How-To Guide
Threads or X: which should a brand invest in?
Threads has more daily mobile users and X still leads on the web. Compare size, engagement, ranking and cost, then run a four-week test before you pick.
No source read shows a single winner, so match the network to your buyers and test. Threads has more daily mobile users (141.5 million against 125 million, Similarweb, January 2026) and a higher median engagement rate in Buffer's 2024 data (6.25% against 3.6%), while X has 145.4 million daily web visits to Threads' 8.5 million and leads in the US. Start where your buyers already are, then post the same eight ideas on both for four weeks and keep the one that brings more signups, with replies per 1,000 impressions as the tiebreaker.
- Size depends on the yardstick: mobile daily users favor Threads, web visits favor X.
- Engagement figures for Threads range from 1.8% to 6.25%, and Buffer warns the networks count differently.
- Superdirector writes and posts X text posts for one account on its $199 plan and does not post to Threads.
Threads or X: what the numbers say
No source read here shows one winner for every brand, and the figures that exist measure different things. Threads has the bigger daily mobile audience: Similarweb counted 141.5 million daily users on iOS and Android against 125 million for X on January 7, 2026, and Meta says Threads reached 500 million monthly active users in June 2026. X still has the web. The same Similarweb data put X at 145.4 million daily web visits against 8.5 million for Threads, and X still leads in the US.
On engagement, Buffer's analysis of 10.2 million posts from 2024 found a median engagement rate of 6.25% on Threads and 3.6% on X, and Buffer warns that the two networks may calculate engagement differently. A sensible plan is to start on the network your buyers already use and test the other for four weeks before you commit.
| Yardstick | Threads | X | Who measured it, and when |
|---|---|---|---|
| Monthly active users | 500 million | No figure found on a readable primary page | Meta newsroom, June 16, 2026 |
| Daily users on iOS and Android | 141.5 million | 125 million | Similarweb, via TechCrunch, January 7, 2026 |
| Daily web visits | 8.5 million (Threads.com and Threads.net combined) | 145.4 million | Similarweb, via TechCrunch, January 13, 2026 |
| US daily users | Behind X, gap narrowing | Ahead | Similarweb, via TechCrunch, reported January 18, 2026 |
| Median engagement rate, 2024 posts | 6.25% | 3.6% | Buffer, February 3, 2025, from Buffer users' posts |
| Change in brand engagement rate, 2025 report | Not in the report | Down 48% | Rival IQ, February 25, 2025 |
Do not add across these rows. Monthly users, daily users and web visits are different units, and the engagement rates come from different samples. The chart below shows three of them side by side to make the point.
Different yardsticks point different ways, and each row carries its own units. Original diagram, from Similarweb figures reported by TechCrunch in January 2026 and Buffer's February 2025 analysis of 2024 posts.
Make it work for your brand
Find it. Remake it. Post it. Superdirector turns the viral formats in your niche into your brand's carousels, X posts and video scripts, and keeps the calendar full.
Match the platform to where your buyers already are
Start from your buyers, not the totals. Pew's 2026 fact sheet, based on responses from 5,511 U.S. adults, has a row for X in its demographic tables and none for Threads. Its question is whether people ever use X, and under-30s say yes most often, at 31%; the share drops to 20% for people in their 30s and 40s, 17% for ages 50 to 64, and 7% for people 65 or over. More men than women say they ever use X (23% against 14%), and so do richer households: 24% where yearly income is $100,000 or higher, against 14% below $30,000.
Pew has no age split for Threads. It does have a topline: in its 2025 wave, 8% of U.S. adults said they ever use Threads, against 21% for X, and its 2025 report says Democrats are more likely than Republicans to use Threads, while Republicans are more likely to use X (24% of Republicans against 19% of Democrats). TechCrunch notes that Threads' mobile growth may come from cross-promotions inside Facebook and Instagram, so a brand whose followers already come from Instagram has a natural first test. Check who follows your competitors on each network before you decide. If you are still choosing among more than these two networks, the guide to which platforms a startup should use compares the wider field.
