I Made a Whole Podcast Just to Find Out If Titles Matter
Brian Stever
2026 · Python, LaTeX, Acast + Apple metrics, observational study
Abstract. After years of making podcasts through Snack Labs, I'd absorbed a lot of industry wisdom about what makes a show successful: storytelling, hooks, trailers, social rollout. Most of it is probably fine advice. But I kept noticing that the shows we produced with the most deliberate metadata strategy tended to have longer discovery tails than the ones with better production value. So I did what any reasonable person would do: I made an entire podcast just to test the theory. Moving to Canada was an AI-generated, search-optimized show published without a dollar of promotion. Over twelve months it reached 6,033 downloads and displayed a slower, more sustained growth curve than two professionally produced documentary shows from the same network. The paper compares all three, and the conclusion is uncomfortable in a useful way: search intent may be doing more work than branding would like to admit.
1.Research Question
I've been titling podcast episodes for a long time now, and I've participated in an embarrassing number of meetings where reasonable adults debated whether a colon or an em dash would perform better in a podcast title.
Most of those conversations ended with someone's gut feeling winning. That always bothered me, not because gut feelings are useless (they're often right), but because nobody could point to actual evidence. Podcast advice is full of respectable-sounding recommendations about hooks, trailers, and brand consistency, and almost none of it addresses the more primitive question: how does anyone find the episode in the first place?
My working theory was that podcast platforms increasingly behave like retrieval systems. If that's true, then metadata isn't decorative packaging applied at the end of production. It's part of the mechanism by which the work becomes discoverable. In plainer terms: titles aren't just names. They're search surfaces pretending to be names. I wanted to test that.
2.Study Design
The focal case was Moving to Canada, an AI-generated informational podcast I built from Canadian government newcomer documents. It had ten episodes, no promotion, and titles written to mirror the kinds of things a real person might type into a search bar rather than sound like tasteful podcast episodes. “How to Get a Canadian Work Permit” is not elegant titling. It is functional titling, which turns out to be a more useful category.
I set that experiment against two professionally produced, promotion-supported narrative documentaries from our network at Snack Labs, then examined download trajectories and Apple Podcasts engagement data over twelve months. That move mattered. It turned “I noticed something weird” into a sturdier argument, or at the very least a more defensible one.
Table 1. Core design choices.
| Element | Implementation |
|---|---|
| Treatment show | AI-generated, search-optimized Moving to Canada |
| Comparators | Two professionally produced, promotion-supported documentary podcasts |
| Primary outcomes | Acast downloads, Apple engagement, and episode-level performance patterns |
| Reproducibility | Data tables, figure-generation script, statistics verification script, and LaTeX manuscript |
3.Comparative Frame
The key comparison is not “AI versus humans” in the broad, annoying sense. It is much narrower and more useful than that. One show was optimized for retrieval and direct usefulness; the other two were optimized for professional storytelling and conventional promotion. The question was which shape of packaging produced more durable discovery over time.
Case A
Moving to Canada
AI-generated, search-optimized, no promotion, practical topic, ten episodes.
Case B
Professional Documentary Show
Human-produced, promotion-supported, narrative format, conventional launch spike.
Case C
Second Documentary Comparator
Same network context, professionally packaged, stronger launch support, faster decay.
The comparison is intentionally unfair in an interesting way: one show had less polish and less promotion, but much tighter alignment between title, listener intent, and information need.
4.Findings
The search-optimized show reached 6,033 downloads across twelve months without promotion and displayed a delayed but sustained growth pattern. The comparator shows behaved more like traditional launches: stronger initial spikes followed by decay. That is not shocking if you think like a search engine, but it is a strange thing to watch happen to shows you helped produce.
I also pulled Apple Podcasts engagement summaries, which mattered because “people clicked it once” is a much weaker claim than “people found it, stayed, and kept finding it.” The takeaway is not that AI audio is secretly superior. It's that tight alignment between a query, a title, and actual informational utility can compensate for a remarkable amount of aesthetic compromise.
Table 2. High-level outcomes from twelve months of observation.
| Metric | Value |
|---|---|
| Treatment show downloads | 6,033 |
| Observation window | 12 months |
| Episodes in treatment show | 10 |
| Marketing spend | $0 |
5.Reproducibility
The part I care about most is that this isn't a vibes-based argument. The project includes the monthly downloads table, engagement summaries, episode-level metrics, a script to regenerate every figure in the paper, and a separate script to recompute the summary statistics. I submitted a double-blind version for review and maintain a credited preprint. That's a satisfying amount of procedural seriousness for something that began with me side-eyeing podcast titles over coffee.
That structure changes the claim. It's no longer “I tried some SEO-ish titles and things went well.” It becomes “here are the numbers, here are the figures, here's the manuscript, and here's the precise narrow claim I'm willing to defend.” I find that more persuasive. Also more fun.
6.Reflection
I don't think this means every podcast should sound like a help article. Not every show benefits from titling itself like a search query from a stressed-out newcomer. But I do think we routinely understate how much distribution logic has crept into editorial decisions.
The broader lesson is uncomfortable and therefore probably useful: metadata is part of the product. Or, if you want the meaner version, metadata is the part of the product that platforms are willing to reward first.