Guide
Upfluence, Grin, CreatorIQ, Aspire, Modash, Heepsy, HypeAuditor, Captiv8, Tagger — every discovery tool prints an engagement rate to two decimal places. The precision is comforting. It's also doing a lot of work to hide how little the number says about the thing you're paying for.
There is no standard engagement-rate formula. Some tools divide interactions by followers, others by reach or by views. Some count saves and shares, some only likes and comments. Some average the last 12 posts, some the last 90 days. That's why the same creator reads 2.1% in one dashboard and 5.4% in another with nothing changed. The practical rule: engagement rate is only meaningful as a relative ranking inside a single tool. The moment someone quotes a rate without naming the formula, it's a decoration, not a metric.
The deeper problem isn't the formula — it's who generates the number. Engagement rate is produced by people who chose to follow the creator: they recognize the face, they get the format, they'll sit through a slow open because they know the payoff. Your campaign runs on the opposite audience — strangers who give the video a couple of seconds before swiping. Fan loyalty and cold-feed hook rate are different skills, and plenty of creators have one without the other. Add that the number is gameable — giveaways, engagement pods, comment-bait captions all inflate it — and a two-decimal rate starts looking less like measurement and more like folklore.
Keep the metric for what it actually detects. A rate near zero on a big following is the classic bought-followers signature, and tools like HypeAuditor and Modash layer real fraud detection on top — audience-quality scores, follower-growth anomalies, geography mismatches. Engagement rate is a fine screening metric: it tells you whether an audience is alive and real. It was never a selection metric. Use it to throw out the fakes, not to pick the winner from the finalists who survive.
The layer missing from the discovery stack is a read on the creative itself. That's what an attention pre-test adds: run a creator's recent videos through PreTestAds and each clip gets a predicted-attention percentile against 76 top-performing TikTok ads, plus Hook Strength for the opening and a full attention curve. The model responds to what's on screen and in the audio — not to follower counts, so a nano-creator and a million-follower account compete on even terms. It's the same measurement layer, applied to hiring, that comparing influencers head-to-head walks through step by step.
The clean division of labor: the discovery platform screens the audience — real followers, right demographics, no fraud flags — and the attention score screens the content — strong hooks, held attention, a consistent median across recent clips. A creator who passes both is a defensible hire. A creator who passes only one is a specific, known risk: great content on a suspect audience, or a great audience watching content that can't hold a stranger. Either way you now know what you're buying — which is more than the two-decimal number ever told you.
Score a creator's recent clips and measure the content behind the metrics — first analysis free, no card.
Score a Creator's Content