| If this is true for you | Lean toward | Why, and how sure it is |
|---|---|---|
| Your buyers and competitors are already active on X | X | Pew shows X use is highest among adults under 30; Pew gives Threads no age split, so confirm with your own search. |
| Most of your followers come from Instagram | Test Threads first | TechCrunch reports Meta promotes Threads inside its other apps; that is a reported possible driver, not a measured effect on brands. |
| Your buyers are mostly U.S. adults over 65 | Neither is proven | Pew shows 7% of that group ever use X; Pew gives Threads no age split, so test before you commit. |
| You post news-driven takes | X | Buffer's view is that X favors timeliness; Buffer's reasoning, not a measurement. |
X use falls steeply with age, and Pew's tables give no matching age split for Threads. Original diagram, from Pew Research Center's 2026 fact sheet, a survey of 5,511 U.S. adults.
Read the engagement numbers by who measured them
Several Threads engagement figures circulate, and they span 1.8% to 6.25%. Every source that shows both networks puts Threads ahead, and they disagree by a wide margin on how far. WebFX also saw Threads engagement fall during 2024, so the date of a figure matters.
| Source | Threads | X | What it measured |
|---|---|---|---|
| Buffer, February 2025 | 6.25% | 3.6% | Median across 10.2 million posts from 2024, made by Buffer users |
| Sprout Social | Roughly 3.8% to 4.5% | Not given | Median Threads engagement rate, as Sprout states it |
| WebFX, 2024 to 2025 | 4.51%, falling from 4.76% in early 2024 to 3.60% by February 2025 | Not given | Average from WebFX's own analysis |
| Lumus Consulting, 2025 | 1.8% for brand accounts | 0.045% for unverified brand accounts | Per-follower rates that Lumus credits to Klaviyo (Threads) and Social Media Examiner (X); two studies of different accounts, originals not found, not used here as evidence |
The gap between Buffer's two medians is about 1.7 times. Lumus pairs figures it credits to two different studies of different accounts, both per follower, to claim a gap of about 40 times. That is not a like-for-like comparison, so we set it aside. Engagement rate also depends on the denominator: Buffer divides interactions by reach or impressions, and says Threads and X might calculate engagement differently. Rival IQ adds the other side of the story for X: its 2025 benchmark of brand accounts found X's engagement rate took the biggest hit at 48%, and brands' posting frequency on Twitter fell 33%.
Use these numbers to expect that Threads posts may draw more interactions per view than X posts, and use your own four-week test, below, to find the size of the gap for your brand. The engagement rate glossary entry explains the formulas.
Compare how each feed decides what to show
X publishes its feed code. Its README says the For You feed mixes posts from people you follow with posts it recommends and ranks them with a transformer model, and its weights file, as we read it in November 2026, gives a reply a weight of 5.0 and a like 0.708. Meta publishes no weights for Threads that we found. It describes controls instead: Your Algo lets people privately choose which conversations they see, and it began rolling out in five countries in June 2026.
| Lever | X | Threads |
|---|---|---|
| What is published about ranking | Code and weights on GitHub | No weights found; Meta describes a viewer control (Your Algo) |
| What a reply is worth | Reply weight 5.0, like weight 0.708, so one reply counts for about seven likes | Buffer's view is that Threads favors meaningful interactions such as replies; no weights published |
| How you label a post by topic | Not covered in the sources read | One community-style tag per post, per Buffer; communities left beta in June 2026 |
The practical point is the same on both: write posts that earn replies and ask one clear question, and on Threads add the one tag that names your topic. For X that rests on published weights; for Threads it rests on Buffer's reading, because Meta has published no weights.
One reply carries about seven times the weight of one like in X's feed. Original diagram, from the weights file in X's open-source feed code, read November 2026.
Count what each network costs to run
Posting by hand costs time on both. Posting through software adds limits and prices that your scheduler absorbs or passes on, so ask yours. X's API is pay-per-usage: creating a post costs $0.015 per request and creating a post with a URL costs $0.200, about 13 times more. The Threads API overview lists limits and no price: each profile may publish 250 posts and 1,000 replies through the API in 24 hours.
