Start with the numbers in order: how many people saw the ad, how many clicked it, how many of those became customers. Most owners look at the last number, the sale, and decide the whole ad was a dud. But a sale is the end of a chain, and the chain usually breaks at one link, not all of them.
Picture a salon running a Facebook ad for a new color service. It gets put in front of 10,000 people. Two hundred click. Three book an appointment. The owner's instinct is to pull the ad and write new copy from scratch. That instinct skips a step. The real question is whether the ad failed to get clicked, or got clicked plenty and then lost people somewhere after.
Find the step where people actually left.
This is what analytics people call funnel drop-off, the specific point in a multi-step process where people quit before finishing, as distinct from someone leaving your site after one glance or wandering off after reading a single page. Amplitude's explainer on the concept lays out the basic math: take the people who started, subtract the people who finished, divide by the people who started, and you get a drop-off rate for each step. Do that step by step, not just start to finish, and the biggest single loss usually jumps out.
For the salon's ad, that means separating two very different problems. If 10,000 saw it and only 200 clicked, the ad itself failed to earn attention or make the offer worth a click. If 200 clicked and only 3 booked, the ad did its job, the website or booking form is where people gave up. Those are not the same fix. Rewriting the ad copy when the real leak is a clunky booking page wastes the next round of ad spend solving a problem that was never there. The smarter move is to find out which part of the funnel actually broke before you touch anything, and Knowing shows whether the weak spot is the first glance, the read or the click.
Clicking is not neutral. It is a sign something already worked.
It is tempting to treat a click as a low bar, barely a decision. The data suggests otherwise. A 2018 study by marketing researchers Vibhanshu Abhishek, Peter Fader, and Kartik Hosanagar modeled an online car-launch campaign as a sequence of hidden stages a consumer moves through, from unaware to considering to ready to buy. The authors report that interacting with an ad, clicking it, substantially raises the odds of eventual conversion, beyond what exposure alone does. A click is not just a data point, it is evidence the viewer moved a real step forward.
That reframes the salon's three-bookings problem. If clicks are consistently underperforming across ad after ad, the issue sits upstream, at the part of the ad that earns a look and a reason to act. If clicks come in fine but bookings don't follow, the ad has already done its hardest job, and the problem has moved off the ad entirely.
A click is not a low bar. It is evidence the viewer already moved a step forward.
Relevance moves people, but most ads never clear that bar.
If the drop happens early, before the click, the next question is whether the ad felt like it was actually for the person seeing it. A 2023 survey of over a thousand YouTube viewers in Ahmedabad, conducted by Manzurfatema Patawat and Swati Dave, found a strong relationship between how relevant a viewer rated an ad and three behaviors: finishing a skippable ad, paying attention to a non-skippable one, and clicking. The authors report correlations in the range of .67 to .75, all strongly significant, though a survey like this can't prove relevance causes the engagement, only that the two move together closely.
What stands out more is the gap they found alongside it: 51.6 percent of respondents said relevant ads show up rarely or never. That is a targeting and message problem showing up before the ad is even judged on its merits. If a salon's ad is going out to a broad local audience with no real reason for most of them to care, low clicks are not a copywriting problem, they are a who's-seeing-this problem.
Where the ad ran changes what counts as success.
Not every low-click ad is actually failing. A 2013 field study by researchers Edith Smit, Peter Neijens, and Robert Heath, covering over 26,000 newspaper readers and 290 print ads, found that attention works in stages: passive exposure is driven mostly by placement, more active attention is driven by the ad's physical features like size and color, and genuine engagement depends on the message and whether the reader cares about the topic. The study looked at print ads, but the same staged pattern shows up online: placement gets an ad seen, design earns a glance, and the message decides whether anyone actually engages. The authors put it directly: advertisers need placement to get seen at all, but need the ad itself to be distinctive and relevant before anyone actually processes it.
That staged model matters for digital ads too, and it connects to a separate finding worth knowing before anyone panics over a banner ad with almost no clicks. The same 2018 car-launch study found that display ads, often dismissed as low-performing because they rarely earn a click, mainly worked early, nudging people from not paying attention to actually engaging, while search ads kept mattering at every stage including the final one. The authors write that this runs against the common assumption that display ads don't do much, and note separately that only a fraction of conversions attributed to any single ad exposure were actually caused by it, a reminder not to over-credit, or over-blame, one placement for the whole sale.
What to check before touching the creative.
Before rewriting an ad that underperformed, lay out the numbers in sequence: views, clicks, and whatever happens after the click, each with its own drop-off rate. Ask where that was placed and what job it was actually given, an awareness banner judged like a direct-response ad will always look like it failed. And look at whether the audience had a real reason to feel the ad was speaking to them, since relevance keeps showing up as the thing that separates an ad people skip from one they finish.
A local gym runs the same video ad on Instagram for a month. Views are strong, clicks are weak, sign-ups from the few who click are good. The leak is before the click: the hook in the first few seconds isn't earning attention, or the offer shown isn't landing with the people seeing it. Rewriting the sign-up page would fix nothing. The ad itself needs a sharper open.
The number that matters is the one at the break, not the one at the end.
An ad that produced three bookings from ten thousand views did not necessarily fail at being an ad. It may have failed to reach the right ten thousand people, or it may have done exactly what it should and handed off to a booking page that lost everyone. Knowing which one happened decides what gets fixed next, and fixing the wrong stage just means paying to repeat the same mistake with different words on it.
Frequently asked
Common questions.
What is the difference between an ad that bounced and one with funnel drop-off?
A bounce is someone leaving after viewing one page without acting. Funnel drop-off is specific to a multi-step process, like seeing an ad, clicking it, then abandoning the booking form. Drop-off points to where in that sequence people quit, which is more useful than a single bounce number for deciding what to fix.
If an ad gets a lot of clicks but few sales, is the ad bad?
Usually not. A click is a real sign the ad earned attention and interest. If people click and then leave without buying, the more likely problem is the landing page, the offer, or the checkout process, not the ad itself. Rewriting the ad in that case treats the wrong symptom.
Does a low-click display ad mean it isn't working?
Not necessarily. Research on a car-launch campaign found display ads mainly move people from not paying attention to engaging, an early-funnel job, while search ads affect every stage including the final purchase. A banner ad judged only on clicks may be doing exactly the job it was placed to do.