Google Ads Benchmarks in 2026: What a “Good” CTR, CPC and Conversion Rate Actually Look Like

Someone in a meeting says your conversion rate is “below benchmark.” Maybe it’s a client, a boss, or a founder who read an article over the weekend. You pull up the account, see 4.1%, and then somebody quotes 8.18% like it’s a speed limit you’ve been ignoring.

That number is real, in the sense that it was calculated from real campaigns. It just probably isn’t telling you what the person quoting it thinks it’s telling you. Benchmarks are useful, but mostly as a way to catch a big mistake, and much less as a way to grade a campaign. This post is about how to use them for the first job and stop using them for the second.

We’ll go through what the 2026 numbers say, why they disagree with each other, what’s baked into them that you can’t see, and what to compare your account against instead. There’s also a short worked method at the end for setting a baseline you can defend in a room full of people quoting averages.

What the 2026 benchmark numbers actually say

The benchmark most people quote comes from WordStream, which is published by LocaliQ. The 2026 edition analyses search campaigns that ran between April 2025 and March 2026. Coverage of the report describes a sample of more than 13,000 US campaigns, and the all-industry averages come out like this:

• Click-through rate: 6.64%
• Cost per click: $5.42
• Conversion rate: 8.18%
• Cost per lead: $66.69

A few other details from the same report are worth keeping in your head. Average CPC is now more than double the $2.32 the study recorded when it started in 2016. Cost per lead fell for the first time in five years, to $66.69, and the write-ups compare that to $59.18 a decade ago. Conversion rate improved in 87% of industries, which means the drop in cost per lead came from better conversion and not from cheaper clicks. And this is the tenth annual edition, so it’s one of the longest-running series in the industry, which is a fair reason it gets quoted so often.

The spread between industries is where it gets interesting. According to the report’s coverage, Arts and Entertainment had the lowest average CPC at $1.63, with Restaurants and Food close behind at $2.05. Roundups of the same data put Attorneys and Legal Services near $9.87 per click and Home and Home Improvement around $8.33. On conversion rate, those roundups list Animals and Pets at 16.22% and Automotive Repair at 15.51%, against 2.64% for Finance and Insurance and 4.50% for Apparel. On CTR, Arts and Entertainment is above 12%, and Finance and Insurance and Travel are both above 9%.

Take a second with that. The same study says a click costs $1.63 in one industry and nearly six times that in another. It says the conversion rate in one vertical is six times the rate in another. “The average” sits in the middle of a distribution that is enormously wide, and almost nobody’s account sits in the middle of it.

There’s one quiet detail in the methodology that matters too. The data is a blend of Google Ads and Microsoft Ads search campaigns. So when someone says “Google Ads benchmark” and quotes 6.64%, they’re quoting a number that isn’t purely Google. That’s not a scandal, and the authors say so openly. It’s just an example of how a headline figure travels away from its footnotes.

Why two “2026 benchmarks” can disagree by a factor of two

Search for the average Google Ads CTR and you’ll find different answers. WordStream says 6.64%. At least one other 2026 roundup puts the average search CTR at 3.52%. Both can be described as “the average Google Ads click-through rate in 2026,” and neither is lying.

The reasons are boring and they matter:

• Different samples. One study draws from advertisers using a particular software platform. Another uses agency-managed accounts. Another uses a panel of accounts that opted in to share data. Each of those groups looks different from the whole population of advertisers.
• Different campaign mixes. Branded search campaigns get very high CTRs and convert well. A study heavy on brand terms will look better on both metrics than one dominated by generic non-brand keywords.
• Different definitions of what counts. This is the biggest one, and it’s worth its own section.

Conversion rate is the clearest case. Commentary on the 2026 numbers points out that an older WordStream figure of 7.52% counted calls and form fills as separate conversions, while a Databox figure of 1.2% skews toward purchase-only ecommerce tracking. One number is inflated by counting several kinds of action; the other is deflated by counting only the hardest one. Neither is wrong. They’re just answers to different questions, and you can’t hold your account up against both.

The same logic applies inside your own account. If you count newsletter signups as conversions, you might show a lovely 9% while your real purchase rate is under 1%. If you count only completed checkouts on an ecommerce store, you might look broken at 1.5% while the business makes money. This is a point that comes up constantly, and it’s the reason we wrote a separate piece on auditing conversion tracking when the numbers don’t add up. Before you compare anything to anything else, you need to know what your conversion column is counting.

Who is in the sample, and who isn’t

Every benchmark is built from accounts that got included. That sounds obvious until you ask who tends to get included.

