Does posting more on X get you more reach? I checked 13 days of my own small account: it helps up to about 10 posts a day, then flatlines

Every "grow on X" thread says the same thing: post more. We tested it by accident. Our small account went from 4 to 7 posts a day in the first week of November to 10, 17 and even 21 posts a day in the second, mostly automated, all in German or English about AI agents and the future of work. Here is what the X API says happened.

Disclosure: I build data tools for AI agents as Fast Data on Apify. This analysis uses only the official X API on our own account; none of our tools are needed.

The numbers

Measured on 15 November for all posts published 1 to 13 November (the last two days are left out because their posts were still collecting impressions). Our own "Source: …" replies under each post are excluded.

Day by day, the two busiest days tell the story best:

Twice the posts on 10 November produced slightly fewer total impressions than 9 November.

What we read from it

1. More posts helped, up to a point. Across all 13 days, posts per day and total impressions correlate strongly (r = 0.82). Going from 5 to 10 posts roughly doubled daily reach. That part of the advice is true.

2. Above about 10 posts a day, total reach stopped growing. Days with 13, 17 and 21 posts all landed between 310 and 361 impressions. The extra posts did not find new readers; they split the same readers.

3. Reach per post fell. The median per post was 13 to 27 on light days and 12 on the two heaviest days (correlation with posts per day: r = -0.48). Over the whole week, posts went up 26 % (58 to 73), total impressions stayed flat (1,411 vs 1,398) and the median per post dropped from 20 to 13.

4. What did well had nothing to do with volume. The three best posts of the two weeks (161, 85 and 74 impressions) were a concrete numbers comparison (months to $5M revenue for SaaS vs AI companies), a post about one developer running several coding agents in parallel, and one about a benchmark where "70 % correct is not done". Specific claims with a number beat generic news summaries.

Pull your own numbers

The X API v2 returns impression_count for your own posts. This is the script behind the table (OAuth 1.0a user context, read only):

import os, statistics, collections, datetime as dt, requests from requests_oauthlib import OAuth1 auth = OAuth1(os.environ["X_API_KEY"], os.environ["X_API_SECRET"], os.environ["X_ACCESS_TOKEN"], os.environ["X_ACCESS_TOKEN_SECRET"]) me = requests.get("https://api.x.com/2/users/me", auth=auth, timeout=30).json()["data"]["id"] posts, token = [], None while True: params = {"max_results": 100, "exclude": "replies,retweets", "tweet.fields": "public_metrics,created_at", "start_time": "2026-11-01T00:00:00Z"} if token: params["pagination_token"] = token page = requests.get(f"https://api.x.com/2/users/{me}/tweets", auth=auth, params=params, timeout=30).json() posts += page.get("data", []) token = page.get("meta", {}).get("next_token") if not token: break by_day = collections.defaultdict(list) for p in posts: day = dt.datetime.fromisoformat(p["created_at"].replace("Z", "+00:00")).date() by_day[day].append(p["public_metrics"]["impression_count"]) for day, imps in sorted(by_day.items()): print(day, len(imps), sum(imps), statistics.median(imps)) 

Two details that matter: exclude=replies drops your own reply threads (otherwise every "source" reply counts as a low-performing post), and you should measure a day only after its posts are at least 48 hours old, or recent days look artificially weak.

What this does not show

• One small account, 13 days. 75 followers, no ads, no engagement pods. A large account may behave differently.
• Not a controlled test. The busy days also had different topics and more short "news in brief" posts, which performed worse (median 5 impressions, but only 3 such posts were measured). Volume and content are mixed.
• Impressions only. Profile visits and follows would be the better goal metric; we gained very few followers either way.

Our conclusion for now: fewer posts with more effort per post. We are planning to go back from 16 scheduled posts a day to 6–8 and measure again.

Try it

The script is above; the chart code and our other small data tools live in the public data-recipes repo.

Question: if you run a small account, where did your daily reach plateau? I would like to know whether ~10 posts a day is a general ceiling or just ours.