A Snapshot is a read on what your public content has actually been doing. It's not a dashboard, and reading it like one is the fastest way to take the wrong thing from it.
Each section answers a different question, and they're arranged deliberately: volume, then rhythm, then the single strongest pattern, then your best work, then how any of it compares. This is a walkthrough of what each part supports, and what it doesn't.
Before the numbers: the evidence tier
Top of the page, next to your brand name, there's a tier label.
That's the most important thing on the Snapshot and the thing most people scroll straight past. It tells you what this read is made of, and therefore which decisions it can carry.
- Tier 1 means public data. Views, likes, comments, upload dates. It can tell you what drew attention. It cannot tell you what held it.
- Tier 2 means your connected platform analytics are in the mix: real reach, watch-through, retention.
If you're on tier 1 and about to make a large creative commitment, the honest move is to connect a platform first. What your evidence is actually worth explains why the gap between those two tiers is bigger than it looks.
Totals: the volume view
Total views, videos analysed, channels live.
Two of these are context. The third, videos analysed, is the one to actually read, because it's the sample size for everything below it.
Thirty analysed videos supports a reasonable read. Six does not. If this number is small, treat every conclusion further down as a direction to watch rather than a finding, no matter how confidently it's phrased. That's not a flaw in the analysis; it's the analysis correctly reporting on a thin channel.
Per-channel sections: the rhythm view
Each connected platform gets its own section, with a cadence heatmap.
The heatmap shows when you actually published, not when you meant to. It's usually the most quietly uncomfortable thing on the page, because publishing plans and publishing reality diverge more than teams remember.
What to look for: gaps, and whether they correlate with anything. A three-week hole in March, followed by a slow recovery, is often the real explanation for a "declining performance" narrative that everyone has been attributing to content quality.
The headline: the strongest single pattern
One statement, in a card, with two things attached to it. Read all three parts.
The statement. The strongest pattern found in your data.
The confidence level. Shown as Well-evidenced, Solid, or Early signal. This is derived from how much data sat behind the claim and how consistent it was, not from how good the sentence sounds.
The evidence tier. Where that data came from.
If the level is Early signal, the correct response is usually to test rather than to commit (or to add evidence by connecting a platform, which is often faster than waiting for more posts).
What's working: your top performers
Your best posts, ranked by format, from real public view counts. Not estimates, not a model's opinion: counted numbers, each linking back to the original.
Two ways to read this well.
Read within format, not across it. A Short and a twelve-minute video are measured differently, by different systems, for different behaviour. Ranking them against each other tells you about the formats, not the films.
Look for the pattern, then check the losers. The instinct is to look at the top three and find what they share. Do that, and then look at your worst posts and check whether they share it too. If your best five all open with a question and your worst five also do, you've found a habit, not a signal. That check is the whole difference between analysis and confirmation.
Outliers, not virality goes deeper on the method.
How you compare: the competitor band
Your triangulation against the competitors you named at intake.
This section is only as good as the cohort you chose, and it's worth being ruthless about that. If you included a channel ten times your size, every dimension will say you're losing, and that's arithmetic rather than insight. If you included only channels you already beat, the reverse.
The comparison worth trusting is [engagement per view](/glossary/engagement-per-view) rather than raw views, because it strips out the distribution advantage that comes with being bigger. A smaller channel out-engaging a larger one per view is a real and useful finding. A larger channel getting more total views is not a finding at all.
The three common misreads
Treating a small sample as a verdict. The most frequent one. Check "videos analysed" before you believe anything, and re-check it before you spend money.
Reading the sentence and ignoring the confidence level. The statement is written to be clear. The level is what tells you how hard to lean on it. Clarity is not confidence.
Assuming public views measure quality. They measure attention, which is a product of the content and of how the platform chose to distribute it. A post with a great title and a weak middle looks identical to a genuinely great post at this tier. That's precisely the blind spot tier 2 closes.
What comes after
Three sensible next moves, depending on what the Snapshot told you.
The headline is Well-evidenced and you have a decision to make → generate a Signal Brief. It turns the read into a decision memo: a recommendation, the options considered, the risks, and every claim carrying its own confidence and evidence tier. The free account includes one generative read, so spend it on a real question.
You're unsure about a specific idea → run a Concept Read. You get a band (Promising, Mixed or Weak) on an idea before you shoot it. It's genuinely good at killing ideas cheaply, which is worth more than it sounds.
The confidence levels are mostly Early signal → connect a platform. Read-only, a few minutes, and it changes the instrument rather than just adding data.
Saving it
Snapshots can be saved and compared over time, which is where the real value shows up. A single Snapshot is a description. Two Snapshots three months apart, with the first one frozen, is a measurement. A frozen benchmark is the only honest way to tell whether something you changed did anything.
Take the first one now, before you change anything. You can't go back and measure a baseline retrospectively.
How to use this tomorrow
Open your Snapshot and write down three things: the videos-analysed count, the headline's confidence level, and one pattern shared by your top posts but not by your worst.
If you can't complete the third one, you have your answer about what to do next, and it isn't "make more content".
Related: Your first 20 minutes in Acumin is the setup that produces this page.