Demand for recommendations in Threads and Telegram: August 2026
August 2026: our Threads and Telegram searches found 377 “recommend” requests across 3 businesses (the sum of weekly counts for the month). Most requests: English schools (132), Dentistry (124) and Travel (121). This is what our searches see, not the whole market.
Any plan · no card · on day 8 you decide
“Recommend” requests for the month, by business
| Business | Requests in the month | Weeks with data | Average per week | Month archive |
|---|---|---|---|---|
| English schools | 132 requests | 2 | 66 | Numbers by week |
| Dentistry | 124 requests | 2 | 62 | Numbers by week |
| Travel | 121 requests | 4 | 30.3 | Numbers by week |
Data updated: 1 September 2026
Report methodology
- The data is the weekly Zapyt demand index. It is updated every Monday for the last full week, and this report is assembled automatically from the same numbers, with no manual edits and no model.
- The monthly figure is the sum of weekly counts. A week belongs to the month its Thursday falls in. The report appears once the month is over.
- The table lists businesses that have at least two weeks with data in the month. A business enters the index for a week if it has at least 10 requests, so the other businesses are not shown for this month.
How we count, and what we do not show
- A request is a public post in Threads or Telegram where someone asks for a recommendation of a specialist or a company, and our system judged it a request for advice. We count no other sources.
- These are the requests found by our clients' searches in these lines of business, so this is a sample, not the whole market. We count requests (posts), not unique people.
- We show a business and its examples for the last full week once it has at least 10 requests. When there are no numbers, we cannot tell why: fewer than 10 requests or data not in yet. The trend chart shows earlier weeks as they were counted.
- Examples are paraphrased by a model: names and details that could identify the author are removed. An automatic filter then drops anything with an @, a link, an email, a phone number or a mention of a messaging app. Intent phrasings are typical phrases a model picks from a sample.
- A week is a Kyiv week, Monday to Sunday. The data refreshes every Monday for the last full week.
- The number depends on which searches we run in a line of business, so a change can reflect a change in our searches, not only in demand.
We do not show: authors, handles, links to posts, client names or verbatim texts.
See the requests in your own line of business and city
We set up the search for your business and city. In 7 days you'll see real requests and draft replies in your voice.
Any plan · no card · on day 8 you decide