I asked a host last month how he found his three favourite podcasts. He thought about it for a while and said: one from a friend in a group chat, one from a clip that ended up on his phone at midnight, one from a guest he already followed. Not one of them came from browsing a chart, searching a category, or reading a directory description.
That is the whole problem in one answer. Most podcast growth advice is built around a browsing behaviour that almost nobody has. People do not go looking for podcasts. They run into them. Your job is not to be findable in a catalogue. Your job is to be run into.
I edit podcasts for a living, from Dubai, for hosts scattered across time zones I mostly see as timestamps in my inbox. I do not run anyone's marketing. But I sit inside the raw footage of a lot of shows, and I watch which ones grow and which ones plateau, and after a few hundred episodes some patterns are hard to ignore. Here is what I think is actually going on.
The chart is a scoreboard, not a door
Every few weeks someone asks me how to rank in a category on Apple or Spotify. I understand the instinct. Charts feel like the front page of a store, and stores are where you find things.
Except charts are mostly a velocity measure. They reward a burst of new subscribers in a short window, which means they reflect growth that already happened somewhere else. Ranking is the receipt, not the transaction. Chasing it directly means asking your existing audience to perform a coordinated action for a badge that brings in very little new attention.
I have watched shows hit a category top ten and see almost no lasting lift. I have watched shows with no chart position at all triple their downloads in a quarter because a single guest reposted a clip. The chart is a scoreboard. Someone else has the door.
Almost everything comes from three doors
When I look at the shows I work on that actually grow, the new listeners come from three places, and it is nearly always in this order.
- A person tells another person. Group chats, DMs, one friend saying you have to hear this. This is still the biggest one and nobody can buy it.
- A short video does the introducing. A clip on YouTube, TikTok, Instagram, or LinkedIn, where the platform is actively pushing content at strangers who did not ask for it.
- A guest's audience crosses over. Someone with their own following mentions the episode, and a slice of their people follow the trail back.
Everything else, the directory descriptions, the category tags, the cross-promo swaps, the paid promos, the newsletter blurbs, sits somewhere between marginal and rounding error for most shows under a certain size. Not worthless. Just not where I would spend my Tuesday.
Notice what all three doors have in common: they are person shaped. A friend, a face in a video, a guest. Discovery happens through people, which is why the packaging of your show matters far more than the plumbing of it.
Nobody has ever subscribed to a podcast because the RSS feed was tidy. They subscribe because thirty seconds of a human being said something they wanted more of.
Clips are not marketing, they are the show
I used to think of vertical clips as a promotional afterthought. Cut the episode, then if there is time, pull a few moments for social. That framing is backwards and I was wrong about it for longer than I would like to admit.
For a lot of shows now, the clip is the first contact and the full episode is the second. That reverses the priority. If the clip is the front door, then the clip deserves the same care as the episode, sometimes more, because a hundred times more people will see it.
Which means the clip has to survive without context. No ramp up, no host introducing the guest, no polite question setup. Start at the claim. If the first sentence is not the interesting part, the clip is already dead. I look for moments where someone says something slightly risky, or answers a question in a way that contradicts the obvious answer. Agreement does not travel. Friction does.
A few things I have learned cutting them:
- Thirty to fifty seconds beats ninety almost every time.
- The strongest clips usually come from the last third of a recording, once people have forgotten they are being filmed.
- Reaction shots matter more than you think. A guest laughing at their own answer sells the moment better than any caption.
- If you have to explain the setup, pick a different moment.
We do clips as part of the workflow for exactly this reason, and vertical clips are $25 each as an add-on. But the real point is not who cuts them. It is that they stop being leftovers and start being the thing you plan the episode around.
Video is where the strangers are, and that is the whole argument
Audio apps are subscription machines. They are excellent at delivering an episode to someone who already decided they like you. They are terrible at introducing you to anyone new, because there is almost no recommendation surface pointing strangers at unfamiliar shows.
Video platforms are the opposite. Their entire business is pushing content at people who never asked for it. That asymmetry is the single most important fact about podcast growth right now, and it is why I keep telling hosts that the video version is not a bonus deliverable.
