Playbooks · 4 min read
Trend Detection We Do Not Have, and the Signal You Already Do
There is no trend detection in this product. Nothing watches hashtags, nothing tracks trending audio, nothing reads the explore page, nothing tells you what your niche is posting about this week. There is no content side to this software at all.
And the tools that do claim it on Instagram are working from public pages rather than from the official API, which is scraping, and scraping breaks Meta's terms. That is the end of that route rather than a detour around it.
What is worth the rest of the page is a different kind of trend, one that arrives in a place you are already permitted to look, and that most accounts read as noise because they are only ever looking at one message at a time.
A trend is a derivative, which is why levels mislead
The reason "top questions in my inbox" is not a trend: it is a level. Price is always the top question. It was the top question last year and it will be next year, and knowing it tells you nothing you can act on.
A trend is a change in rate. The question that was not being asked in March and is being asked eleven times in June. That one is information, because something caused it — a new competitor, a policy change, a video somebody else made about your category, a rumour.
The useful signal is never the biggest number. It is the number that moved.
The three places a rising question shows up
You do not need a detector for this. You need to notice three things you are probably already experiencing as mild irritation.
- The keyword you had to add. Every time you open the builder to add a word to a trigger list, something changed in how people talk to you. That is the single cleanest trend signal in the product and it is generated by your own maintenance work. Keep a note of the date you added each one.
- The replies you find yourself typing by hand. Automation covers your common cases. So the questions you are personally answering are, by definition, the ones your templates did not anticipate. If you type the same unanticipated answer three times in a week, that is a trend with a sample size.
- The lead notes that start rhyming. Four rows in a row saying some version of "asked whether the price includes GST" is a real change in what buyers are worried about.
Read the words, not a summary of them
There is no analytics product here — no chart of question volume, no topic clustering, nothing that will surface any of the above on your behalf. That is a real limitation and it means the practice is manual.
It is also, for most account sizes, better than the automated version would be. Reading fifty comments as raw strings takes about ten minutes and gives you exact phrasing, which is the thing you actually need. A chart telling you that shipping enquiries rose does not tell you that people started writing kitne din lagenge instead of delivery time, and it is the second fact that changes your keyword list.
Leads export as CSV. A month of notes in a spreadsheet, read end to end on a Sunday, is the whole reporting layer and it costs nothing.
Two ways to be wrong about a trend
Mistaking your own posting for the world. You made a reel about wedding orders and now everybody is asking about wedding orders. That is not a market shift, it is your own content coming back. Check what you published before you conclude anything. Your content decides what people ask you is the fuller version of that loop.
Mistaking a spike for a direction. One influencer mentions you and forty people ask the same thing in a day. Answer them well, add the keyword, and wait a fortnight before you rebuild your product line around it.
What a rising question is worth doing about
Three moves, in increasing cost.
Add the words to an existing trigger, so the next person asking gets answered instantly. That is free and takes a minute.
Write a template for the answer, because if you are typing it by hand repeatedly it deserves to be one of your free unlimited replies rather than a generated one.
And, if it keeps rising, put the answer in a caption or a pinned post, so the question stops arriving at all. The best outcome of noticing a trend in your inbox is that the inbox gets quieter.
None of that requires knowing what is trending on the platform. It requires knowing what changed in the small number of people who are already talking to you, which is the only audience whose behaviour you can both observe and affect. If you want the version of this aimed at what buyers compare you against, that is a separate read.
