# Selling on Threads: what 901 posts and 893 replies showed

> Selling on Threads was measured in Replyvero's study by likes and replies, not sales: roll calls and milestones took off most often, direct offers least. The sample: 901 posts from Threads search in October 2026, including 478 selling posts (268 Russian, 210 English). In English, posts addressed to the author's own group stood out: 48% of those that took off, 14% of the rest.

Updated: 2026-10-21 · https://replyvero.com/guides/threads-selling-posts-study/

## Selling on Threads: what the study counted

This study of selling on Threads rests on three terms. A selling post is a post where the author talks about their own business, product or service, or invites others to show theirs. A roll call is a selling post that asks a named group ("small business owners", "fellow illustrators") to show what they do in the replies, usually with the author going first. A post took off if it reached at least 100 likes or at least 60 replies.

Likes and replies are the only outcomes in the data. There are no clicks, orders or revenue, so the study can't tell you what sells on Threads; it shows which selling posts people liked and answered.

## How the posts were collected and labeled

We collected the posts from Threads search results in October 2026: 32 queries in the Top tab, 16 in English and 16 in Russian, with 30 posts taken for each. After duplicates and posts without text were dropped, 901 posts remained, 431 in English and 470 in Russian.

The queries were phrases a small business owner, maker or founder would type. In English: small business owners, shop small, my small business, I built an app, just launched, my first sale, commissions open, booking now, drop your business, link in bio, my startup, client results, I help, my product, solopreneur, business owners on threads. The Russian queries were local equivalents, such as "looking for clients" and "first sales".

The labels came from MiniMax-M2.7, an AI model separate from the one that writes Replyvero's own texts. The schema covered the kind of selling post, type of first line, form, call to action, tone, who is speaking, proof and whether the post ends with a question. Of the 901 posts we collected, 478 came out as selling posts (268 Russian, 210 English), in seven kinds.

111 of the 478 selling posts took off, 61 Russian and 50 English. For each kind of post we counted the take-off rate and the median likes and replies, then compared first lines, tone, length, emoji and closing questions between the selling posts that took off and the rest.

For replies, we took the 10 most-discussed roll calls in each language and collected up to 60 replies under each. A reply's likes were ranked only against other replies under the same post, as a percentile where 0.50 is an average reply for that post. One of the 20 posts, labeled a roll call by the model, turned out to be an author showing their own naming, logo and packaging project and asking what people thought. That is a request for feedback, so its 59 replies were left out. The tables use the other 19: 427 Russian replies under 9 posts and 466 English replies under 10.

## Which selling posts take off on Threads?

Roll calls and milestones took off most often among the 478 selling posts in Replyvero's October 2026 sample. 65% of English roll calls and 38% of Russian posts about a milestone or client result reached 100 likes or 60 replies. Direct offers did worst in both languages: 4% in English and 11% in Russian.

English selling posts (210):

| Kind of selling post | Posts | Took off | Median likes / replies |
|---|---|---|---|
| Roll call: "show me your business" | 34 | 65% | 73 / 116 |
| Milestone, result or client case | 17 | 29% | 44 / 25 |
| Intro: "here's what I do" | 35 | 20% | 19 / 18 |
| Admission: the hard side of the business | 40 | 18% | 14 / 19 |
| Soft sell | 35 | 17% | 14 / 4 |
| Launch or new product | 23 | 9% | 2 / 1 |
| Direct offer | 26 | 4% | 1 / 0 |
| All English selling posts | 210 | 24% | |

Russian selling posts (268):

| Kind of selling post | Posts | Took off | Median likes / replies |
|---|---|---|---|
| Milestone, result or client case | 29 | 38% | 55 / 13 |
| Roll call: "show your business" | 49 | 29% | 18 / 29 |
| Admission: the hard side of the business | 35 | 23% | 14 / 11 |
| Intro: "here's what I do" | 41 | 22% | 15 / 20 |
| Soft sell | 60 | 20% | 12 / 9 |
| Launch or new product | 36 | 14% | 17 / 7 |
| Direct offer | 18 | 11% | 10 / 4 |
| All Russian selling posts | 268 | 23% | |

The two languages agree at the bottom and disagree at the top. Offers and launches spread least in both, and soft sells, where the product comes up inside a story, did a little better. English roll calls had a median of 116 replies, more than their 73 likes, most likely because people reply to introduce themselves. Russian milestones got the most likes (median 55) and far fewer replies, which looks like congratulations more than conversation. Rows of 17 or 18 posts only show a direction.

