Podcast cover art: the specs, and the part that decides a tap

Apple Podcasts will take a show cover anywhere from 1,400 by 1,400 pixels up to 3,000 by 3,000, and almost every podcaster submits the big one. Then the file gets rendered at roughly the size of a postage stamp in a browse row, and the three-line show name that looked balanced in Canva turns into grey mush. Podcast cover art is the only piece of marketing most shows make once and never look at again, which is strange, because it is the single asset standing between a stranger and a tap.

This is the practical version: the specs that will actually get a feed accepted, the design decisions that survive being shrunk, and the honest limits of what a new image can do for a show that nobody is finding in the first place.

Two numbers Apple checks before anything else

Apple publishes its show cover requirements in Apple Podcasts Connect, and they are short enough to memorize. The target size is "3000 x 3000 pixels." For shows delivered the normal way, through an RSS feed, Apple states that it "accepts Show Cover artwork ranging from 1400 x 1400 to 3000 x 3000 pixels." The accepted formats are "PNG or JPG."

One more rule catches people out. Apple says show covers "cannot contain transparency and should not contain an alpha channel," and that the file needs "a solid background before exporting." A logo designed on a transparent canvas and exported straight to PNG will carry that alpha channel along with it, and the result is either a rejection or a cover that renders black on some surfaces and white on others. Flatten it onto a real background first.

Spotify handles the same file through its own pipeline. Spotify's help documentation for creators notes that a cover art change usually takes about 24 hours to show up, so a new image is not a same-day fix. If a podcaster is swapping art ahead of a launch or a press mention, that is a day of lead time to build in, not an afternoon.

The gap between the export size and the viewing size

The file is square and enormous. The place it gets seen is small and crowded. In every major app the show cover appears first as a thumbnail in a search result row, a category shelf, or a "you might like" carousel, sitting beside a dozen competitors at the same size.

Apple's own artwork guidance reads like it was written by someone who has watched this go wrong. The guidance is to keep the podcast title "clear and prominent," to "prioritize legibility" with strong contrast between text and background, and to use high resolution source images rather than upscaled ones. Apple also tells podcasters not to add "Apple logos or hardware" to the art, and to avoid black or white backgrounds that blend into the app interface, which is a real risk now that most listeners are in dark mode and a black-backed cover dissolves into the screen around it.

The working test is cheap. Export the cover, drop it into a document, and scale it to about 55 pixels wide. Look at it on a phone at arm's length. If the show name is unreadable at that size, the cover has no title on it as far as a browsing listener is concerned, and whatever subtitle, tagline, host name, or microphone illustration is competing for that space should come out.

Three things survive the shrink: a short word or two of the title set large, one high-contrast color pairing, and a single recognizable shape or face. Nearly everything else is decoration that only exists on the designer's monitor.

The faces question, and when a host belongs on the cover

Photography on a cover works when the face is the product. Interview shows built on a known host, true crime shows with a recognizable narrator, and any podcast whose audience arrived from a personal brand all get real recognition value from a portrait, because the thumbnail is doing the same job a book jacket does for a famous author.

It works badly for everyone else. A tightly cropped photo of an unknown person at thumbnail size is indistinguishable from the twenty other tightly cropped photos of unknown people in the same category shelf, and it buys none of the keyword legibility that a type-led cover gets for free. A show two years old with four hundred monthly listeners is usually better served by a cover that says what the show is about in two readable words than by a picture of a host that nobody has met yet.

Where the art and the title fight each other

A cover and a show title are solving two different retrieval problems, and podcasters routinely make them do the same job twice.

The title field is what search inside the apps reads. It is also the field that gets cut off in a narrow app row, so the front of the title has to carry the subject. The cover is what the eye reads. Setting the full title in full on the artwork, including the subtitle that already lives in the metadata, spends the entire canvas on text that the app is printing underneath the image anyway.

The split that works: let the title field hold the searchable words, and let the cover hold the shortest form of the name that a listener could recall and say out loud. A show called "The Mid-Career GPS Podcast: Job Search, Promotion and Career Clarity" wants the long version in metadata and something closer to "Mid-Career GPS" on the art.

Keyword work in the title field is a separate exercise with its own rules, and overloading that field has its own consequences inside the apps. That side of the work is covered in our podcast SEO guide, and the Google-facing half of it in how to get your podcast on Google.

Three cases for a different image on one episode

Apple supports per-episode artwork, and Spotify lets creators customize episode covers too. Most shows should skip it. A different image on every episode erases the one benefit a consistent cover earns over a hundred publishes, which is that a returning listener recognizes the show in a feed without reading anything.

Episode art earns its place in three situations. A named guest whose face or logo carries weight with the audience. A series inside the show, where a visual marker helps someone find all six parts. A one-off episode being promoted off-platform, where the image needs to carry a hook that the permanent cover does not. Outside those cases, the same cover on every episode is the stronger choice.

How to test a cover without a split test

Podcast apps do not offer A/B testing on artwork. Impressions and tap data at the directory level are not available to podcasters either, so the honest answer is that nobody can cleanly measure a cover change against downloads. What a podcaster can do is test the decision before it ships.

Three methods cost nothing. Put the candidate covers in a screenshot of a real category shelf, at real size, surrounded by actual competitor art from the same category, and ask five people in the target audience which show they would tap and what they think it is about. The second question matters more than the first. Run the same candidates as image variants in a small paid social test, where thumbnail click-through is at least a measurable signal, even though the context is not a podcast app. And look at the covers of the five shows ranking above the podcast in its own category, not to copy them, but to find the visual slot nobody in that category is occupying.

When a cover does change, note the date somewhere permanent. Downloads move for a dozen reasons at once, and without a dated record, a podcaster will end up crediting or blaming artwork for something an episode topic did. Our first-year growth strategy covers the discipline of keeping that kind of log.

Impressions first, conversion second

A cover converts attention that already exists. It cannot create it. A show appearing in nobody's search results, nobody's category shelf, and nobody's recommendations has an impressions problem, and a new cover applied to zero impressions returns zero additional listeners.

That is the part of the work most design advice leaves out. The art determines what share of the people who see the show decide to try it. Everything upstream determines how many people see it at all, which is a question of where the show appears outside the podcast apps: search results, YouTube, other people's audiences. Those channels are the subject of how to promote a podcast and how to get more podcast listeners, and the ratings and reviews that feed the apps' own recommendation surfaces are covered in how to get more podcast reviews.

The order to work in is boring and reliable. Fix the cover so it is legible and compliant, because that is a one-day job with a permanent payoff. Then spend the rest of the quarter on the thing that actually scales, which is getting the show in front of people who are already searching for what the episodes contain.

What we do with the episodes once the art is sorted

PodAnswer exists for the upstream half of that problem. Every episode of a podcast contains answers to questions people are typing into Google right now, and those answers are sitting in an audio file where no search engine can read them. We pull the transcript, find the question the episode genuinely answers, and publish an article built around the real excerpt, with the speaker named, the timestamp shown, and a link back to the episode.

Growth is $149 a month for four articles. Network is $349 a month for twelve, which is the plan for shows publishing weekly that want a full year of answer pages built out rather than a trickle. Both plans use the podcaster's own words, not a paraphrase, and both send the traffic to the show.

The pricing page has the full comparison, and for podcasters explains how the question selection and the transcript sourcing work before anyone pays for anything.

A legible cover gets the tap. The articles are what put the show in front of someone who would never have scrolled past it.