Impressions vs Clicks, and What to Do With Both

Fourteen thousand people saw your ad yesterday. Nine of them clicked.

That's not a broken campaign.

It's two different numbers doing two different jobs. Impressions count who saw the thing. Clicks count who acted on it. Neither one means much sitting alone.

Most dashboards report both side by side and let you draw your own conclusion. Most people draw the wrong one.

A flat click line under a tall impressions bar reads like failure. Sometimes it is.

Sometimes it's an audience that's exactly right, paired with a headline that isn't. Sometimes it's a bid winning cheap placement nobody notices. Sometimes the denominator itself is padded with traffic that was never a person at all.

The fix differs for each. The two numbers together are what show you which one you're looking at.

Here's the read: what each number counts, the formula that connects them, and what a mismatch between the two is usually trying to tell you. Then what counts as normal by channel in 2026, with the reports to check it against, and where to spend the next hour based on your own account right now.

What Impressions and Clicks Actually Count

An impression is an appearance. The moment your ad or your organic result shows on a screen, the counter moves, whether or not anyone actually looked at it.

Google is specific about this in its own documentation. An impression on a search-generated surface counts once per session when the result scrolls into view, whether the person reads it or scrolls straight past. Scroll back and see it twice, and only one impression is recorded (Google Search Central help).

A click is different. Someone acted: they tapped or pressed and asked to go somewhere else.

The same documentation counts a click only when the visitor follows the link to open your result. Abandoning the results page without following through doesn't register. Neither does interacting with an on-page widget instead of the link.

Paid platforms complicate the picture further. An impression in a paid auction counts the ad being served, full stop, regardless of whether the placement ever sat inside the visible part of the screen.

Reach is a related, separate number: unique people who saw something, where impressions count every view including repeats from the same person. A campaign with 10,000 impressions and 4,000 reach means the average viewer saw it two and a half times. That's a frequency number, not a visibility one.

None of that is academic.

If your traffic report shows impressions jumping on a page while Search Console shows the same page flat, you're comparing two different counting rules, not two different weeks.

Check this today:

1. Open whichever report you actually act on: Search Console, Google Ads, or a Meta Ads Manager view.
2. Confirm which of the three it's showing you: impressions, reach, or a click counted a specific way.
3. Write that definition down next to the number.

A week-over-week comparison against the wrong baseline isn't a comparison at all.

The Formula That Connects Them

Click-through rate turns two raw counts into one number you can actually judge.

1. Pull the last seven days of impressions and clicks from whichever platform you're checking.
2. Divide clicks by impressions.
3. Multiply by 100. That's your CTR.

Say a search campaign served 8,400 impressions this week and picked up 92 clicks. That's 92 divided by 8,400, times 100, which comes out to 1.1 percent. On its own, that number means nothing.

A CTR of 1.1 percent on search and a CTR of 1.1 percent on a Facebook feed aren't the same result. Measured against the table further down, one is badly underperforming and the other is sitting close to average.

Run your own numbers before reading further. The Board's ad impressions calculator does the arithmetic if you'd rather not do it by hand, and what CTR actually means and what counts as good walks through the ratio in more depth than this post has room for. Pull your last seven days, run the three steps above, and hold onto the number. The rest of this only matters against yours, not an example.

High Impressions, Low Clicks: What It's Actually Telling You

A CTR under benchmark with plenty of impressions to work with is usually one of four things.

The first is broad or irrelevant matching: the ad is showing for searches or placements that were never a real fit, so the impression counted but the person had no reason to click. The second is a message mismatch, where the audience is right but the headline doesn't answer what they typed or scrolled past looking for.

The third is invalid traffic inflating the denominator itself. Spider AF's 2025 Ad Fraud White Paper found that up to 46.9 percent of traffic on certain ad networks is invalid, racking up impressions without ever producing a click that was going to happen anyway (Spider AF, 2025).

The fourth is simply losing the position battle: showing, but low enough on the page or feed that most people never really register it before scrolling past.

Take the 1.1 percent example from the last section. Search ads across industries averaged 6.64 percent over the twelve months to early 2026, in the tenth edition of WordStream by LocaliQ's benchmark report (WordStream by LocaliQ, 2026). That gap, 1.1 against 6.64, isn't noise.

The failure mode here is spending more to fix a problem that was never about volume. Raising budget when CTR is already low just buys more of the same wrong exposure, faster.

Do this instead, this week:

1. Pull the search terms or placement report.
2. Sort by impressions.
3. Cut anything that's racked up 100 or more impressions with zero clicks.
4. Rewrite the ad copy on the three queries or placements eating the most impression share.
5. Check again in four days.

A genuine message-match fix moves CTR inside that window, because the audience hasn't changed, only what they're being shown.

