the first 7 days?". The edits you will need: the model may write "over 6,000 listeners per episode", which treats downloads as people, so change it to "downloads". It may describe the audience as "decision-makers" when your survey only shows owners and managers, so use their words. And it may add the phrase "highly engaged", which you cannot support unless you have a completion figure. Strike every adjective that has no number behind it.
Ad formats and what each should include
Describe each format by what the sponsor gets and what they have to supply, so there is no confusion later.
| Format | What the sponsor gets | What you need from them | Things to state |
|---|---|---|---|
| Pre-roll, 30 seconds | A short read at the start of the episode | Key message, offer, link or code | Whether inserted dynamically or baked in; how long it stays in the back catalogue |
| Mid-roll, 60 seconds, host-read | A personal read in the middle of the episode | Talking points, product access for the host | Whether the host will say they use the product, and only if true |
| Newsletter mention | A short paragraph with a link | Copy, link, image | Subscriber count and measurement method |
| Sponsored segment, up to 15 minutes | A conversation or story featuring the product | Brief, approval timeline | Disclosure wording and editorial independence |
| Series sponsorship | Name and read across several episodes | Longer commitment | Number of episodes, dates, renewal option |
Say whether ads are inserted dynamically or baked into the audio. Dynamic insertion lets you update or remove ads in older episodes, and baked-in ads stay forever. Sponsors ask, because it affects how long their message runs and how you count downloads after the first month.
Building a rate card from your own numbers
Pick a method and explain it in the kit. A simple one: choose a target CPM for each format, multiply by your median 30-day downloads in thousands, and round to a clean figure. Here is an illustrative card using the show above, with a 5,900 median (5.9 thousand downloads) and example CPMs that stand for your own choice, not market rates.
| Format | Illustrative CPM | Maths | Price per episode |
|---|---|---|---|
| Pre-roll, 30 seconds | $18 | 5.9 x $18 = $106 | $105 |
| Mid-roll, 60 seconds | $25 | 5.9 x $25 = $147.50 | $150 |
| Newsletter mention | $40 per thousand subscribers | 3.1 x $40 = $124 | $125 |
Then offer packages: three mid-rolls for the price of 2.5, or a series of six episodes with a pre-roll and a newsletter mention bundled. Ask the AI to check your arithmetic and to flag any package whose discount makes the effective CPM lower than your lowest format. That is a good use of a model: it reads the table and reports inconsistencies, and you decide what to do.
If you are unsure whether to publish prices, publish "from" prices for each format and a line saying packages are quoted to the campaign. Whichever you choose, the sponsor should be able to reproduce your price from your audience figures in a line or two.
Past results without invented testimonials
Evidence of past results is the section most kits fake, and the one sponsors trust most when it is real. If you have run sponsorships before, collect what you actually have: click counts on a tracked link, redemptions of a promo code, or a short written comment from the sponsor, with their permission to quote it. State the period and the method. If you have no past sponsors, do not invent them. Use first-party evidence: how many listeners replied to a survey, how many clicked a link you shared in the newsletter, how a previous affiliate promotion performed. You can also offer a first-partner arrangement, such as a reduced rate in exchange for a case study you both approve.
A filled example of a past-results line: "Newsletter link to a sponsor's free guide, March: 214 clicks from 3,000 subscribers (tracked link). Sponsor reported 31 sign-ups." Label it as an illustration here; in your own kit, the numbers would come from the sponsor's report and your link tracker. Our piece on sponsor reports after an event shows how to gather and present this kind of result.
One filled example page
This is how the first page of a finished kit might look, with all figures illustrative. A real kit should list the date and the period covered at the top.
[Show name]: sponsorship overview (data period: 13 episodes, [start date] to [end date]; updated [date])
A weekly interview show for owners and managers of small businesses. New episodes every Tuesday.
Audience (source: hosting dashboard, YouTube Studio, email platform): median 5,900 downloads per episode at 30 days; 78,000 downloads over the period; 2,400 YouTube views a month; 3,100 newsletter subscribers. Listener survey (62 responses, self-reported): most respondents run or manage a business with fewer than 20 staff.
Formats and prices per episode: 30-second pre-roll $105; 60-second host-read mid-roll $150; newsletter mention $125; series packages on request.
How we count: downloads are as reported by our hosting provider at 30 days after publication; we publish the method on request.
Past results: [add real results only].
Contact: [name and address].
Notice what it leaves out: no adjectives, no "premium audience", no figure without a source. That restraint is what makes the kit credible, and AI helps you keep to it if you tell it so. Keep a "sources" tab in your spreadsheet that matches each number in the kit to the dashboard page it came from.
Three situations that change what the kit should say
A small show in a tight niche
Say a show for independent gift-shop owners averages 900 downloads at 30 days. A general sponsor will pass on those numbers, but a supplier of shop fixtures or a payments company serving small retailers may value a concentrated audience of its exact customers. Lead with the survey of who listens, the niche and the specific problem the show solves. State the CPM honestly. A niche show can ask for a higher CPM than a general one, but only if it can justify the price with audience fit, not with adjectives. A sponsor who sells to that audience can run a direct test: a code, a tracked link, one episode.
A show whose audience is mostly on video
Some shows get more consumption on YouTube than in podcast apps. Present it that way. Give video views, average view duration and subscriber count from YouTube Studio as their own block, with a note that the audio and video figures are separate and are not added together. Offer a format that suits video, such as a 30-second integration with an on-screen product shot, and price it with your own CPM logic on views.
