Read your creative the way it actually performs

Thumbstop, hook & hold rate — in plain terms

Each metric is a simple ratio built from how many people saw your video and how far they watched. On their own each is a single checkpoint; together they map exactly where attention leaks.

Thumbstop rate — does the ad stop the scroll?

Thumbstop rate is the share of people who scrolled past your ad and actually stopped to start watching it — 3-second video views divided by impressions. It answers one question: is your opening frame strong enough to interrupt a fast-scrolling feed? A low thumbstop rate means most people never even gave the ad a chance, so nothing downstream matters until you fix the hook frame.

Example. Your ad got 50,000 impressions and 6,000 people watched at least 3 seconds. Thumbstop rate = 6,000 ÷ 50,000 = 12%. A typical strong short-form ad lands around 25–35%; 12% says the first frame isn’t earning the stop, so test a new opening before touching anything else.

Hook rate — how many keep watching past the intro?

Hook rate is the share of people who, having started the video, are still watching at the 25% mark — 25%-watched views divided by video plays. Where thumbstop asks “did they stop?”, hook rate asks “did the first few seconds earn the next few?”. A strong thumbstop with a weak hook means a great thumbnail wrote a cheque the opening line couldn’t cash.

Example. 8,000 people started your video and 3,200 were still there at the 25% mark. Hook rate = 3,200 ÷ 8,000 = 40%. If your thumbstop is healthy but the hook is low, the fix is the spoken or on-screen promise in seconds 1–4, not the thumbnail.

Hold rate — do they stay to the message?

Hold rate is the share of viewers who watch the video to (or near) the end — completed views (or ThruPlays) divided by video plays. It tells you whether the body of the creative holds attention all the way to the offer and call-to-action. High thumbstop and hook with a low hold means people are intrigued but the middle drags before the payoff.

Example. 8,000 plays and 1,440 watched to completion. Hold rate = 1,440 ÷ 8,000 = 18%. Pair it with hook rate to read the whole curve: 40% hook but 18% hold means strong open, saggy middle — tighten the body or move the offer earlier.

Read all three together as one drop-off curve

Admaxxer lines the checkpoints up per ad — stop, then 25%, 50%, 75%, complete — so you see the exact frame people abandon. ThruPlay (Meta’s term for a meaningful watch, generally 15 seconds or to completion) is shown alongside so you can compare against the number Meta optimizes toward.

Example. Ad A: 30% thumbstop, 45% hook, 20% hold. Ad B: 12% thumbstop, 60% hook, 35% hold. Ad B’s creative is stronger once people watch — it just needs a better opening frame to get them there. Without the curve you’d have killed Ad B for low thumbstop and missed the winner.

Each metric points at a different fix

The value of separating the three is that each one has a different lever:

• Low thumbstop → rework the first frame / thumbnail / pattern interrupt.
• Strong thumbstop but low hook → fix the promise in the first few seconds.
• Strong hook but low hold → tighten the middle and surface the offer sooner.

You stop guessing which part of the ad to change and start editing the exact segment that’s losing people. One merchant with a 28% thumbstop, 22% hook ad re-shot only the first spoken line; hook climbed to 41% on the same footage and CPA dropped — because the metric pointed at the opening, not the whole ad.

Why ad spend is currency-converted before ROAS

ROAS is revenue divided by spend — but only if both numbers are in the same currency. Ad accounts and stores very often aren’t: a US-based ad account can bill in USD while the store it advertises sells in CAD, EUR, or GBP. If those two numbers were divided as-is, the ROAS would be fiction. Admaxxer converts every ad’s spend into your store’s reporting currency first, using the rate from the day the spend happened, so creative comparisons and ROAS are always apples to apples.

Your ad account and your store may not share a currency

The currency an ad account bills in is fixed when the account is created and is independent of where your store sells. Left unconverted, dividing CAD revenue by USD spend produces a ROAS that’s off by the exchange rate — and the error is invisible because both numbers look like normal money.

Example. Spend US$1,000, revenue C$2,000. Divided raw, ROAS reads 2.0×. But at 1 USD = 1.36 CAD that spend is really C$1,360, so the true ROAS is 2,000 ÷ 1,360 = 1.47×. The raw number overstated efficiency by ~36%.

Converted at the rate from the spend’s own date

Admaxxer doesn’t apply today’s exchange rate to last month’s spend. Each day of spend is converted using the reference rate from that day, so a campaign that ran across a month of moving rates is summed from per-day-accurate conversions — not flattened to one snapshot. Reference rates auto-update; your finance team can also pin a fixed rate to match the books.

Native amount is always one hover away

Conversion is for comparison, never concealment. Every converted money tile keeps its original amount, the applied rate, and the rate’s as-of date available on hover — so a finance reviewer can spot-check any tile against the ad platform’s own billing screen. A converted spend tile reads C$1,360 and its hover shows the native US$1,000, the 1 USD = 1.36 CAD rate, and the as-of date — enough to reconcile without leaving the page.

