Maintenance
Why platform trends converge on the same handful of shapes
Why social platforms keep converging on three shapes, the feed, short vertical video and ephemeral stories, with dated examples, three durability tests and a way to read a trend claim.
The word "trend" covers three different things, and writers switch between them mid-sentence, which misleads beginners. First, trends in what platforms build. Second, trends in how people use them. Third, trends in what circulates on them.
These trends move at different speeds and are evidenced differently. Of the three, the product layer is the one you can act on this week, because you can test a new feature in your own account.
What to take away
- Separate the three layers before you believe a claimproduct, behavior and content. A claim that does not say which layer it belongs to is not yet a claim.
- Platforms converge on the same shapes because they face the same constraints. Similarity is expected. It does not prove that the feature works, and it does not prove that anyone copied anyone.
- Test a trend for durability with three questions.
The three layers
| Layer | What changes | How fast | How you can see it | What you can do about it |
|---|---|---|---|---|
| Product | Surfaces, formats, tools, policies the platform ships | Months | Announcements, your own interface, a test account | Adopt, wait, or ignore, by a rule |
| Behavior | How people spend time, what they open, when, on what device | Quarters to years | Only through aggregate reporting, which is usually someone's marketing | Very little; plan around it |
| Content | What kinds of things are circulating and being copied this month | Days to weeks | Your own feed, which is a biased sample of one | Join, adapt, or sit it out |
The layers interact, and the direction of cause is the useful part. A product change can enable a content trend. A behavior shift can make a product change viable. A content trend that lasts can become a behavior. But a content trend on its own is weather. The mistake is reading a content trend as a behavior trend and reorganizing around it.
Three layers of platform trends
Product
- What changes
- Surfaces, formats, tools
- How fast
- Months
- How you see it
- Announcements, interface
- What to do
- Adopt, wait, ignore
Behavior
- What changes
- How people spend time
- How fast
- Quarters to years
- How you see it
- Aggregate reporting
- What to do
- Plan around it
Content
- What changes
- What circulates and copies
- How fast
- Days to weeks
- How you see it
- Your own feed
- What to do
- Join, adapt, sit out
The product layer is the only one with a paper trail. Keep a change register with the date observed, the platform, the feature and what changed. Separate what the platform announces from what you can verify in the product. Several changes in a row that point the same way are a more reliable trend claim than a single announcement.
The three shapes platforms keep converging on
Three product shapes recur across the major platforms.
- The algorithmic feed.A ranked stream of items chosen for the user without a search. Instagram's feed, TikTok's For You feed and YouTube's home page all work this way.
- Short vertical video.Clips shown full screen, one after another. TikTok popularized the format. Instagram Reels followed in 2020, and YouTube Shorts was announced the same year.
- Ephemeral stories.Posts that disappear after a set time. Snapchat launched Stories in 2013. Instagram added Stories in 2016, a public example of a format spreading from one platform to another.
These three shapes share a cause. Every feed-based product has too much supply, too few slots and no query from the user, and it has a fraction of a second to choose. Those constraints favor formats that are cheap to evaluate, surfaces that need no query and features that make a measured objective easier to count. Short video gives the platform a stream of clips to rank. A feed supplies the next item without a search.
The shared constraint explains the similarity. It does not rule out copying. Instagram's Stories launch is a documented case of one product taking a format from another, so convergence and copying can both be true.
Three durability tests
A trend claim that survives all three tests is more likely to describe something real. Most claims in the feed fail at least one of them.
Three durability tests for a trend
- Does it reduce a cost for someone?
- Does it serve the platform's countable objective?
- Would it survive without promotion?
- Does it reduce a cost for someone?Durable shifts make something cheaper to make, to consume or to find. A trend that only makes something more fashionable has no floor under it.
- Does it serve the platform's countable objective?If the trend moves the measure the platform optimizes, the platform keeps feeding it. If it delights people without moving that measure, it loses distribution when the measure dips.
- Would it survive without promotion?Anything a platform is pushing looks like a trend, because the push makes it visible. Ask what the trend would look like if the push stopped. If you cannot imagine it existing, it is a campaign. Paid promotion and platform recommendation both make a format look popular, so a recommender's drift can look like a trend of its own from outside.
These tests map onto an older observation about how new things spread. Everett Rogers's diffusion of innovations model describes an adoption curve: a small group tries early, a larger group follows once the thing is proven, and the rest arrive late or never. Where you sit on that curve decides whether joining is early, on time or late. By the time a trend is written up, the early group has usually moved on.
Reading a trend claim
Before accepting a claim, ask who made it and what they sell.
- A platformsays a surface is growing. It benefits from your investment in that surface. Read the claim as an intention, not a measurement.
- A tool vendorsays a behavior is shifting. It benefits from you needing the tool. Ask for the sample behind the claim.
- An agency or consultantsays a content trend is the future. It benefits from your uncertainty. Ask what they predicted last year.
- Your own numberssay something changed. They benefit nobody. Check the inputs that could have moved them before looking outside.
The attention pattern around new technologies, with a rise, a disappointment and a quieter, smaller settled role, is described by Gartner's hype cycle model. It is a description, not a law, but it is a useful check. If a trend is at its coverage peak, that coverage is the least reliable guide to its eventual size.
What a beginner should actually do
- Keep a product-layer record. Three months of it teaches more than a year of trend articles.
- Pick two behavior claims a year to plan around, chosen on evidence you can check, not on volume of coverage.
- Treat content trends as optional. Join one when the cost is low and it fits what you already make, and skip the rest without guilt.
- Re-read your own record every quarter and cross out the trends that did not happen. The crossing out is the education.
Common questions
Everyone says a certain format is taking over. Should I switch?
Run the three tests. If it fails "would it survive without promotion," you are being recruited, not informed. If it passes all three and the cost is low, try it alongside what you do now rather than instead of it. Watch whether people keep using it after the promotion fades.
How do I know a behavior trend is real when I cannot see the data?
You mostly cannot. Look for the same claim from parties with opposed interests, look for it in your own numbers, and hold it loosely until both agree.
Why do platforms keep changing things that worked?
Because a change is measured against the platform's objective, not yours. A feature that worked for you and did nothing for that objective has no defender inside the company.
Is convergence bad for people who publish?
It lowers the cost of being on several platforms at once, and it raises the cost of standing out on any one of them. Whether that nets out well depends on which platforms you were relying on.
In this guide
- Platform trends guide: Track updates without the noisePlatform trends guide to tracking algorithm changes, emerging trends, and outages with named sources, discovery tools, and a weekly half-hour routine.
- Trend-tracking platforms compared: seven habits and how seven tools scoreTrend-tracking platforms differ in what they measure and how fast they report it. The good ones name their source, show their method and let you export the data.
- Platform trend research is an argument, not a measurementPlatform trends research examined as an argument: three premises, what the loop pushes toward, the forces pushing back, and what none of it can give you.
