Guide
Everyone optimizes for the first 3 seconds. The problem is that most of your viewers have already decided to swipe before second one is over.
Platforms report the 3-second view because it's a clean line to draw, and the industry adopted it as the standard for a "hook." But 3 seconds is where the platform starts counting, not where the viewer starts deciding. On a fast-moving feed, the swipe reflex fires almost immediately — the useful mental model is that you have about a second to earn the next second, and that second to earn the one after it. By the time someone has "watched 3 seconds," a large share of your served audience is already gone.
It's tempting to imagine a cliff — viewers stay, then leave at some fixed moment. Real drop-off looks more like a steep curve that bleeds people continuously from the first frame. A slow logo intro, a beat of dead air, a shot that takes too long to reveal its point — each one accelerates the bleed. The ads that survive don't have a magic retention trick later; they simply lose fewer people early, because every second they keep is a second they can build on. Learning to read that shape is what reading an attention curve is about.
Feeds are engineered to make scrolling frictionless, and viewers are trained to keep moving. In that state people aren't evaluating your offer — they're pattern-matching. If the opening frame reads as "an ad," the reflex is skip. This is the mechanism behind hook rate and thumb-stop ratio: both measure whether your opening breaks the pattern fast enough to buy a real look. The fix isn't to cram more into the first second — it's to make the first frame visually unexpected so the reflex hesitates.
A live campaign will tell you exactly how fast people swipe — after you've bought the impressions to find out. The pre-spend version is a prediction. Upload your ad and PreTestAds scores Hook Strength for the opening and maps where predicted attention falls across the rest, so you can catch a slow open or a saggy middle before it burns budget. It predicts attention, not conversions, and it's a screen that complements live testing rather than replacing it — but fixing the swipe problem before launch is far cheaper than diagnosing it after. If you want the tactical fixes for the opening itself, start with how to write a hook.
Upload your ad and see predicted attention second by second — first analysis free, no credit card.
Score Your Ad's OpeningOn feed and short-form video, the decision to keep watching or swipe is often made inside the first second — well before the classic 3-second mark most people optimize for. The 3-second view is just the first checkpoint platforms report; most of the drop-off has already happened by then. The realistic frame is: you have roughly a second to earn a second.
Feeds train people to scroll on autopilot. Every post competes with an endless supply of the next one, so the default action is to keep moving unless something interrupts it. Viewers aren't judging your ad on its merits in that moment — they're pattern-matching it as 'ad, skip' unless the opening frame breaks the pattern fast enough to buy a closer look.
Live campaigns tell you after you've paid, through hook rate and watch-time drop-off. Before spending, you can predict it: PreTestAds scores Hook Strength for your first 3 seconds and maps where predicted attention falls across the ad, so you can fix a slow open before it costs you impressions. It predicts attention, not conversions.