Last month I cut an episode that was genuinely excellent. Ninety minutes, two cameras, a guest who told a story about losing a business and rebuilding it that made me stop scrubbing and just watch. Then the host sent over the title: "Episode 47: A Conversation with Guest Name."
That is not a title. That is a filename.
I have edited enough shows now to notice something uncomfortable. The gap between the shows that grow and the shows that plateau is very often not audio quality, not pacing, not guest booking. It is the eight words sitting on top of the episode. Those eight words are the only part of your work most people will ever see, and most hosts spend about forty seconds on them after spending six hours on everything else.
So here is the formula. Not a magic phrase, an actual structure I use when a host asks me to help fix a title before we publish.
The title is the whole funnel
Think about where your episode actually gets chosen. A podcast app feed where the artwork is a thumbnail the size of a stamp. A YouTube grid. A share in a group chat. A search result. In every one of those places the title is doing more persuasive work than the artwork, the description, and the trailer combined.
Your episode does not compete with other podcasts. It competes with the decision to keep scrolling.
That reframe changes what a good title looks like. You are not labelling a file for your archive. You are making a promise specific enough that someone believes it is worth forty minutes of their life. Most titles fail because they make no promise at all.
A title that describes what happened in the room is a label. A title that describes what the listener walks away with is an offer.
I am not asking anyone to write bait. Bait works once and then poisons the well, and hosts who lean on it end up with great click rates and terrible completion. The point is specificity, not exaggeration.
The four parts
When I rewrite a title with a host, I am looking for four things. Not all four always fit, but three of them should.
- A subject. The concrete thing the episode is about. Not "marketing", but "cold email that gets replies". Not "mental health", but "the week I stopped answering my phone".
- A tension. Something that suggests the obvious answer is wrong, or costly, or harder than it looks. Words like "why", "the case against", "what nobody tells you", used sparingly and honestly.
- A stake. What it costs or gains. Numbers work here, and real numbers work best. "The $40k mistake" beats "a costly mistake" every time.
- A person. Only when the person is genuinely the draw. A famous guest is a stake. An unknown guest in the title is dead weight.
A title like "Why I Fired My Best Salesperson" has a subject, a tension, and a stake in six words. "Episode 112: Sales, Hiring and Team Culture" has a filing system.
Where hosts get this wrong
The most common failure I see is the guest name lead. "Sarah Whoever on Building Teams" reads as a title only to people who already know Sarah. For everyone else it is a name they must evaluate before they get to the substance, and evaluating a stranger's name is friction. Put the idea first and the name second, if at all.
The second failure is the topic soup. Three subjects joined by commas, because the conversation covered three things and it felt dishonest to pick one. It is not dishonest. Pick the strongest ten minutes and title the episode after it. The other two subjects are still in there.
The third is the numbered prefix. "Ep. 214" burns the first characters of a title that gets truncated on nearly every surface. Episode numbers belong in the metadata, not in the visible line. I have watched hosts lose the entire back half of a good title to a truncation because a number was hogging the front.
And the fourth, the sneaky one: titling for the guest rather than the listener. Hosts do this because they will send the episode to the guest and want them to feel flattered. Understandable. Also expensive. The guest is one listener.
Titles and retention are the same problem
Here is where this stops being a marketing topic and starts being an editing topic, which is where I live.
A title sets an expectation. The first ninety seconds of your episode either pays that expectation off or breaks it. When the title promises "why I fired my best salesperson" and the episode opens with four minutes of weather chat and a sponsor read, the listener who came for that promise is gone before the story starts. That is not a title problem or an edit problem. It is a mismatch, and mismatches are the single biggest source of the drop-offs I see when hosts share their analytics with me. I wrote more about that in the piece on how pacing decides retention, because the fix is usually structural.
So I have a rule now. Write the title before the final edit, not after. If the title is decided first, the cold open can be built to deliver it. We move the line that justifies the title into the first thirty seconds. The episode then keeps the promise it made in the feed, and the retention curve looks completely different.
This is also why titles and trailers should be written in the same sitting. They are the same argument at two lengths.
The title has to survive the crop
Practical constraints, because they matter more than theory.
- Apple Podcasts and Spotify will truncate visibly somewhere around 40 to 60 characters depending on device and view. Assume your title dies at 45.
- YouTube shows roughly 60 characters in most grids and search views.
- Front-load the hook. The first five words carry the click. Everything after is confirmation.
- Say it out loud. If you would not say the title to a friend in a bar, it is written in press release language and nobody will tap it.
- One idea. If you need an "and", you probably need two episodes.
I test this in the crudest possible way. I put the title in a phone note, screenshot it, and look at it at arm's length. If the meaning does not survive that, it does not survive a feed.
The packaging layer nobody budgets for
If your show has any video presence, and most do now, the title is only half of it. The thumbnail is the other half, and the two need to argue for the same thing. A title with a sharp tension and a thumbnail of two people smiling at a table cancel each other out.
This is the one place I will point outside my own studio. Our sister studio ClickTheory does the thumbnail side, eight for $250, and the reason I mention it is that hosts who treat titles and thumbnails as one job get results that hosts who treat them as two separate afterthoughts do not. Same for vertical clips, where the on-screen hook text is really just your title compressed to five words. If the title is weak the clip has nothing to build on.
And if you are considering renaming the whole show, not just an episode, the same logic applies at a bigger scale, with more risk attached. I put down what I have seen work and fail in rebranding a podcast mid-run.
What I would do if I ran your show
I would write five titles per episode and throw away four. That sounds excessive until you try it. The first title you write is almost always a description. The third is usually the one with a point of view. The fifth is often too clever and should be discarded, but writing it teaches you where the ceiling is.
I would stop titling episodes at the moment of upload, when I am tired and just want it live. Title on a different day than you edit.
I would publish less and title harder. A show that ships twice a month with real titles beats a show that ships weekly with filenames, which is a case I have made at length in the argument against weekly episodes and still believe.
And I would be honest about which parts of this a tool can do for you. Automated tools will happily generate a title from your transcript, and it will be grammatically fine and completely inert, because it is summarising rather than choosing. Choosing requires knowing which ten minutes of the conversation actually landed, which requires having watched the whole thing. That is the same reason I do not trust auto-cutting, and I laid out the whole comparison in Descript versus a human editor.
The short version
A good title is a specific promise, front-loaded, in the words you would actually use, about one idea, that your first ninety seconds can pay off.
That is it. No formula survives contact with every episode, and occasionally the right title is just a strange sentence the guest said that you cannot stop thinking about. Trust that instinct when it shows up. It is rarer than you think.
We watch every episode we cut all the way through, which means by the time we deliver a first cut we usually have three or four title candidates pulled from the transcript, and hosts are welcome to ask for them. It costs nothing and it takes about five minutes. If you want that as part of your workflow, or you just want to argue with me about a title you are stuck on, I am at samy@pixelsamy.com.