Which platform is a separate and genuinely contested question, and I wrote out my thinking on choosing between YouTube and Spotify for the video version rather than repeat it here. Short version: they reward different things and you should not treat them as one upload.
The part hosts underrate is packaging. A full episode on YouTube lives or dies on its title and thumbnail before a single frame plays. I am an editor, not a thumbnail designer, which is why our sister studio ClickTheory handles that side, eight thumbnails for $250. I mention it because I have seen the same episode, the same cut, get wildly different numbers on a re-upload with different packaging. The edit was not the variable.
Search still works, just not the way people sell it
I am generally sceptical of podcast SEO advice, mostly because a lot of it is recycled blog tactics wearing a headset. But there is a real version of it, and it is narrower and less glamorous than what gets sold.
The real version is this: episodes that answer a specific question, with a title that names that question in plain language, get found by people typing that question into YouTube or Google months later. That is a slow, compounding trickle rather than a spike, and it only works if your titles stop being inside jokes. I went deeper on the parts of podcast SEO that genuinely earn their keep if you want the detail.
Show notes and chapters help here more than most hosts expect, partly for search and partly because chapters change how people watch. Someone who can jump straight to the section they care about is far more likely to stay than someone facing an unmarked hour. We do notes and chapters for $35 an episode, which is roughly the least interesting sentence in this post and one of the highest leverage ones.
Consistency beats cleverness, but only if the tape is good
Here is the uncomfortable part. Most shows that do not grow do not have a distribution problem. They have a listenability problem, and distribution is the more fun thing to blame.
If the audio is uneven, if one camera is greener than the other, if the first four minutes are throat clearing, then every marketing effort is pouring water into a bucket with a hole in it. You get the click. You do not get the second episode.
The fixable version of this is mostly craft. Getting a real answer out of a nervous guest, for one, which is why I keep arguing that the second take is usually the one worth keeping. And on the picture side, mismatched cameras are a quiet credibility tax, the kind of thing viewers never name but definitely feel. If you shoot with more than one body, getting cameras that do not match to look like one system is worth the effort, and it is a large part of what podcast color grading is actually for.
Every episode we touch is watched end to end by a human editor. Nothing is auto-cut by silence detection, because silence detection has no opinion about which pause is dead air and which pause is a guest deciding whether to tell you the truth. That distinction is the entire job.
What I would actually do with a limited week
If a host had five hours a week for growth and asked me where to put them, this is my honest answer, in order:
- Two hours booking guests whose audience overlaps with yours but is not identical to yours. Guests are the closest thing to a growth cheat code that exists.
- One hour on titles and thumbnails for the full episode. Not the edit. The packaging.
- One hour choosing clip moments yourself, because you were in the room and you know which answer landed.
- One hour making sure the episode actually ships on the day it says it ships.
Notice what is not on the list. No chart pushes. No directory optimisation. No swapping promos with a show that has the same fifty listeners you do. Those are not scams, they are just where effort goes to be busy.
And the shipping line matters more than it looks. Half the growth stories I see are not clever, they are just uninterrupted. A show that lands every Tuesday for eighteen months builds something a brilliant show with a six week gap never does. That is partly why we work on a 48 hour first cut with revisions the next working day, and why hosts who batch tend to buy the 4-episode pack at 15% off or go monthly at 25%. Predictability is a growth tactic wearing overalls.
The thing nobody wants to hear
Discovery is downstream of being worth recommending. You can improve the odds, you can widen the surface area, you can make the clip good and the thumbnail sharp and the schedule reliable. All of that helps and I would do all of it.
But the mechanism at the bottom is still one person telling another person. Everything above it is just making that easier. When I look at the shows on my drive that grew this year, none of them found a trick. They found a format they could sustain, made the first thirty seconds worth staying for, and kept showing up long enough for the compounding to notice.
If you want a sense of how we fit into that, how it works covers the process without any of my opinions attached. And if you would rather just argue with me about clip length, I read every email at samy@pixelsamy.com.