Accounts that are connected to a tool, or run through an agency that shares aggregated data, or spending enough for the vendor to care about, are over-represented. Accounts that spent $200 in a month, tried Google Ads for six weeks and quit are not. The dead ones drop out of the data quietly, so the survivors set the average. Commentary on these reports makes the same point: benchmarks skew toward larger spenders, so treat them as directional and then normalise against your own lifetime value, margin and close rate.

There’s a second, subtler effect. The report from WordStream is US-only. If you run in the UK, Australia or Germany, your CPCs and conversion behaviour may look nothing like it. Auction density, language, checkout habits and even the typical purchase price are different. A US average is a poor yardstick for a Dublin plumber.

None of this means the benchmark is useless. It means it answers a narrower question than it appears to. “What does a typical, reasonably managed US search campaign in this industry look like?” is a fair question that the benchmark can help with. “Is my campaign good?” is a different question, and it needs different evidence.

The arithmetic that holds every benchmark together

Here’s a small piece of maths that makes benchmarks easier to think about. Cost per lead is roughly CPC divided by conversion rate.

Take the 2026 all-industry averages. $5.42 divided by 0.0818 comes to about $66.26. The reported cost per lead is $66.69. That’s very close, and the small gap is what you’d expect when you average ratios across thousands of campaigns instead of computing one ratio from totals. So the four numbers are more or less internally consistent, which is reassuring.

The useful thing about this relationship is what it tells you about where to look. If your cost per lead is higher than you’d like, there are only two levers in the equation:

• You pay too much per click, or
• Too few of those clicks turn into leads.

That’s it. Everything else you might tinker with, such as ad copy, landing pages, match types, bid strategies, audience layers or scheduling, works by moving one of those two numbers. When you’re looking at a benchmark gap, the first question is which of the two you’re actually off on. A campaign that’s paying a lot per click but converting well can be perfectly healthy. A campaign with cheap clicks that never convert is the more worrying pattern, and no CPC benchmark would ever flag it.

Suppose your CPC is $3.20 and your conversion rate is 2.5%. Your cost per lead is $128. A blended benchmark might tell you your CPC is 40% under average and your conversion rate is far under average, and the honest reading is that cheap traffic isn’t turning into leads. The likely investigation is the search terms you’re matching, the landing page, or the definition of the conversion, and the CPC is a distraction. (That’s a made-up example to show the method, not data from any account.)

CTR: the benchmark most likely to mislead you

CTR is the most quoted number and the least connected to money. It tells you how often people who saw your ad clicked it. It doesn’t tell you whether those people were any good.

Three things move CTR a lot without changing anything about how well your account is doing:

• Brand versus non-brand mix. People searching your company name click your ad at very high rates. If you add a brand campaign, your blended CTR goes up. If you pause it, CTR falls. Neither move says anything about your non-brand ads.
• Position. Ads in the top spots get clicked more, so a campaign with high impression share at the top of the page will show a higher CTR than one that appears lower down for the same ad copy.
• Query breadth. Broad targeting generates impressions on marginal searches. Those impressions dilute CTR. Tightening targeting raises CTR and may lower total conversions.

The Search Engine Journal write-up of the WordStream study says something sensible on this: benchmarks shouldn’t be treated as universal goals, because expectations vary with budget, competition, geographic targeting, conversion actions, sales cycle, landing page quality and business model. CTR is exactly where all of those variables pile up.

If you want to use CTR properly, use it at the ad group or keyword level, on non-brand traffic only, and against your own history. A drop in CTR on a stable keyword after an ad change is information. A drop in blended account CTR after you launched a broad campaign is arithmetic.

Quality Score sits close to this topic, since expected CTR is one of its components. We covered what it does and doesn’t measure in our piece on Quality Score, and the same warning applies. A number Google reports about your account is a diagnostic, and it stops being useful once it turns into a goal.

CPC: high isn’t bad, and low isn’t good

The average CPC of $5.42 will make some people wince and others laugh, depending on their industry. If you sell legal services, a click at that price would be a bargain. If you run a restaurant, it would be a disaster.

CPC is set by an auction, and your part of the auction is shaped by things a benchmark can’t see: your bids, your ad relevance, your landing page experience, the competitors bidding on the same terms in the same locations at the same hours, and how much each conversion is worth to you. Comparing your CPC to an industry average tells you whether you’re in the ballpark. It doesn’t tell you whether you’re bidding well.

The better question is whether you can afford the CPC you’re paying. That depends on the rest of the funnel:

• What share of clicks become leads?
• What share of leads become customers?
• What’s a customer worth, after costs?

If a click costs $9 and one in ten clicks becomes a lead, a lead costs $90. If one in four leads closes, a customer costs $360. If that customer is worth $2,000 in gross profit, the campaign works and the CPC is irrelevant. If the customer is worth $300, it doesn’t work at any CPC you’d get from this keyword. The benchmark could say your CPC is 60% above average and it wouldn’t change either conclusion.