| Cost or limit | X | Threads |
|---|---|---|
| Create one post by API | $0.015 per request | No price on the overview page |
| Create a post with a link by API | $0.200 per request, about 13 times a plain post | Not stated on the overview page |
| The limit you plan around | The meter: every call is billed, and link posts cost the most | The caps: 250 posts and 1,000 replies per profile in 24 hours |
| Ads | Not covered in the sources read | Global rollout announced January 21, 2026, low ad load at first |
Here is the X meter at one founder's volume. Say you schedule three X posts a day through a tool that bills X's API at cost, using the prices X lists. With no links, a month of 90 posts costs 90 × $0.015 = $1.35. If one post a day links to your product page, the month costs 60 × $0.015 + 30 × $0.200 = $0.90 + $6.00 = $6.90. If every post carries a link, it costs 90 × $0.200 = $18.00. The same 90 posts on Threads sit far below the cap of 250 a day, and the overview lists no price.
| Three X posts a day for 30 days | Plain posts | Link posts | API cost for the month |
|---|---|---|---|
| No links | 90 × $0.015 | 0 | $1.35 |
| One link a day | 60 × $0.015 | 30 × $0.200 | $6.90 |
| A link in every post | 0 | 90 × $0.200 | $18.00 |
These are API charges only. Your scheduler's own price is separate, and whether it passes X's charges on is a question for its vendor. Paid reach is a separate question again. On January 21, 2026, Meta announced ads on Threads for all users worldwide, rolling out gradually, with ad delivery staying low at first. The sources read give no X ad figures, so compare ad costs inside each network's own ad tool.
Link posts carry almost the whole X API bill at these volumes. Original diagram, from X's listed API prices of $0.015 per post and $0.200 per post with a link, read November 2026.
Run both for four weeks, then keep one
The studies cannot settle this for your brand, so run the two networks side by side. Post the same ideas, adapt each first line, and compare replies per 1,000 impressions, a ratio both networks let you compute. The crossposting glossary entry covers adapting one idea for several networks.
- Write eight posts that stand alone on both networks: a claim, a result, a question, a short list.
- Post each one on both networks on the same day and time, adapting the first line, and add one topic tag on Threads.
- Seven days after each post, log impressions, replies, profile visits, link clicks and signups.
- Add up replies and impressions for each network and divide, to get replies per 1,000 impressions.
- Keep the network that brought more signups. If signups are level, keep the one with more replies per 1,000 impressions, and if that is level too, stay where more of your buyers already follow you. This is a house rule with no statistical test behind it.
Here is how the arithmetic runs for an example brand: a two-person DTC shop selling pour-over coffee kettles. The numbers are invented to show the method; they are not a result or a benchmark. After four weeks, its eight Threads posts had 9,600 impressions and 58 replies, and its eight X posts had 14,000 impressions and 42 replies. Replies per 1,000 impressions come out at 58 ÷ 9,600 × 1,000 = 6.0 on Threads and 42 ÷ 14,000 × 1,000 = 3.0 on X. Signups, counted from a different tracking link on each network, were 3 from Threads and 7 from X.
| Measure after four weeks (invented) | Threads | X |
|---|---|---|
| Impressions, 8 posts | 9,600 | 14,000 |
| Replies | 58 | 42 |
| Replies per 1,000 impressions | 6.0 | 3.0 |
| Link clicks | 41 | 66 |
| Signups | 3 | 7 |
The two measures point different ways. Signups decide, so this brand keeps X and posts on Threads only now and then, by hand. With counts this small, a gap of four signups can be chance, so if your own result is close, run four more weeks before you drop either network.
The test sheet holds the columns and formulas, and the weekly content plan guide gives a weekly rhythm for the four weeks.
Signups decide, and here X wins them 7 to 3. Original diagram of the test this article describes; the figures are invented to show the arithmetic, not a result or a benchmark.
Where Superdirector fits
Superdirector, which publishes this guide, offers autonomous social media growth for founders, app owners and DTC brands. For $199 a month its AI runs one social account in the direction you set: it watches the accounts, competitors and keywords you name across Instagram, TikTok, YouTube and X, drafts X posts and Instagram carousels in your brand voice, and eligible Instagram carousels and single link-free X text posts can be published to a connected account under its standing posting permission. Threads is not one of the networks it posts to, so a Threads test means publishing there by hand. A $29 a month plan covers hands-on work from Claude or ChatGPT. AI video and full autopilot have not launched; both are on the way. See what runs today.