View-through vs click-through truth

When you rank creatives by the conversions the ad platform reports, you’re ranking partly on view-through credit — sales from people who saw the ad but never clicked it. That’s not wrong, but it flatters ads that get a lot of cheap impressions and can crown a “winner” that real clicks don’t support. Admaxxer shows the platform’s claim next to the click-through truth from your first-party pixel and the orders your store actually banked.

Click-through — a real, deliberate action

A click-through conversion is credited when someone clicks your ad and then buys. It’s the conservative, defensible read: there was an intentional click in the path. Your first-party pixel follows these real clicks, which is why its revenue number is lower than the platform’s and closer to what an accountant would accept.

View-through — saw it, didn’t click, bought later

A view-through conversion credits the ad for a sale by someone who only saw it (often for a second or two) and never clicked. Platforms include these by default, which inflates the conversions a creative claims — especially broad, high-impression video that gets seen a lot. Useful for understanding reach, but it can make a mediocre ad look like a star.

All three numbers, side by side, per creative

Admaxxer shows each creative three ways: what the platform claims (view-through + modeled, highest), what your pixel followed from real clicks (conservative), and what your store actually banked. A plain-English note on every row explains the gap.

Example. Platform: 120 purchases. Pixel (clicks): 70. Store banked: 74. The ~50-purchase gap is mostly view-through. Ranked by banked orders, this creative beats a sibling that looked better only because it had cheaper, broader impressions.

True ROAS decides which creative to scale

Once spend is currency-converted and you can see click-through next to view-through, the decision metric is True ROAS — the revenue your store actually banked for a creative divided by what you actually spent on it. It’s the one figure neither the platform (never sees your spend) nor your store (never sees ad spend) can compute alone. A creative with US$1,000 converted spend and C$1,360 of banked, click-attributed revenue has a True ROAS of 1.0× — not the 2.0× its raw, view-through-inflated, unconverted numbers implied. Scale on the 1.0×.

How it fits together

The three lenses answer different questions, and they’re strongest used in sequence:

1. Diagnose the creative. Read thumbstop, hook, and hold as one drop-off curve to find the exact frame people abandon — and which segment to re-edit.
2. Make the money comparable. Convert each ad’s spend into your store’s currency at the day’s rate, so ROAS compares revenue and spend on the same ruler.
3. Scale on the honest number. Compare the platform’s claim against pixel clicks and banked orders, then scale the creative that wins on True ROAS — not the impression lottery.

A creative with a strong hook but soft hold, a healthy currency-converted ROAS, and banked orders that hold up without view-through is a real winner worth scaling. A creative that only looks good on the platform’s unconverted, view-through-inflated number is not — and now you can tell them apart at a glance.

Frequently asked questions

What are thumbstop, hook rate, and hold rate?

They are three simple ratios that show where a video ad wins or loses attention. Thumbstop rate is 3-second views divided by impressions — does the opening frame stop the scroll? Hook rate is 25%-watched views divided by video plays — do the first few seconds earn the next few? Hold rate is completed views (or ThruPlays) divided by plays — do people stay to the offer? Read together they map the exact point in the video where viewers drop off, so you know which part of the creative to fix.

Do I need to set anything up to get creative-quality metrics?

No. Admaxxer reads the same video-view counts the ad platform already records for every ad — 3-second views, 25/50/75/100% watched, and ThruPlays — and turns them into thumbstop, hook, and hold rates automatically. There's no extra tag, no extra cost, and no change to your ads. Connect the ad account and the metrics appear per creative.

What's a good thumbstop or hook rate?

Benchmarks vary by product, audience, and placement, so treat them as directional. For short-form video, a strong thumbstop rate is often in the 25–35% range and a healthy hook rate frequently sits above 40%. The more useful comparison is against your own library: line your creatives up on the same drop-off curve and double down on the openings and structures that consistently retain better.

Why does Admaxxer convert ad spend into my store's currency before ROAS?

Because ROAS is only meaningful when revenue and spend are in the same currency. Ad accounts often bill in a different currency than the store sells in — a USD ad account against a CAD store, for example. Dividing CAD revenue by USD spend without converting produces a ROAS that's wrong by the exchange rate. Admaxxer converts each day's spend at that day's reference rate into your reporting currency first, so ROAS and creative comparisons are always apples to apples. The original native amount stays visible on hover for finance reconciliation.

What is the difference between view-through and click-through conversions?

A click-through conversion is credited when someone clicks your ad and then buys — a real, deliberate action your first-party pixel can follow. A view-through conversion is credited when someone only saw the ad (often briefly) and never clicked, but bought within the platform's look-back window. Platforms include view-through by default, which inflates the conversions a creative claims, especially broad high-impression video. Admaxxer shows both so you don't crown a winner that only won the impression lottery.

Which creative should I actually scale?

Scale the creative that wins on real clicks and banked revenue, not the one with the highest platform-reported number. Admaxxer shows each creative three ways — the platform's claim (view-through + modeled, highest), your pixel's click-through count (conservative), and what your store actually banked — and computes True ROAS (banked revenue ÷ real spend, currency-converted). True ROAS is the honest basis for the scaling decision because neither the platform nor your store can calculate it alone.

Related documentation

Meta attribution truth · Sources & Attribution · Workspace currency · Metric glossary · Connect your ad account · Documentation home