Your content is up 40% this quarter. Whether that's good news or bad news is not a question your own numbers can answer.
Forty percent means one thing if your category was flat and something else entirely if every account in it doubled. The first is a win. The second is you losing ground while the dashboard turns green. The number is identical either way, which is exactly why a number on its own is neither good nor bad. It's good or bad against a baseline, and choosing the baseline is the whole job.
There are two baselines worth having. Most teams have one.
The two baselines
The first baseline is your own past. Your YouTube this month against your YouTube last quarter, the same metric measured the same way. That comparison is clean because it holds your audience, your format and the platform's counting rules constant, so a move in the number is more likely to be a move in the work. It's the subject of benchmarking a platform against its own history, and if you only ever build one baseline, build that one.
But it has a blind spot the size of your category. Measured only against yourself, you can improve every quarter and still be falling behind, because "better than we were" and "better than the field" are different questions and only the second one decides whether you're winning.
The second baseline is the category. You against the accounts you actually compete with for the same attention. That's this article, and it's the one most teams never build, partly because it's more work, and partly because the first baseline feels like enough right up until a competitor eats your lunch while your own trend line is pointing up.
Comparing like with like
Before any category number means anything, the comparison has to be fair, and most aren't.
The rule is the same one that governs every honest comparison in content: same platform, same window, same metric. A competitor's YouTube against your YouTube, over the same thirty days, counting the same thing. Compare your TikTok views to their YouTube views and you've produced a number that measures the difference between two platforms' counting rules, not the difference between two brands. A view is a platform-specific unit (each platform documents its own definition in its analytics help pages, and they change it without harmonising with anyone), so a cross-platform comparison is arithmetic on mismatched units. Why cross-platform averages lie is the long version of why that number is worthless.
Pick the platform that matters most to you, fix the window, fix the metric, and do the whole exercise there first. One honest platform beats four blended ones.
The category median, not the category leader
The instinct is to benchmark against the biggest account in your category. It's the wrong reference point.
The leader's numbers are a product of things you can't replicate this quarter: a head start, a budget, an audience compounded over years. Measured against them, you'll always look like you're failing, and the comparison tells you nothing you can act on.
The median of your competitive set is the more useful line. Rank the accounts you compete with by whatever you're measuring, take the middle one, and that's the number that says what "normal" looks like for an account in your category right now. Beating the median is a real, reachable bar. Beating the leader is a multi-year ambition, not a quarterly benchmark.
Use the median deliberately rather than the average, for the same reason outlier analysis does: one giant in the set drags the average up until everyone below it looks like they're underperforming, which is backwards. The median is robust to the giant. It describes the middle of the field, which is where you actually live.
Share of voice: the whole-category version
The median tells you where you sit. Share of voice tells you how the whole pie is divided and, more importantly, which way the slices are moving.
Your share of voice is your portion of the total measurable attention across a named set of accounts in a named window: your views over everyone's views. Its absolute level is close to arbitrary, because it's an artefact of who you put in the set. Its trend, with the set held fixed, is one of the few numbers that catches you losing ground while your own views rise: if your views climb and your share of voice falls, the category grew faster than you did. Share of voice without a media budget is the full method, including the dominant-outlier trap that makes a set with one giant in it hard to read.
The pairing is the point. Your own trend says whether you're improving. Share of voice says whether improving was enough.
Finding the open lane
Benchmarking against the field isn't only about the score. It also shows you where the field isn't.
When you map your competitors' content and find a format or topic nobody's making, you've found a lane, which is either an opportunity or a graveyard, and telling them apart is its own skill. An empty space that's empty because nobody's got to it yet is worth entering cheaply; one that's empty because several teams tried it and quit is a warning you're getting for free. Finding the format gap in your category is the read for that, and it's the constructive half of category benchmarking: not just "how do we rank" but "where is there room".
This is where competing against the field stops being defensive. The median tells you the bar. The open lane tells you where clearing it is cheapest.
The base-rate question
Come back to "up 40%". The discipline this whole article is arguing for is a single habit: before you celebrate or panic over a change, ask what the base rate is.
A base rate is what normally happens (to the category, to the season, to accounts like yours). Growth of 40% against a category growing faster than that is underperformance. A drop in a quarter when the whole category dropped further is a strong relative result. Engagement "down" in January might be January, not you. Without the base rate, every number is a Rorschach test that people read as whatever they already believed, which is the same failure that makes "grow the brand" an ungoal: nothing can contradict it.
You won't always have a clean base rate. Category data is tier-one evidence (public, outside-in, blind to everyone's paid pushes and internal context), so treat it as directional, not precise. Directional is still enough to stop you misreading your own good quarter as a bad one, or the reverse.
The traps
The flattering set. The single biggest lever on every category number is who's in the comparison cohort, and it's the easiest one to rig without noticing. Pick weak competitors and you'll always win; pick only aspirational giants and you'll always lose. Choose the set for a written reason (same platform, same rough scale, same category) and then leave it fixed, because a set you re-pick each quarter measures your choices, not your content.
The moving window. Comparing your best month to their average month is a cherry-pick with a percentage sign on it. Same window for everyone, decided before you look.
The single breakout. One competitor's post that broke out can swing a small set's numbers for a whole window and mean nothing about the trend. Check whether a move is one outlier before you conclude the category shifted.
Confusing attention with market. Share of voice is attention, not revenue. A competitor can win the feed and lose the business. Category benchmarking reads the content layer; it doesn't read the P&L.
Where Acumin fits
A competitor watchlist is the part that makes this sustainable: it collects the public numbers for a fixed set continuously, so the set stays frozen and the trend accumulates without anyone maintaining a spreadsheet. The Snapshot's share-of-voice read is computed live across that set, and where one account dominates it, the reading is reframed around that fact rather than reporting a technically-true-but-useless "you have 4%".
The numbers are public-data estimates, labelled as such, never blended with your first-party figures into one confident line. That's the same doctrine as everywhere else: the number is real, its provenance is on the label, and the honest interpretation is part of the output rather than left to you. How to research your competitors on YouTube is the manual version of the same read.
How to use this today
Take the one platform that matters most to you and five competitors you genuinely compete with. Write down why each is in the set. For the last thirty days, get each account's median views, line them up, and find the middle of the field.
Then put your own number next to the median, not next to the leader. Whatever the gap is, that's your actual benchmark, and it's the one your own trend line was never going to show you.
Related: Benchmarking a platform against its own history is the other baseline: you against your past self. Leading vs lagging indicators is how to pick metrics that let you steer before the quarter's over, rather than just scoring it after.