Twitter Competitor Analysis: AI Winning Tweets & Profile Posts on the Apify Store →

How to do Twitter competitor analysis with AI: find the tweets that beat each account's average

Most "Twitter competitor analysis" ends in a spreadsheet: you export a competitor's tweets, add up likes, reposts and replies, sort, delete the retweets and replies, and then stare at the top rows trying to guess why they worked. Next week you do it again — and pay for the same tweets again. This tutorial shows how to get the same answer in one run, with the why written for you, using the Apify Actor Twitter Competitor Analysis: AI Winning Tweets & Profile Posts.

The problem

A social media manager watching 10 competitor accounts reads ≈ 1,000 tweets a quarter to find the 100 or so that stood out. Raw scrapers make the reading cheap — about $0.0005 per tweet — but not the thinking: a big account's ordinary post outranks a small account's breakout, retweets and promoted posts pollute the list, and nobody writes down which hook or format did the work. By hand that is about 5 minutes per winning tweet (our estimate), ≈ 8 hours per 100.

Step by step

  1. Open the Actor and click Try it. The form is pre-filled with a free demo: 54 sample tweets from 3 fictional accounts (a SaaS founder, a growth lead, a developer-relations account). Press Start — in seconds you get their 10 winners, nothing charged.
  2. Look at the Shortlist view. Each row is a tweet that beat its own account's median engagement over the window — likes + reposts + replies + bookmarks — with its multiple ("5.13× this account's median"), format, hook and a one-line why. The Shortlist view
  3. Type your own accounts. In "X accounts to analyse" type handles such as levelsio, marc_louvion, dvassallo (or pick the SaaS founders preset). Leave the window at 90 days and the multiple at 2. Add your brand in one sentence — "Acme Analytics — product analytics for early-stage SaaS founders" — to get remix drafts.
  4. Press Start. The Actor runs the Twitter Profile Scraper on your Apify account (capped by "Max source spend", $0.50 by default), drops retweets, replies and ads, computes each account's median and keeps the outliers. With AI on, every winner gets a hook type, topic, call to action, tone, a why-it-worked that quotes the tweet and names its numbers, and a remix idea plus a draft post that reuse the structure — never the words.
  5. Export or connect. Download CSV / Excel, or open the Remix drafts view to brief your writer. The Remix drafts view

What a run costs

5 accounts × 90 days ≈ 500 tweets read → ≈ 60 winners (≈ 12 %, an estimate until the first live run is measured):

Your first 25 winning tweets are free, so the first run costs ≈ $0.36 with rules. Below-median tweets, retweets, replies, ads and tweets you already received are never charged by the Actor.

Automate it

Schedule the run weekly (Actions → Schedule) with your e-mail in "E-mail me the new winning tweets". The cross-run memory delivers only new winners; a scheduled run keeps the 90-day median but returns only the winners posted since the last run. From code:

from apify_client import ApifyClient

client = ApifyClient("<YOUR_API_TOKEN>")
run = client.actor("<your-apify-username>/x-winning-tweets-ai").call(run_input={
    "accounts": ["levelsio", "marc_louvion", "dvassallo"],
    "lookbackDays": 90, "outlierMultiple": 2, "maxQualified": 50, "maxDiscoveryChargeUsd": 0.5,
}, timeout_secs=3600)
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item["outlierMultiple"], item["hookType"], item["url"], item["whyItWorked"])

A weekly run on 5 accounts costs ≈ $1.15 a month with rules and ≈ $1.50 with AI — a fraction of a $49 Tweet Hunter seat, and about the accounts you choose rather than a topic-wide library.

Twitter Competitor Analysis: AI Winning Tweets & Profile Posts on the Apify Store →