## What do selling posts that took off have in common?

88 characters was the median length of the English selling posts that took off, against 119 for the rest; in Russian the gap was smaller, 120 against 128. The posts that took off also ended with a question less often, and in English they were far more often written to a group the author belongs to.

| Feature | English: took off | English: rest | Russian: took off | Russian: rest |
|---|---|---|---|---|
| Median length, characters | 88 | 119 | 120 | 128 |
| Ends with a question | 4% | 9% | 5% | 8% |
| Has emoji | 46% | 39% | 48% | 47% |
| Personal tone | 38% | 48% | 48% | 37% |
| In-group tone (label "community") | 48% | 14% | 33% | 27% |
| Expert tone | 2% | 8% | 8% | 14% |
| Salesy tone | 0% | 4% | 2% | 2% |

Tone is where the languages split. In English, the in-group tone, writing to a group you belong to such as "fellow illustrators", made the difference: 48% of the selling posts that took off had it, against 14% of the rest. A personal tone on its own was less common among the English posts that took off (38%) than among the others (48%). In Russian, both tones were a little more common among the posts that took off. Taken together, personal or in-group posts made up 86% of the English posts that took off and 100 of the other 160; in Russian, 80% and 64%. Expert and salesy tones were rare among posts that took off in both languages. The remaining posts were labeled neutral or humorous.

Emoji barely moved anything, and length mattered far more in English (a 31-character gap) than in Russian (8).

> **Let Replyvero find the conversations for you.** Describe what you sell. Replyvero searches Threads for people who need it, filters posts by meaning and drafts a short reply for each one. By default you review every draft before it goes out. The first hour is free. Try it free in Telegram: https://t.me/Replyverobot?start=md_threads-selling-posts-study

## Which first lines work in selling posts?

The first line split the two languages more than any other feature. English selling posts took off most often when the first line addressed a group (18 of 40), Russian ones when it opened with a milestone or result and a number (9 of 24). Each cell shows how many selling posts with that opening took off:

| First line | English | Russian |
|---|---|---|
| Addresses a group ("small business owners…") | 18 of 40 (45%) | 10 of 33 (30%) |
| Asks a question | 5 of 21 (24%) | 2 of 28 (7%) |
| An admission | 6 of 31 (19%) | 11 of 40 (28%) |
| A general statement | 4 of 31 (13%) | 7 of 39 (18%) |
| A milestone or result with a number | 1 of 10 (10%) | 9 of 24 (38%) |

The table shows five of the opening types; the remaining posts opened with an announcement, a contrarian line, a request for help or something rarer. An opening question took off more often in English than in Russian, though both counts are small. Only 10 English selling posts opened with a number and 1 of them took off, which is too few to say anything about numbers in English first lines.

## Which replies get liked under roll calls?

Saying what you make was the best-liked kind of reply. English replies that did so averaged the 0.59 like percentile among 466 replies, against 0.49 for the rest, and Russian replies that opened with the craft averaged 0.63 among 427. Replies containing a question did worst in both languages.

Russian replies (427 under 9 roll calls):

| Reply | Replies | Average like percentile |
|---|---|---|
| Opens with what the person does ("I weave…", "I sew…") | 38 | 0.63 |
| All other replies | 389 | 0.49 |
| 50 characters or longer | 240 | 0.54 |
| Link, @mention or "see my profile" | 30 | 0.50 |
| Ad words ("full cycle", "high quality") | 17 | 0.49 |
| Shorter than 50 characters (often just "hi") | 187 | 0.44 |
| Contains a question | 42 | 0.32 |

English replies (466 under 10 roll calls):

| Reply | Average like percentile |
|---|---|
| Says what the person makes or sells | 0.59 |
| All other replies | 0.49 |
| Links to a shop | 0.50 |
| Shorter than 50 characters | 0.43 |
| Contains a question | 0.23 |

Links were neutral in both languages (0.50); in English, roll-call authors often ask people to drop their links. The most-liked Russian reply was one plain sentence from someone who weaves garden furniture and teaches the craft: 2,549 likes under a post that had 1,000.

## Comparison with an earlier sample of 668 posts

The earlier sample agrees on calls to share and can't settle much else. Before this study we analyzed 668 unique posts from the Top tab of Threads search, collected in October 2026 for 18 queries about SaaS, indie products and AI tools ("build in public", "micro saas"); 468 were in English. Simple text rules labeled them instead of an AI model, and follower counts weren't collected there either.

| Earlier sample, English posts (468) | Posts | Median likes | Median replies |
|---|---|---|---|
| Asks people to share their own work | 13 | 42 | 112 |
| No call to action | 434 | 14 | 11 |
| Points to a link or profile | 12 | 1 | 1 |
| First line with numbers or money | 35 | 37 | 18 |

Calls to share match the roll-call result: in both samples, posts asking people to show their work drew more replies than likes. Links are harder to compare. Posts pointing to a link got a median of one like there, while a link in a reply under a roll call was neutral. In the earlier sample, a figure in the first line went with high median likes (37); this study had too few English posts opening with a number to check that.