A show with strong seasonality
A sports club podcast that peaks during the season and slumps in summer should show both. Give the median for the season months and the median for the off months, and offer packages timed for the peak. Sponsors prefer an honest seasonal pattern to a flat average that hides it, because they can plan a campaign around it. Pull the monthly numbers from your dashboard and ask the AI to draw out the pattern in one sentence, then check it against the chart.
A realistic mistake and how it showed up
A producer pastes only the headline from the dashboard (the total downloads since launch, 410,000) and asks the AI for "a persuasive audience section". The draft says "Our 410,000 listeners tune in every month". It confuses lifetime downloads with monthly listeners. The sponsor's agency asks for a monthly figure, finds 78,000 downloads in the last quarter and a median of 5,900 per episode, and the discrepancy ends the conversation. The fix has three parts: give the AI period-specific data only, add the rule "never present lifetime totals as current audience", and have someone not involved in writing the kit compare each sentence with the source before it goes out.
Keeping the kit current with little effort
Treat the kit as a document with a refresh routine. Once a quarter, export the same columns from the same dashboards, paste them into the same prompt and ask for the same table. Keep the prompt, the data block and the output together in one folder named by quarter, so you can see how the numbers moved. Ask the AI to list what changed against the previous quarter and to flag any figure that moved by more than a set amount, say 20 per cent, so you can check it before it reaches a sponsor. Then update the date on the first page.
This routine takes about an hour once it exists, and it has a second benefit: when a sponsor asks for a mid-quarter update, you can answer in an afternoon with figures you trust. The effort is in building the first version well. After that, the AI handles the reformatting, and you stay in charge of the numbers and the promises that go with them.
Checking the kit before a sponsor sees it
- Every number matches the export or the dashboard on the day you send it.
- Every figure has a source and a period.
- Downloads are called downloads, not listeners, unless your platform measures listeners.
- The rate card can be reproduced from the audience numbers.
- No testimonial, client name or result appears unless it is real and approved.
- The date is on the first page.
Ask a colleague, or a second AI session with no memory of the first, to try to find any number that does not trace to the data block. If it cannot, you are in good shape. For a different type of one-page kit, aimed at press and partners, see how to build a one-page media kit with AI.
Using the kit in a pitch
A kit is a reference, not a pitch. Send a short email that names a specific reason you think the sponsor's audience matches yours, attaches the kit, and proposes one format and one date. A tiered approach works well for first conversations: offer a single episode trial at your standard rate, then a package on renewal. For the structure of tiers and pitch emails, our tutorial on event sponsorship proposals with AI has wording you can adapt, and a strong guest line-up from your guest outreach gives you something concrete to point to.
Also plan the other side: what you will report afterwards. Tell the sponsor in the pitch that you will send downloads at 30 days, link clicks and any code redemptions, and when. A sponsor who knows they will get a report is easier to renew, and the report becomes the next entry in your past-results section. Over a year, that loop of real numbers, honest pricing and real results is what builds a kit sponsors take seriously.
Video and growth figures worth adding as they become real
If you are building a video audience, add it to the kit as its own line with its own source, since YouTube Studio and your hosting dashboard count different things. Our guide to turning audio into video covers how to do that without the result looking cheap. Show growth only when it is real and recent: a quarter-on-quarter change in median downloads is more useful to a sponsor than a headline lifetime total, and a flat but steady figure can be a selling point when you describe it as stable. Whatever you show, put the period next to it.
Questions podcasters ask about media kits
What if my numbers are small?
Show them honestly and lead with what is strong: audience fit, completion rate, a niche sponsors cannot reach elsewhere, and replies or clicks from past promotions. A small, engaged audience in a tight niche can be worth more to the right sponsor than a large general one. Never round up or borrow numbers.
Should the media kit list prices?
A rate card builds trust and saves time, but some producers prefer a range or a 'from' price with a note that packages are tailored. Choose one approach and keep it consistent. Whatever you publish, make sure the maths behind it is something you can explain in one line.
Can AI write the kit from my analytics?
It can organise the numbers, draft the wording and lay out a rate card, as long as you paste in real figures and tell it not to add any. You must check every number against the source dashboard before the kit leaves your hands.
How often should I update it?
At least quarterly, and whenever a figure changes materially. Put the date and the period covered on the first page, so a sponsor knows how recent the numbers are. An out-of-date kit with a confident date on it is worse than a modest one that is current.
Further reads
- How to Write Podcast Show Notes With AI: A Template for Producers — Show notes are where sponsor links and codes live.
- AI for Podcast Clips: Find, Cut and Caption Social Clips — Clips give sponsors extra reach and you extra evidence.
- Descript vs Riverside for Podcast Producers: Which to Standardise On — Pick the tools that produce the exports you will quote.
- AI That Creates Podcasts: Can It Replace a Production Company? — Understand where AI-made audio fits with sponsorship.
- Can AI Edit a Podcast? What It Does Well and Where It Fails — AI handles fillers, silence, noise and levelling well. Crosstalk, tone and meaning still need a human ear. Which edits to trust, and a listening test.
- AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.
Sources: Apple Podcasts for Creators help (analytics metric definitions), Spotify for Creators, Apple's Mail Privacy Protection help page; your own hosting, YouTube and email analytics. No market rates are quoted because none could be verified from a primary source.