The quick answer
To compare Instagram posts by engagement, sort posts by the metric that matters, then normalize the read by format, date, caption, CTA, and source link. Likes alone can be noisy. Public comments, available views and post format provide additional context. Private saves are not available for a competitor.
Across accounts, raw counts reflect several differences, including audience, format, exposure and observation time. A ratio can describe relative counts, but it does not remove those differences or establish the cause of a result. State the numerator, denominator, eligible posts and observation date.
The headline is not "this post got likes." The headline is "this format plus this hook made the audience do something visible."
Step-by-step
- Pick one primary metric. Use likes and comments as visible interaction counts and views as an available playback count. None establishes unique reach or intent.
- Sort the public posts. Put the strongest posts at the top, but keep their dates and formats visible.
- Group by format. Compare carousel to carousel, reel to reel, and image to image before making cross-format claims.
- Read the first line. Read the first line as context for a hypothesis; it does not reveal why a person clicked, saved or commented.
- Mark the CTA. A soft/no CTA winner teaches a different lesson than a post that says "shop now" or "comment below."
- Open comments. Look for repeated questions, objections, praise, confusion, and user language.
- Check the source link. Do not use a screenshot-only example as the source of truth.
- Compare the nearest weaker posts. Ask what the winner had that the weaker post did not.
What to inspect before calling a post a winner
| Signal | Question to ask | What it can reveal |
|---|---|---|
| Format | Was this a reel, carousel, or single image? | The creative container may be doing more work than the caption. |
| Caption hook | What is the exact first line? | The audience promise, tension, or proof angle. |
| CTA | Did the post ask for a direct action? | A possible prompt to investigate, without attributing engagement to it. |
| Comments | What are people actually saying? | Demand, confusion, objections, and language to reuse. |
| Posting date | Was the post tied to a launch, season, or news moment? | Whether timing made the result hard to repeat. |
A real comparison: adidas vs. Glossier
The retained Adidas and Glossier samples each contain 20 posts and were saved on June 9, 2026 UTC. Their windows differ. The samples report averages of 486,951 likes and 2,640 comments for Adidas, and 5,252 likes and 86 comments for Glossier.
Using those rounded averages, 100 × comments ÷ likes is approximately 0.54% for Adidas and 1.64% for Glossier. This describes comments relative to likes in each sample. It is not a conversion rate, a follower-normalized measure or evidence that a particular prompt worked.
Inspect the retained Adidas sample and source posts. Do not treat saved sample values as current account metrics.
How InstaSeer helps
InstaSeer gives you a structured public post list with captions, dates, formats, engagement fields, and source links. Sort the loaded posts, open the strongest post, then inspect nearby weaker posts before writing the pattern.
When you need a cleaner read, use compare mode to put two public handles side by side and review which account wins on cadence, format mix, and top post proof.
InstaSeer comparison read
| Step | Report area | Decision rule |
|---|---|---|
| Start with highest public signal | Best post card and post list. | Treat it as the first example to inspect, not the final answer. |
| Filter by format | List, detail, and grid views with type filters. | Compare reels to reels before mixing formats. |
| Open the source | Open source button. | Verify comments, creative context, and caption framing. |
| Compare lower-signal posts | Worst post and nearby post cards. | Look for the difference that might be repeatable. |
The takeaway
Do not ask, "Which post won?" Ask, "Which visible differences are worth testing, and what evidence would distinguish the explanations?" That is the useful part.
FAQ
What engagement metric should I use first?
Start with the metric that maps to your goal: public likes, comments or available views for a descriptive comparison, and first-party clicks or conversions when you have appropriate owned data.
Should reels and carousels be compared together?
Only after you first compare within format. Format affects reach, behavior, and how people read the caption.