## Limits of the study

The main limit is the sample: 901 posts that Threads search showed in its Top tab for business phrases, which is a narrower slice than a random draw from the feed. The results show what search surfaced for these queries in October 2026 and say nothing about memes, news or celebrity posts. The other limits:

- Likes and replies stand in for results. Clicks, orders and revenue weren't measured.
- Follower counts weren't recorded, so a new shop and a large creator count the same, and part of any gap may come from who tends to write each kind of post.
- One AI model did the labeling. Single labels can be wrong, so we report shares across dozens of posts.
- We worked with text only. Photos in replies, which sellers often attach, weren't part of the data.
- The tables show what the posts that took off had in common. They can't show what made them take off.
- Two languages, one month. English and Russian already disagree on first lines; other languages may differ again.

## How to sell on Threads, based on these results

Lead with the group or the person and keep the offer for your profile. A roll call where you go first, a milestone with a true number or a plain "here's what I do" gave readers something to answer, and in English the first line worked best when it named a group you belong to. The [how to go viral on Threads](https://replyvero.com/guides/go-viral-on-threads/) guide turns these formats into five post patterns with examples.

Write one or two sentences and stop. Half of the English selling posts that took off were 88 characters or shorter. For topics in these formats, the [Threads post ideas for small businesses](https://replyvero.com/guides/threads-post-ideas/) list starts from the same data.

Under someone else's roll call, open with what you make, skip the question and add a link only if the author asks for one. The same habit helps when you look for buyers in replies, as the guide on [how to get clients on Threads](https://replyvero.com/guides/get-clients-on-threads/) explains.

Count your own orders while you test. These formats are the ones to try first if you want Threads posts that sell, but only your sales can show which of them does.

We rewrote the post generator in Replyvero, our Chrome extension for Threads, after this study. Before, 21 of 24 test posts ended with a question to the audience; afterwards none of the 24 did. The Posts tab has formats taken from these kinds, including Milestone, Why I do this and Show what you're working on, and the ones about your own business ask for your facts first, so the AI doesn't invent a result. Posts are drafts by default; if you turn on the autopilot, it publishes one of your own posts a day without review. The extension also drafts replies, which by default wait for you to send them; with the autopilot on, they go out without approval ([Threads auto reply](https://replyvero.com/threads-auto-reply/), [pricing](https://replyvero.com/#pricing)). Threads' [Terms of Use](https://help.instagram.com/769983657850450) restrict automated access, so automation carries a risk of restrictions at any pace.

## How to cite this study

The Replyvero team collected and analyzed the data in October 2026. If you quote these numbers, keep the sample next to them, for example: "in 478 selling posts out of 901 posts collected from Threads search in October 2026 (Replyvero)". The figures describe that sample only, and they count likes and replies rather than sales.

Suggested citation:

> Replyvero, "Selling on Threads: what 901 posts and 893 replies showed", October 2026, https://replyvero.com/guides/threads-selling-posts-study/

## Questions and answers

### Can you sell on Threads?

Yes, and plenty of people do: 478 of the 901 posts Replyvero collected from Threads search for business phrases in October 2026 were selling posts. Whether they led to sales, the study can't say, because it counted likes and replies. By that measure, roll calls, milestones and intros spread, while only 1 of 26 English direct offers took off, and 2 of 18 Russian ones.

### Why do direct offers rarely take off on Threads?

The study counts how often direct offers took off; it can't say why. A likely reason is that most readers don't know the seller yet and have nothing to say to a bare offer, while a result, a roll call or an admission gives them a reason to answer. Pin your offer to your profile and let those posts bring people there.

### How many selling posts took off in the study?

111 of the 478, counting a post as taken off at 100 likes or 60 replies: 61 Russian and 50 English. That is 23% of the Russian and 24% of the English selling posts. By kind of post the rate ran from 4% for English direct offers to 65% for English roll calls.

### Did follower count affect which posts took off?

The study can't tell, because follower counts weren't recorded. A shop with 300 followers and a creator with 50,000 weigh the same in these tables, so some of the gap between kinds of posts may come from who usually writes them. Treat the rates as a direction to test.

### Will these results work in my niche?

Use them as a starting point. The queries were phrases makers, shops, freelancers and founders use, so the numbers fit people selling their own work. Try one pattern for a couple of weeks and compare each post with your own median likes and replies; the rates on this page come from other people's accounts.

---

> **Let Replyvero find the conversations for you.** Describe what you sell. Replyvero searches Threads for people who need it, filters posts by meaning and drafts a short reply for each one. By default you review every draft before it goes out. The first hour is free. Try it free in Telegram: https://t.me/Replyverobot?start=md_threads-selling-posts-study