There are a few CPC situations where the number does carry a warning, though:

• A sudden jump with no change on your side. That usually means a new competitor, a change in match behaviour, or a shift in what queries you’re matching. It’s worth checking the search term report before you assume it’s the market.
• A CPC that’s high because you’re bidding on the wrong geography. Location settings that include people who are merely interested in a place, not present in it, can pull in expensive and irrelevant clicks. We wrote about that in the location targeting post.
• A CPC that’s mysteriously low. Cheap clicks can mean you’re matching queries nobody else wants, and there’s usually a reason nobody else wants them.

Conversion rate: the number everyone quotes and nobody defines

Conversion rate has the widest spread and the worst definitional problems. Even the tidy 8.18% average blends industries where it’s 2.64% with industries where it’s over 16%.

Before you compare your rate to anyone’s, answer these questions about your own account:

• What is counted as a conversion? Look at your conversion actions and check which ones are set to “primary” and included in the Conversions column. Page views, button clicks and micro-events inflate the rate and make any comparison meaningless.
• Is each action counted once or every time? For leads, “one” is normally right. For purchases, “every” is normal. If you count every form submission when one person submits three times, your rate overstates.
• Is tracking actually working? An account where the tag fires on half of real conversions will look terrible against a benchmark, and the campaign may be fine. This is where broken enhanced conversions quietly wreck comparisons.
• How long is your conversion window and sales cycle? If your leads take three weeks to become anything, checking a campaign after four days makes the conversion rate look low.

Also, calls. If phone calls from ads count as conversions in some benchmark and not in yours, you’re comparing different things. It’s the same thing that explains why the 7.52% and 1.2% figures from different sources are both quoted, and both defensible.

One published recommendation stuck with me: manage against your own trailing 90 days and use the industry average as a sanity check. I think that’s roughly right, with the caveat that 90 days only works if the tracking was sound during those 90 days. If it wasn’t, the baseline is contaminated and you’d be optimising toward a phantom.

Cost per lead and ROAS: closer to money, still not the answer

Cost per lead is a better yardstick than CTR or CPC because it combines both. The 2026 average of $66.69 is a useful anchor for one specific reason: it lets you see whether you’re in the right order of magnitude. If you’re at $400 and the industry is at $80, something structural is going on. If you’re at $72, your problem is probably not that you’re behind the average.

But cost per lead has a problem of its own, and it’s the reason so many lead-gen accounts stall out at “good CPL, no revenue.” A lead is only as valuable as what happens next. A campaign generating $40 leads that never answer the phone is worse than one generating $110 leads that close. If you can feed real outcomes back into Google, through offline conversion imports or conversion value rules, the account can start optimising toward what matters. Our post on conversion value rules goes into how to tell Smart Bidding which conversions carry weight.

For ecommerce, the equivalent is return on ad spend, and it has the same trap. A ROAS of 4 can be great on a product with 70% margin and a loss on one with 20% margin. Nobody’s benchmark knows your margin. Break-even ROAS is 1 divided by your margin (a 25% margin means break-even at 4.0), and that number is a better target than any industry figure. The choice between target CPA, target ROAS and maximise conversions is worth its own thought, and we broke that down here.

How to build a baseline you can defend

If your own data is the better benchmark, here’s how to build it so it holds up.

Step 1: Fix the definitions first

Open your conversion actions list. Decide which action or actions represent the outcome the business cares about. Mark those as primary. Everything else can stay as a secondary action for observation, but shouldn’t drive bidding or appear in your headline conversion rate. If you change these settings, note the date, because your history before and after won’t be comparable.

Step 2: Segment before you average

Split the account at least three ways before you calculate anything:

• Brand and non-brand. Never blend these. Brand traffic will flatter every metric.
• Campaign type. Search, Performance Max, Shopping and Demand Gen behave differently, and averaging them together produces a number that describes none of them.
• Device and geography, if the volume supports it. Mobile conversion rates often differ from desktop, and a national average can hide a strong metro and a weak one.

The point of segmenting is to end up with several small, honest baselines and not one big, blurry one.

Step 3: Use a window long enough to be stable, and short enough to be current

Ninety days is a decent default for accounts with reasonable volume. For low-volume accounts, a rolling 90 days might contain 12 conversions, and a rate calculated on 12 conversions moves a lot based on two or three of them. We looked at the volume question in how much conversion data Smart Bidding needs, and much of that thinking applies to baselines too. If you don’t have the volume, extend the window or judge by cost per lead over a longer horizon and don’t read weekly swings as signal.

Step 4: Record the spread, not just the average

Look at your weekly conversion rate for the last 13 weeks. Note the lowest week, the highest week and the median. That range is your account’s normal wobble. A week that falls inside it isn’t news. A week that falls well outside it deserves a look.

