You shipped something good on Tuesday. You wrote a changelog entry. Eleven people read it, nine of them were already customers, and by Friday the release effectively never happened. This is the most common form of wasted work at an early-stage company: the building is fine, the announcing is nonexistent.
A changelog is a receipt. It proves the work happened for people who already care. An announcement is a different act: it goes out to people who are not looking, and it has to earn a second of attention before it can explain anything.
A feature announcement video is a short, silent, text-led clip that makes one new capability legible in a couple of seconds to someone who is not currently thinking about your product. Make one only when an existing user would change what they do because of the release, open on the pain the feature removes rather than its name, show the actual screen as proof, and end on a single action. ShipTeaser generates that clip from your own landing page, so announcing a release stays a same-day task instead of a production.
Not every ship deserves an announcement
The fastest way to make announcements meaningless is to make one for everything. If every bug fix gets a post, people learn to ignore your posts. Be honest about which category a release falls into.
| Type of ship | What it deserves |
|---|---|
| Bug fix, copy change, perf work | A changelog line. Nothing else. |
| Quality-of-life improvement | A changelog line plus a reply to whoever asked for it. |
| New capability an existing user would switch behaviour for | A post, an email, and a video. |
| Something that changes who the product is for | A full launch moment. Treat it like a new product. |
If you can't name a person who will change what they do because of this release, it's a changelog entry.
What a feature announcement video is for
It has one job: make a capability legible in a couple of seconds to someone who is not currently thinking about your product. Not a tour. Not a full explanation. One capability, one reason it matters, one action.
That constraint is why the format tends to look the same across good announcements. Short, so it survives a feed. Silent-first, because most people see it muted and never turn sound on. Text-led, so the message lands without audio. On-brand, so it reads as coming from you rather than from a template.
The structure that works
Announcements fail when they open with the feature name. Nobody is emotionally invested in your feature name. Open with the thing the feature removes.
- The old pain, stated plainly. "Exports took four clicks and a spreadsheet."
- The change, in one sentence. "Now it's one button."
- The proof. The actual screen, the actual number, the actual before-and-after.
- The action. Where to go, what to click, what it costs.
That is the same spine as a launch teaser, which is not a coincidence: a feature announcement is a launch at smaller scale. The five-scene shape behind it is covered in how to make a product launch video.
Where the announcement actually goes
The video is not the campaign. Founders make one asset, post it once, and call it distribution. A release deserves a short sequence instead, and none of these steps take long:
- Your main social account: native upload, not a link out, with the video first and the explanation in the replies.
- The customers who asked for it: a direct reply to the original request beats any broadcast. This is the highest-conversion message you will send all week.
- The product email: one release per email, with the video near the top rather than buried under three paragraphs of preamble.
- In-app: a small note where the feature lives, for people who never read email.
- The landing page: if the capability changes what you sell, the page has to change too, not just the changelog.
The last one gets skipped constantly. A feature significant enough to announce is usually significant enough to belong on the page, and a video that lives on the page keeps working long after the post has scrolled away. Landing page video covers where it should sit.
Make it repeatable or it will not survive
The real problem with feature announcements is not quality, it is cadence. Anyone can produce a good one when there is time. The hard part is producing one for every meaningful ship, indefinitely, while also building the product.
So judge your process by the boring version of the question: can you do this again in three weeks, on a Tuesday, when nothing is on fire but nothing is calm either? If the answer requires a designer, a briefing round and an editing session, the answer is no, and your announcements will quietly stop after the second one.
A good announcement you can only make once a quarter is worse than a decent one you can make every time you ship.
That is the case for generating the asset from your own page rather than assembling it by hand. Your landing page already carries the current copy, the current screenshots and the current brand, so it can act as the source instead of a folder of assets you have to re-gather each time. The cadence argument in full is in video content for startups.
FAQ
How long should a feature announcement video be?
Between 10 and 20 seconds, and silent. A feature announcement carries one capability, and one capability does not need more than that. Anything longer starts explaining, which is what the changelog and the docs are for. ShipTeaser produces a 15-second, 1080p silent motion-graphics teaser, which is the length a feed will actually watch to the end.
Which releases deserve a video and which do not?
Make a video when an existing user would change their behaviour because of the release. Bug fixes, copy changes and performance work belong in a changelog line and nowhere else. Quality-of-life improvements deserve a changelog line plus a reply to whoever requested them. If you cannot name one person who will do something differently, you do not have an announcement.
How should a feature announcement video open?
Open on the pain the feature removes, not on the feature name. Nobody is invested in the name of something they have never used, so "exports took four clicks and a spreadsheet" earns attention that "introducing One-Click Export" does not. State the old pain, then the change in one sentence, then show the actual screen, then give one action.
Do you need a designer for a feature announcement video?
No, and depending on one is why most founders stop after their second announcement. The bottleneck is cadence rather than quality: a process that needs a brief, a designer and an editing round cannot run every time you ship. ShipTeaser builds the teaser from your landing page, which already holds your current copy, screenshots and brand, so the same asset can be produced on the day of every release.
Where should the announcement video be posted?
Native upload on your main social account first, then a direct reply to the customers who requested the feature, then the product email with the video near the top, then an in-app note where the feature lives. If the capability changes what you sell, put the video on the landing page too, where it keeps working long after the post has scrolled away.