How I ran a competitor analysis of an Instagram account for under a dollar

This week I needed to work out why a small Instagram account I help with — a little cat rescue I volunteer for — gets plenty of likes but almost none of the action it keeps asking for. Instead of guessing, I pulled the real numbers: the account’s last fifty posts, plus three similar rescues to measure it against. I handed the lot to an AI and asked it to find the pattern. The whole thing cost under a dollar and took an afternoon, and it found the problem in minutes. Here’s exactly how, so you can run the same analysis on your own account, or a competitor’s.

💡 The trick: rent a ready-made tool to copy the public numbers off a page, then let an AI read them back to you. You choose and judge; the machine fetches and counts.

What fifty posts told me that scrolling couldn’t

I used Apify to pull two things: the rescue’s own account (profile, plus the last fifty posts with their likes, comments, views and dates), and three comparable rescues to use as a yardstick. Then I pasted the raw numbers into Claude and asked it to compare them:

One of the other accounts had three times the followers on fewer posts — so the difference was approach rather than effort.
The account’s own likes were fine, but comments and shares were near zero — the posts were being seen, yet weren’t moving anyone to act.
The real killer only showed up when I traced the path a follower takes: the link in the profile went to a donations page, while every post told people to “fill in the form.” The thing the account was asking for was unreachable.

None of that came from an opinion. It came from the numbers.

What a “scraper” actually is

A scraper is a small program that visits a public web page and copies the information off it into a tidy table. You don’t build one — you rent one that already exists. Apify is a shop full of them, one for each major platform:

https://apify.com/store

Nothing sneaky is going on: it’s the same numbers anyone could read by hand, just gathered in one go.

A guess feels like knowledge — until you check it

Most marketing decisions get made on a guess: “their stuff does better than ours,” “we should just post more.” Real public data turns “I think” into “I know,” and the answer is often something you’d never spot by scrolling — as the rescue showed. For under a dollar, you can check the assumption instead of guessing at it.

Run your own audit, step by step

You won’t need to code. Start small — one competitor is enough to learn the moves.

Make a free Apify account. The free tier covers a small audit; a full run costs pennies.
Open the scraper for the platform your competitors live on. For Instagram I used the official Instagram scraper — open it and click Try for free.
Run it. Paste the account’s profile URL into the input box (the Direct URLs field), set the number of posts to about 50, and click Start. It runs for a minute or two. When it finishes, use Export and choose CSV to download the results as a spreadsheet.
Do it again for three to five comparable rivals — same size, same niche, plus your own account. The comparison is what turns a hunch into a fact.
Hand the spreadsheets to an AI assistant (Claude, ChatGPT) and ask it to find the pattern. Use the prompt below.
Follow the customer’s whole path: click the bio link, find the form, check the website. The biggest problems hide off the feed.

The prompt I used to analyse it

Paste your scraped data where the brackets are, and give the AI this:

Here is the public data from my Instagram account and three competitors:
[paste each account's profile stats and its last ~50 posts — likes,
comments, views, post type, captions, hashtags, dates].

Work only from these numbers. Don't give me generic advice. Tell me:
1. Engagement per follower for each account, ranked.
2. What my top 5 posts have in common, and what my bottom 5 share.
3. How my posting frequency and format mix compare to the others.
4. The single biggest structural difference between me and the best performer.
Finish with the one change you'd make first, and the evidence for it.

To have it pick fair competitors for you first:

I run a [type] account in [niche / location]. Suggest five real, comparable
[platform] accounts to benchmark against — similar size and focus, not the
giant national brands. For each, say in one line why it's a fair comparison.

Let the agent pull the data for you

Once you’re comfortable, you can skip the copy-paste: Apify plugs into AI agents through a connector called MCP, so the agent fetches the data and reads it in one move.

https://docs.apify.com/platform/integrations/mcp

Where people get this wrong

Confirming something you already believe is just a slower guess. Run this on questions you genuinely can’t call.
Blame the visible thing last. When results disappoint, the content gets blamed first; the real cause is usually structural — a broken link, a form no one reaches.
Public data only. Don’t reach behind logins you’re not entitled to, and follow each platform’s terms.