ROAS, CPM, Hook Rate: The Paid-Social Metrics That Actually Matter

Most paid-social dashboards show you forty numbers. In a decade of running and auditing accounts for clients, I have never needed forty. I have needed about six, and I have needed to know which one to look at first when a campaign starts misbehaving. The rest are noise that makes a report look thorough while hiding the one thing that is actually broken.

This is the short, opinionated list I give every team I work with. For each metric I will tell you what it measures, what good looks like, and crucially where in the funnel it points so you know what to fix. Benchmarks vary wildly by industry, country, and offer, so treat the numbers below as rough goalposts, not gospel. Your own trailing thirty-day average is the only benchmark that truly matters.

The one number that pays the bills: ROAS

Return on ad spend is revenue divided by spend. Put two dollars in, get six dollars of revenue back, and your ROAS is 3. It is the closest thing paid social has to a north-star metric because it folds everything downstream of the click, conversion rate, average order value, and cost, into a single ratio.

What good looks like. A blended 3 to 4 is healthy for most e-commerce; lower can still be profitable if your margins or repeat-purchase rates are strong, and a subscription business might happily run at a first-order ROAS below 1. The trap is judging a whole account on a single platform-reported ROAS. Attribution windows lie in both directions, so I always sanity-check against actual revenue in the bank.

Why it is not enough. ROAS tells you that something is wrong, never what. A ratio of 1.2 could be a weak creative, a broken landing page, or simply a bid that is too high. To diagnose it you have to walk back up the funnel, which is what the rest of these metrics are for.

CPA: the version your finance team understands

Cost per acquisition is spend divided by conversions. Where ROAS is a ratio, CPA is a hard dollar figure, and that makes it the metric your CFO actually feels. If you know your customer is worth 120 dollars over their lifetime and your blended CPA is 45, you have room to scale. If CPA drifts to 130, you are buying customers at a loss no matter how pretty the creative is.

What good looks like. There is no universal CPA. The only honest target is one anchored to your unit economics: CPA should sit comfortably below your contribution margin or your payback-period ceiling. Set that number before you launch, not after the spend report arrives.

CPM: the price of attention

Cost per mille is what you pay for a thousand impressions. It is the rawest signal of how expensive your audience is to reach right now, and it moves with auction competition, seasonality, and your audience definition. Black Friday CPMs can double; a narrow, high-intent B2B audience on LinkedIn will always cost more than a broad consumer one on Meta.

What good looks like. Less a target than a baseline to watch. Track your own CPM trend and react to swings. A sudden CPM spike with flat results usually means audience fatigue or rising competition, not a creative problem. A slowly climbing CPM across weeks on the same audience is the platform telling you the well is running dry.

The subtle part. A high CPM is not automatically bad. If an expensive audience converts at a high rate, you will gladly pay more per thousand. CPM only becomes a problem when it rises and your downstream metrics do not follow.

CTR: is the ad earning the click?

Click-through rate is clicks divided by impressions. It is the first true measure of whether your creative and offer resonate, because it is the first action a person takes. Use the link-click or outbound CTR, not the all-clicks version that counts likes and expands, which flatters every report.

What good looks like. On Meta, a link CTR around 1 to 2 percent is solid, and anything north of 2 is strong. LinkedIn runs structurally lower, where 0.4 to 0.6 percent is respectable, one of several reasons the two platforms reward different creative, as I broke down in this Meta versus LinkedIn comparison. A CTR that is healthy while CPA is bad points the finger past the ad: people are clicking, so look at your landing page and offer, not your creative.

Hook rate: the metric most teams ignore

Here is the one I wish more teams obsessed over. Hook rate, also called the thumb-stop rate, is the percentage of impressions that become three-second video views. It measures one thing: did the opening frame stop the scroll? On platforms where the feed moves at the speed of a thumb, the first second decides whether anything else you made gets seen at all. If you want to get systematic about that first second, the patterns I pulled from a thousand ad hooks are the best place to start.

What good looks like. Roughly 30 percent is a decent hook rate, and the best creatives clear 40 to 50. If your hook rate is weak, nothing downstream can save the ad, because almost no one is watching long enough to reach your offer. This is why I treat it as the earliest leading indicator of creative quality, well before ROAS has enough data to be trustworthy.

Why it matters more every year. As targeting gets more automated and audiences get broader, the creative is doing more of the work the targeting used to do. Hook rate is the cleanest read on whether the creative is pulling its weight. Pair it with hold rate, the share who watch past fifteen seconds or to the end, to see whether the ad keeps the attention it grabs.

Frequency: the silent killer of good campaigns

Frequency is the average number of times each person has seen your ad over a date range. It is the metric people forget to check until a winning campaign mysteriously dies. What actually happened is that the audience saw the same creative one too many times and stopped responding, while CPMs crept up to reach the few who had not yet fatigued.

What good looks like. For prospecting, a frequency creeping past 2 to 3 over a week is your cue to refresh the creative or widen the audience. Retargeting tolerates higher frequency because intent is higher, but even there, rising frequency with falling CTR is the classic fatigue signature. When you see it, the fix is almost never a new audience. It is a new ad.

How to read them together

Numbers in isolation mislead. The skill is reading them as a chain, top of funnel to bottom, and stopping at the first link that breaks.

• ROAS or CPA is bad, CTR is fine. The ad is working; the problem is past the click. Audit the landing page, the offer, and the checkout.
• CTR is bad, hook rate is fine. You stopped the scroll but the body of the ad or the offer is not compelling enough to earn the click.
• Hook rate is bad. The opening frame is failing. Nothing downstream matters until you fix the first second.
• Everything was fine, now it is sliding. Check frequency and CPM first. You are almost certainly looking at fatigue, not a structural problem.

Notice how often the answer lands on the creative. Bidding and audiences matter, but in a world of automated delivery, the ad itself is the highest-leverage variable you control. That is also the part most teams iterate on the slowest, because producing a genuinely new on-brand creative is the expensive, friction-heavy step.

Why better creative moves all of these

Look back at the list and you will see that creative quality is upstream of almost every metric that matters. A stronger hook lifts hook rate, which lifts views, which lifts CTR. A more relevant, on-brand ad earns more clicks and conversions, which improves CTR, CPA, and ROAS in one move. And a steady supply of fresh creative is the only real antidote to the frequency fatigue that quietly kills your best campaigns.

The bottleneck is rarely knowing this. It is the cost and slowness of producing the next good ad. This is where tooling earns its keep. Adkumo was built around that bottleneck: you start from proven ads that already work, adopt them to your brand so the voice and look stay consistent, generate variations across formats and languages, and launch them to Meta and LinkedIn without tool-hopping. The point is not the tool. It is that the fastest path to better numbers is more shots on goal with better creative, and removing the friction from that loop is what lets a lean team keep frequency low and hook rate high. More shots only helps when each one is distinct, which is the case I make for why raw variation count can hurt rather than help.

Whatever you use to make them, watch these six. ROAS and CPA tell you whether the machine is profitable. CPM and CTR tell you whether the market and the ad are healthy. Hook rate tells you whether the creative earned its first second. And frequency tells you when the well is running dry. Master the chain between them and you will spend less time staring at dashboards and more time fixing the one thing that is actually wrong.