The idea
You can buy likes and comments for a post in minutes. Some people say social proof on an ad’s post makes Meta deliver it more cheaply. Others say Meta spots it and punishes the ad. I want to see which one happens, if either.
So I’m running two identical ads. The only difference: one post gets bought likes and comments before launch, the other gets nothing.
To be upfront: buying engagement breaks Meta’s rules. That’s why this runs on a throwaway ad account and a throwaway Page, not the Game of Algorithms Page or any Page from another experiment.
Why this test: Should You Buy Likes and Comments for Your Meta Ads?
I don’t know yet. Nothing has been bought or launched. The design is locked. No ads, no engagement, and no result yet.
Hypothesis
An ad whose post has bought likes and comments will get different delivery from an identical ad whose post has none. That could mean a reward (cheaper CPM, more reach) or a punishment (worse ad quality ranking, higher CPM, less reach).
Null: Bought engagement does not change delivery when creative, audience, budget, and schedule match.
- Design locked
- Not started ← you are here
- Result
How I’m setting this up
1. Two identical ads
What I did: Same creative, same audience, same budget, same schedule.
Why: If anything else differs, a gap in delivery could come from that instead of the bought engagement.
2. Same budget, fixed length
What I did: $10/day per ad for 14 days. $280 in ad spend total.
Why: Equal money over the same window keeps the comparison fair.
3. A throwaway account and Page
What I did: A separate ad account and Page used only for this test.
Why: Buying engagement violates Meta’s rules. I’m not risking the Game of Algorithms Page or any Page from another experiment.
4. A coin flip picks the treatment
What I did: A coin flip decides which post gets the bought engagement.
Why: It stops me from picking the post I think will win.
5. One purchase, treatment only
What I did: The treatment post gets 250 likes and 20 comments, bought once for $30 from a third-party engagement vendor. The control post gets nothing bought.
Why: One fixed purchase means there is one clear difference between the two ads.
6. What’s not in this experiment
What I left out on purpose: Whether the comments look real to a person. Organic reach.
Why: This test is only about what Meta’s ad delivery does next.
Scoreboard
Meta Ads Manager
Delivery, not likes.
- CPM
- Reach
- Cost per result
- Ad quality ranking
Does engagement move rankings?
Live · Baseline upcoming