I’m 28, and near the end of 2025 I posted my first TikTok video. Until that point I had treated short-form video mostly as something other people used for distribution. I wanted to stop theorizing about it and run a small experiment of my own.
This was not a launch campaign and I had no evidence that TikTok would become a useful acquisition channel. I simply wanted to take one real product flow, turn it into a short vertical demo, publish the same creative across multiple platforms, and measure what happened.
That sounds simple. The surprising part was how many product and measurement decisions appeared before I even pressed publish.
A product demo is a separate interface
For the clip, I chose a representative flow from an older web side project and recorded it as it actually worked. I wanted a real interface with real states and real transitions, not a mockup created only for the video.
That decision exposed the first problem immediately: an interface designed for interaction is not automatically an interface designed for observation. A person using the product can pause, read, scroll, hover, retry, and decide where to look. A viewer watching a short video cannot.
Once I saw the recording inside a phone-shaped frame, I stopped thinking of the video as “a screen capture.” It was another product surface with its own hierarchy, pacing, constraints, and failure modes.
The 9:16 frame changes what is important
The most immediate lesson was not about the underlying code. It was presentation.
Trying to show a desktop web interface inside a vertical mobile video exposes every weak assumption in the layout. Text that feels comfortably readable on a monitor can become tiny. Secondary controls compete with the primary action. A wide information hierarchy gets compressed into a narrow column. Cursor movement that feels natural during use can look slow and confusing in a recording.
The solution is not necessarily to redesign the product for TikTok. The better approach is usually to redesign the demonstration: choose one visual idea, crop aggressively, increase the scale of the important state, remove dead time, and make the action understandable even with the sound off.
This distinction matters because otherwise it is easy to draw the wrong conclusion. A bad vertical demo can make a good desktop product look bad. That is a presentation failure, not automatically a product failure.
What I actually tested
I cross-posted the result to Instagram Reels, TikTok, and YouTube Shorts. The experiment was intentionally narrow: use the same basic creative on three large short-form platforms and see whether any of them produced measurable downstream behavior.
I did not have the answer when I published the article. Posting on three platforms is an action, not validation. Views are not visits. Visits are not activated users. And activated users are not necessarily retained users.
That sounds obvious, but short-form platforms make vanity metrics unusually seductive. A video can look successful in the app while producing almost nothing useful outside it.
The measurement funnel I care about
I now think about this kind of experiment as a chain rather than one metric:
- Reach: how many people were shown the video?
- Attention: did they stay long enough to understand the core idea?
- Intent: did they visit the profile, open a link, search for the product, or take another measurable step?
- Arrival: how many sessions actually reached the site or product?
- Quality: did those visitors engage with the thing I wanted them to try?
A high number near the top of the funnel cannot compensate for zero movement underneath it. If a clip gets reach but nobody expresses intent, the distribution may be fine while the creative or proposition is weak. If people click but immediately leave, the problem may be the landing experience rather than the video.
Tracking matters before the first post
If I run the experiment again, I want the measurement plan ready before publishing. At minimum I would keep platform-specific campaign parameters, use the same landing destination, record the publishing time, and define the conversion event before seeing the results.
I would also write down the hypothesis in advance. Something as simple as “a vertical demo of one clear workflow will produce measurable visits from at least one platform” is better than looking at the numbers afterward and inventing a story that fits them.
For small experiments, perfect attribution is unrealistic. People may see a video and visit later through search, another device, or a direct URL. But imperfect attribution is still better than no measurement discipline at all.
What I learned about making the creative
- One clip should communicate one idea. A complete product tour is usually too much for a short vertical video.
- Show the result early. If the interesting state appears only after ten seconds of setup, many viewers will never reach it.
- Design for silent viewing. The core action should remain understandable without narration.
- Use scale deliberately. The viewer should not have to inspect tiny UI to understand what changed.
- Remove interaction that matters only to the operator. Loading, cursor travel, menu hunting, and setup may be authentic, but authenticity is not the same as clarity.
- Do not confuse editing with deception. Cutting dead time and focusing the frame is not falsifying the product as long as the demonstrated behavior is real.
Publishing the same video on three platforms is useful—but limited
Cross-posting gives a cheap first comparison because the creative is roughly controlled. If one platform produces substantially more useful downstream behavior, that is a signal worth investigating.
It is not a clean scientific benchmark, though. The platforms have different audiences, recommendation systems, compression, caption conventions, profile layouts, and linking friction. Even the same file is not necessarily being tested under equivalent conditions.
So I would treat the first cross-post as reconnaissance. If a platform shows promise, the next experiment should adapt the hook, pacing, text placement, caption, and call to action specifically for that platform instead of endlessly reposting the same generic clip.
The result was still unknown, and that was fine
My first TikTok did not prove that TikTok, Reels, or Shorts would become a meaningful traffic source for me. At the time of writing, I did not have enough evidence to make that claim.
What it gave me was a concrete experiment: choose one representative product flow, adapt the presentation for vertical video, publish the same creative on three platforms, keep the tracking consistent, and wait for enough data before deciding whether the channel was useful.
That is already more valuable than an opinion about whether short-form video “works.” Channels do not work in the abstract. A specific creative, for a specific audience, with a specific proposition and landing experience either produces useful behavior or it does not. The interesting part is measuring that chain honestly.