Low Impressions, High CTR: The Opposite Problem

This one gets ignored because it looks fine on paper. A CTR of 3 percent against a 1.58 percent industry average for social feeds looks like a win (Focus Digital, 2026 Facebook Ads CTR Benchmarks Report).

It is, for the people who saw it. The problem is almost nobody did.

Three things usually cap impressions while CTR stays healthy. A budget or bid too low to win enough of the auction is the most common one. An audience defined so narrowly that it's simply small is the second. The third is a platform actively throttling the objective picked, and that one's documented, not a guess: reach-oriented campaigns on Meta saw CTR decline slightly year over year in the same 2026 benchmark report, which reads as the algorithm favouring engagement-first delivery over pure visibility objectives even when an advertiser explicitly asks for reach.

The failure mode in this half of the mismatch is the mirror image of the last one: rewriting an ad that's already converting, because the raw totals look small next to a bigger campaign elsewhere in the account.

A high CTR on a tiny base isn't a creative problem. Small multiplied by nothing is still nothing, and the lever is reach, not message.

This week:

1. Check your impression share or reach setting, not the ad copy.
2. If delivery is capped by budget or by bid, raise one of the two.
3. Hold the creative steady for a week.
4. Watch what CTR does as the audience widens.

Holds steady, and you found real headroom. Drops fast, and you found the edge of a genuinely narrow, genuinely good audience worth protecting rather than expanding.

What Counts as Normal, by Channel and Sourced

"Good CTR" is a channel-specific question, not a universal one. Averaging across the three numbers below and asking whether your account beats "the average" would be comparing a search click to a feed scroll, which were never the same behaviour to begin with.

Three channels, three completely different baselines. Search sits an order of magnitude above feed-based social because the person searching already typed intent into a box. A feed scroller didn't ask for anything; they're being interrupted, which is why 0.61 percent on LinkedIn isn't a failing number the way it would be on search.

The direction matters as much as the level. All three reports show CTR moving with cost, not against it: LinkedIn's cross-industry CPC climbed 9 percent year over year to $5.74 in the same period its CTR ticked up 0.04 percentage points. Impressions there are getting more expensive and marginally more effective at the same time, not one instead of the other.

The rest of what's tracked under Traffic covers the cost side of that same trade in more depth, including what CPM actually means and how it moves.

Turning the Two Numbers Into a Decision This Week

CTR alone answers whether people click. It says nothing about whether the click was worth having.

Go back to the 8,400 impressions and 92 clicks from earlier. If that campaign spent $420 to get those 92 clicks, the cost per click is $420 divided by 92, or $4.57. If six of those 92 clicks turned into a lead, the cost per lead is $420 divided by six, or $70.

Set both against the same 2026 benchmark report: average CPC across search industries was $5.42, and average cost per lead was $66.69 (WordStream by LocaliQ, 2026). This account's CPC sits under benchmark and its CPL sits just over it, which is a genuinely different problem than the low CTR flagged two sections back. The clicks that are happening convert at close to a normal rate, so the real lever here is CTR itself, not the funnel underneath it.

That's the actual test, and it splits four ways:

• CTR low, CPL fine: the fix is upstream, in the ad or the targeting.
• CTR fine, CPL high: the fix is downstream, on the landing page or the offer.
• CTR low, CPL high: fix both, starting with the ad.
• CTR fine, CPL fine: leave it alone and scale the budget instead.

Chasing CTR by itself is how an account ends up with a great click rate and a mediocre cost per lead. The ad is clickable. The offer underneath it isn't converting. CTR never once shows you that, because CTR only measures the step before it.

What CAC actually costs, worked with real numbers and the ROAS formula explained the same way both pick up exactly where this leaves off, once clicks have turned into either a lead or a sale.

Calculate your own CPC and CPL for the last seven days right now, using the two divisions above. Set both against the benchmark for your channel from the table.

Fixing the wrong end of that chain is the single most common way a budget gets spent twice on the same unsolved problem. How much a website visitor actually costs to acquire is worth reading before you touch the budget at all.

Where a Flat Price Skips the Guessing Game

Every number in this post assumes an auction that resets overnight. Bid and position move on their own. Delivery moves with them, against competitors you can't see and won't be told about.

The Board runs on a different mechanic entirely. It's a public leaderboard of brands, not an auction: rank is the amount paid, sorted, nothing else, disclosed as a paid placement rather than an organic result the whole way down. Every listing publicly shows its own impressions and clicks for the current holder's reign, right next to what it paid to get there.

None of the earlier math stops applying once a listing goes live. A spot on the board still earns a real CTR, still delivers an impressions number worth checking against the table above, and a reader who clicked through still needs a page underneath that actually converts.

What a fixed placement removes is the week-to-week repricing. The number due at checkout is the number, until somebody else pays more for the same rank.

Whichever channel you're running today, the read stays the same: two numbers, one ratio, and a decision that only makes sense once you know which side of it is actually the broken one. Pull this week's figures before you touch next week's budget.