Planning Instagram content without data is guesswork, and most tools sold to fix that problem cannot see the numbers that decide distribution. Understanding which metrics a tool can actually access explains why so many predictions turn out to be worthless.
The short version: the three signals Instagram has publicly named as most important are all ratios involving reach, and reach is not visible from outside an account.
Key Takeaways
- Instagram’s named ranking signals are watch time, likes per reach and sends per reach.
- Third-party viewers see likes and comments only, so they cannot compute any of the three.
- Competitive Insights benchmarks up to 10 rival accounts inside the app.
- Insights holds roughly 90 days, so anything longer needs exporting yourself.
- Automated collection of Instagram data without permission risks your own account.
Why Most Prediction Tools Cannot Predict
In January 2025, Adam Mosseri named the signals that matter most across Instagram’s ranking systems: watch time, likes per reach, and sends per reach. Note that two of the three are ratios, and the denominator in both is reach. Instagram’s own explanation of how ranking works sets out the wider picture, that there is no single algorithm but a separate system for Feed, Stories, Explore, Reels and Search, each weighing signals differently.
Now consider what a third-party analytics viewer can see when it looks at a public profile. Likes. Comments. Post timing. Follower count. That is the complete list, because it is everything the public interface displays.
It cannot see reach, views, sends, saves, profile visits, watch time or skip rate. Which means it cannot calculate a single one of the three named signals, and its “predicted likes” figure is an extrapolation from a post’s like history. That is modelling the weakest of the three signals using the only data available, then presenting the output with a confidence the input does not support.
This is the structural reason such predictions feel arbitrary. They are not measuring the mechanism. Two posts with identical like counts can have wildly different reach, and the one with lower reach is performing better on the ratio Instagram is actually reading.
What Your Own Insights Shows That Nothing Else Can
A Creator or Business account gets the data the ranking systems use, free, from the Professional Dashboard.
Insights was rebuilt into three tabs in April 2026, adding skip rate, share rate and views over time. Instagram also consolidated most reach-style metrics into a single Views number that year, deprecating separate media impressions, Reel plays and Story impressions. The distinction still matters when reading your own numbers: reach counts unique accounts that saw something, while views counts every display including repeats by the same person.
The figures worth watching are accounts reached, accounts engaged, sends, saves, profile visits and follower growth. The value sits in the relationships between them rather than in any one. A post reaching 4,000 accounts with 30 sends is doing something a post reaching 12,000 with 8 sends is not, regardless of which shows the bigger number publicly.
One practical limit: the dashboard retains roughly 90 days. If you want to compare this quarter to the same quarter last year, you have to export and keep the data yourself, and almost nobody does until they need it.
Competitor Research Without the Scraping Problem
Wanting to see what similar accounts are doing is reasonable. The tools sold for it are the problem, and there is now a first-party alternative.
Competitive Insights inside the Professional Dashboard tracks up to 10 accounts you nominate, showing follower growth side by side, posting frequency, the mix of Reels against feed posts, and like counts, including on accounts that have hidden likes publicly. That covers most of what anonymous viewer tools promise, from inside the app, using data Instagram is choosing to show you.
The alternative carries a cost worth stating. Instagram’s Terms of Use state that you cannot attempt to access or collect information in unauthorised ways, including collecting information in an automated way without express permission, and Meta restricts accounts for suspected data scraping. Third-party viewers work by collecting exactly that way, so using one puts the risk on your account rather than on the service. Paying for the tool does not transfer the exposure.
A Weekly Planning Routine That Uses Real Signals
Thirty minutes a week is enough, and the structure matters more than the duration.
Sort last month by sends, not by likes. Sends carry the most weight for reaching people who do not follow you, so your highest-send posts are the ones that found new audiences. Those are the templates worth repeating.
Check non-follower reach separately. A post that performed well entirely within your existing audience taught you something different from one that broke outside it. Both are useful. They are not the same result.
Read skip rate on video. It tells you where attention breaks, which is more actionable than an aggregate watch time figure. If skips cluster at four seconds, the problem is the opening, not the subject.
Use your own follower activity data for timing. Insights shows when your audience is active by hour and day. Generic best-time-to-post windows are averages across every account in every category and time zone, which describes nobody in particular.
Change one variable at a time. Swapping format, hook style and posting time in the same week produces a result you cannot attribute to anything. This is the discipline that separates planning from superstition.
What the Data Cannot Tell You
Worth keeping honest about the limits, because analytics gets oversold in both directions.
Small accounts have small samples, and a post reaching 800 people produces numbers too noisy to draw conclusions from in isolation. Look at rolling averages over several weeks rather than reacting to individual posts. Insights also tells you what happened, not why, so the interpretation is yours and the temptation to over-explain a good week is strong.
And no dashboard predicts what will resonate. It narrows the range of sensible guesses, which is genuinely valuable and considerably less than the promise attached to most tools in this category.
Conclusion
The useful test for any Instagram analytics tool is simple: ask whether it can see reach. If it cannot, it cannot compute the ratios Instagram ranks on, and whatever it shows you is a proxy dressed as a prediction.
Your own Insights can see all of it, costs nothing, and does not put the account at risk to collect it. Start there, track sends rather than likes, and change one thing at a time.
Read Next
- How the Instagram Algorithm Works in 2026: Signals, Surfaces, and What Actually Changed
- What Instagram Views Actually Measure, and Why Buying Them Fails
- Building Your Brand on Instagram: Strategies to Attract Followers
