Trends are cheap to spot and expensive to catch late. By the time a format shows up in your feed three times, other brands have already shot their version. A trend radar is a weekly routine that tells you which patterns in your niche are early enough to act on and which have passed.
It is a routine, not a dashboard. The output is one page, reviewed once a week, with a decision beside every row.
Four lenses#
Look at four kinds of trend separately. They behave differently, and mixing them blurs the picture.
| Lens | What you track | Example |
|---|---|---|
| Formats | Repeatable video structures | “Three things I stopped buying” |
| Sounds | Audio that creators in your niche reuse | A voiceover meme, a song snippet |
| Topics and claims | What people are talking about | An ingredient, a routine, a debate |
| Products | Items showing up across creators | A gadget, a flavor, a bundle |
Formats and sounds tend to move fast. Topics and products often last longer. That is a working assumption to test on your own niche, since every category differs.
Four states#
Give every item one state each week. These definitions are starting rules, so adjust the thresholds to the size of your niche.
| State | Signs | Action |
|---|---|---|
| Emerging | Fewer than 5 creators, but recent videos run 3x or more their creators’ medians | Test cheaply within 72 hours, if it fits the brand |
| Rising | 5 to 25 creators, ratios above 2 holding for a week, large accounts mostly absent | Produce your best version this week |
| Peaking | More than 25 creators, brands and large accounts joining, ratios falling | Join only with a clear twist, or skip |
| Fading | New videos fall below creators’ medians, imitation outnumbers originals | Retire it from the content plan |
A state is a judgment based on three things you can check: how many distinct creators use it, how its recent videos compare with their creators’ norms, and how crowded it is. The comparison against each creator’s norm matters because it separates a trend from a creator who always gets views.
The 45-minute weekly routine#
Minutes 0 to 10: Pull. Run one research pass over the last seven days, across TikTok, Instagram Reels and YouTube Shorts. Use a prompt that asks for the four lenses explicitly (an example follows).
Minutes 10 to 25: Label. Give each candidate a lens and a state. For each, note the evidence: two or three links and the creator count. If you cannot find evidence, the row does not make the page.
Minutes 25 to 35: Compare. Open last week’s radar. Mark which items changed state. The movement matters more than the list. An item that went from Emerging to Rising is your best action. An item that stayed Rising for four weeks is probably a staple.
Minutes 35 to 45: Decide. Choose one item to test, one to watch and one to retire. Assign an owner and a date. More than three decisions and nothing gets done.
The one-page template#
Copy this into a doc or a sheet.
| Item | Lens | Last week | This week | Evidence (links) | Brand fit (0 to 2) | Action | Owner |
|---|---|---|---|---|---|---|---|
Three columns carry the weight. “Evidence” stops guesswork. “Brand fit” stops you from chasing trends you cannot do well. “Action” forces a decision.
A fit filter for every candidate#
Ask three questions before anything goes into production:
- Audience fit. Would your buyer actually watch this?
- Proof fit. Can your product show something real inside the format?
- Speed. Can you shoot, edit and post in 72 hours?
Two out of three is a maybe. One or none is a skip. A trend that does not fit your product turns into content that gets views and no sales.
Worked example (illustrative)#
A skincare brand runs its radar in a week of late September. The rows below are invented to show the format.
| Item | Lens | Last week | This week | Action |
|---|---|---|---|---|
| “Barrier repair” explainers | Topic | Emerging | Rising | Produce a dermatologist-style explainer with a patch test |
| “Night routine in 3 steps” | Format | Not seen | Emerging | Test one version, three openings |
| Skin-fasting challenge | Topic | Rising | Peaking | Skip, because the claims are hard to support |
| Glass-skin routine | Topic | Peaking | Fading | Retire from the plan |
The brand tests one new format, produces one topic video and drops one stale theme. That is a good week.
Prompt for the pull#
A Deep Search prompt for the first step:
For the sensitive-skin skincare niche, find what changed in the last 7 days on TikTok, Reels and Shorts. List formats, audio, topics and products that appear in at least three unrelated creators’ videos. For each, give the creator count, how recent videos compare with their creators’ median views, and links. Flag anything new since last week.
ContHunt returns the analyzed set with links, so every row on your radar has evidence behind it. The public Viral Now boards are a good second source. They show top viral videos by niche, which helps you spot a trend from an adjacent niche before it arrives in yours (see Viral Now). Save each week’s evidence to a board so you can compare it with the next one.
Common failures#
- Treating the radar as a list of ideas. It is a ranking with evidence. If a row has no links, delete it.
- Chasing peaking trends. By then your version is the fortieth. Peaking items need a real twist or a pass.
- Ignoring fading signals. Retiring is a decision too, and it frees capacity.
- Skipping the comparison. A single week’s list tells you what is popular. Two weeks tell you what is moving.
Use the radar alongside the winning-product scorecard if you sell physical products, and the hook taxonomy to brief whatever you choose to test.
Get started with a first radar pull on your niche this week, and fill in the template before your next content meeting.
