ExpertsFMThis Is How You Should Actually Make Launch Videos
Y Combinator Startup Podcast · October 23, 2026
5 Selected Lessons:
heuristic · 2 reasoning_pattern · 2 priority · 1
Downloads/retention over views
Source
Y Combinator Startup Podcast — This Is How You Should Actually Make Launch Videos (October 23, 2026) · at 7:01 ▸
When this applies
when deciding which metric should drive conviction to continue investing in a product or campaign
The takeaway
Rank metrics by how much genuine user commitment they require, and let the highest-commitment metric drive strategic conviction.
Why it works
Views are a vanity signal of reach; downloads and multi-day retention indicate genuine product-market pull because they require real effort/commitment from the user.
One level up
Optimize decisions on proof of real behavior change, not on proxies for attention.
Where it stops holding
This stops working when the product has no download/install step (e.g., pure content, where views ARE the end behavior) or when retention data isn't available yet.
Detail
hierarchy:
- views (least important)
- downloads
- day-4/5 retained daily active users (most important)
rationale: 7,000 downloads and ~100 people retained daily for five days gave Farza the conviction to keep building, not the 3 million views.
In their words
I was like, How many people downloaded it? And so, after this video, I think 7,000 people downloaded it. And then I looked at who's like retained four days later. And if you ask me why I'm working on the product today, it's not because it went viral. It's because 7,000 people downloaded it.
Reframe 'launch' as continuous 'demos'
Source
Y Combinator Startup Podcast — This Is How You Should Actually Make Launch Videos (October 23, 2026) · at 3:31 ▸ · at 17:45 ▸
When this applies
when deciding the cadence and framing of any announcement about progress — product releases, research results, hires
The takeaway
Reduce the perceived size of each unit of communication so that the cadence of sharing can be continuous rather than rare and high-stakes.
Why it works
Treating every update as a 'mini launch' lowers the psychological stakes of posting, enabling high frequency, which compounds into audience-building over time; a single 'big launch' framing creates pressure and infrequency.
One level up
Frequency and compounding repetition often outperform singular, high-stakes bets in building distribution.
Where it stops holding
This stops working when your audience or market genuinely expects and rewards rare, high-ceremony announcements (e.g., regulated product approvals, major funding milestones).
Detail
approach: Call every shipped feature or idea a 'demo' and post about it, rather than reserving communication for a big coordinated 'launch'.
signals:
- asking 'what are we launching this week' as a running company question
- working backward from the caption/hook before building the feature
rationale: The more demos shipped, the more shots on goal for finding what resonates, and it normalizes sharing as routine rather than a rare event.
In their words
And so yeah, I think that now we keep calling these launches. And I think that that is what these are. These are mini launches all the time.
Behind-the-scenes process content as a mini-story
Source
Y Combinator Startup Podcast — This Is How You Should Actually Make Launch Videos (October 23, 2026) · at 22:43 ▸ · at 25:50 ▸
When this applies
when deciding what aspect of your work to make public to build audience affinity
The takeaway
Showing authentic process, not just polished outcomes, creates narrative pull because audiences crave insider access and a story to follow, even absent a conventional 'hook'.
Why it works
Most people have no visibility into what actually happens inside a company; showing the mundane, authentic process (not polished conclusions) satisfies curiosity and builds parasocial investment/rooting interest, functioning as a story even without a traditional hook.
One level up
Transparency about ordinary work can substitute for traditional marketing hooks by satisfying a universal curiosity about how things really get made.
Where it stops holding
This stops working when the process itself is proprietary/sensitive, legally restricted, or genuinely uninteresting even in authentic form.
Detail
approach: Post simple 'what we shipped this week' videos with no hook or story framing, just showing the team and the work.
signals:
- audience loves it even without going viral
- some even go viral despite zero traditional hook
rationale: What we are doing on a day-to-day basis at our companies is so far-fetched from what most people think is actually happening at a company... it gets them excited.
In their words
So, there's no hook, there's no attention, there's no story, it's just we're a group of people and we want to tell you what we're doing.
Casual beats polished on algorithmic platforms
Source
Y Combinator Startup Podcast — This Is How You Should Actually Make Launch Videos (October 23, 2026) · at 0:00 ▸
When this applies
when choosing between a high-production approach and a cheap, authentic approach for any channel whose distribution is algorithmically driven rather than gatekept
The takeaway
When the distribution mechanism shifts from gatekeepers to algorithms, optimize for native authenticity and frequency over polish, because polish was only a signal gatekeepers valued.
Why it works
Algorithmic feeds reward content that stops a scroll and feels native to the platform; old media's value came from gatekeepers (press, curation) which have been replaced by algorithms that don't care about production value.
One level up
Match your production investment to what the actual distribution mechanism rewards, not to what the old mechanism used to reward.
Where it stops holding
This stops working when your channel is still gatekept by humans (e.g., enterprise sales decks, press that requires polish) rather than algorithmically distributed.
Detail
rule: A casual phone video saying 'hi, this is my product' will likely outperform a $20,000 produced video in engagement and traffic on algorithmic platforms.
threshold: Phone + laptop is enough; no PR/press required
rationale: Old media (TechCrunch, NYT) is gone as the primary distribution gate; algorithms now serve content directly to interested audiences regardless of polish.
In their words
But now all the old media is kind of just gone. And so now what's working is like a video where I put, I say hi and I talk about my product is probably going to perform way better than like a $20,000 produced video by the New York that I post with the New York Times in terms of actual engagement and traffic and like other things.
Audio quality is the one non-negotiable production value
Source
Y Combinator Startup Podcast — This Is How You Should Actually Make Launch Videos (October 23, 2026) · at 40:49 ▸
When this applies
when allocating scarce production resources across the quality dimensions of any piece of content (visual, audio, pacing, resolution)
The takeaway
Identify the single quality dimension that is a hard gate for audience retention and invest there first, even if everything else stays cheap.
Why it works
Viewers can mentally tolerate poor visual quality if the content is engaging and audible, but bad audio creates immediate cognitive friction that causes abandonment regardless of content quality.
One level up
Not all quality dimensions are equally load-bearing; find the one whose failure causes total abandonment and protect it above all else.
Where it stops holding
This stops working for purely visual mediums (e.g., silent demo GIFs) where audio isn't part of the experience.
Detail
rule: Invest in a cheap dedicated mic (e.g., DJI mic) over any other gear upgrade; never rely on laptop or AirPod mics.
threshold: Even a 720p video in a 'weird basement room' retains viewers if audio is crisp and content is interesting.
rationale: The most important thing in any video you make, doesn't matter if it's a shitty loom or a shitty little video or your iPhone, is audio quality. If your audio is bad, I'm bouncing.
In their words
The most important thing, this is actually so important. I can't believe I missed this. The most important thing in any video you make, doesn't matter if it's a shitty loom or a shitty like little video or your iPhone, is audio quality. If your audio is bad, I'm bouncing.