People skip this step and end up reacting to noise. If your weekly conversion rate normally bounces between 3% and 6%, a 3.4% week isn’t a problem, and a diagnosis at that point will invent a cause that isn’t there.

Step 5: Use the industry benchmark as a smoke alarm

Now bring the external number back in, with a narrow job. If your segmented non-brand conversion rate is a third of your industry’s reported rate, or your CPC is triple, that’s worth an investigation. If you’re within a fairly wide band of the reported numbers, the benchmark has nothing more to tell you, and your own trend is what to watch.

What to do when you really are below benchmark

Sometimes the gap is real, and you’ll want a sequence for chasing it down. The equation from earlier helps: cost per lead is CPC divided by conversion rate, so work out which side is off.

If clicks are too expensive:

• Check which queries you’re paying for. The search term report usually explains a high CPC faster than any setting will.
• Look at match types. Since the definitions have shifted over the years, our match types post is a good refresher on what broad, phrase and exact actually do now.
• Check location and audience settings for expensive waste.
• Only then look at bids and Quality Score components.

If clicks don’t convert:

• Confirm tracking works, by placing a test conversion yourself and watching it arrive.
• Check that the landing page loads fast on a phone and matches the promise of the ad.
• Read the search terms again, this time asking whether the people behind them wanted what you sell.
• Check whether the conversion action is set up sensibly, and whether the sales cycle means you’re judging too soon.

The answer often turns up in the first two items of one of those lists. It’s rarely an exotic setting, and a query mismatch or a tracking problem is a common culprit. That’s a reason to fix the definitions and the search terms before you start rewriting ads.

What to do when you’re above benchmark

It’s tempting to relax when your numbers beat the industry average. Resist that a little.

Being above average can mean your account is excellent. It can also mean you’re capturing only the easiest demand. A campaign that only bids on exact-match brand terms and a couple of high-intent phrases will post a wonderful CTR and conversion rate, and it will also be tiny. It’s leaving growth on the table, and a benchmark will never say so because it measures efficiency and not reach.

If you’re beating the benchmark, check impression share alongside it. High efficiency with low impression share because of budget or rank means you may be able to spend more at an acceptable cost per lead. We wrote about how to tell whether impression share is being lost to budget or rank, which is the natural next question.

Why benchmarks decay, and why your baseline does too

Benchmarks are a snapshot. The WordStream figures cover April 2025 to March 2026, and by the time you read them the auctions have moved on. CPC rose from $2.32 to $5.42 over a decade, so on average it’s trending one way, and any number you saved from a previous year is stale.

Your own baseline decays too, and faster than most people expect. Seasonality shifts it. A competitor entering or leaving shifts it. A change to how Google matches queries, or how it attributes conversions, shifts it in ways that show up as a step change in your charts. If you recorded a baseline in January and you’re comparing to it in October, be aware that you’re comparing two different worlds.

A practical habit: refresh your baseline at a regular interval, and refresh it after any major structural change to the account. Write the date next to the number. A baseline with no date attached is a rumour.

How continuous monitoring helps with benchmarks

The hard part of everything above is not the arithmetic. It’s that a baseline is only useful if someone is looking at it regularly enough to notice when the account leaves its normal range. A weekly conversion rate that drifts from 5% to 3% over five weeks doesn’t announce itself, and it’s easy to miss when you look at the account once a fortnight. That’s the situation Growera’s Google Ads manager is built around: it watches the account every day against its own history, so a change in CPC, conversion rate or spend pacing gets flagged when it happens and not when the monthly report is due. It’s the same idea as the habit we described in how often you should actually check your account, just done consistently, and you can see how it’s priced on the pricing page.

A short list of things to say when someone quotes a benchmark at you

You’ll end up in that meeting eventually. A few responses that are honest and don’t turn into an argument:

• “Which benchmark is that, and what does it count as a conversion?” Often this alone changes the conversation.
• “Is that brand and non-brand blended? Ours is split, and non-brand is the number that matters.”
• “Our own trailing 90 days is X, with a normal range of Y to Z. This week is inside it.”
• “The industry average for cost per lead is a fair sanity check, and we’re within it. What we’re working on is the lead-to-customer rate, which no benchmark covers.”

None of those are excuses. They’re the questions a careful person asks before deciding a number means something.

Summary

The 2026 WordStream benchmarks give an all-industry average of 6.64% CTR, $5.42 CPC, 8.18% conversion rate and $66.69 cost per lead, from more than 13,000 US campaigns between April 2025 and March 2026. They’re a decent sanity check and a poor scorecard.

Different studies report different averages because they define conversions differently, sample different advertisers, and mix brand and non-brand traffic in different proportions. Industry ranges are wide enough (CPC from about $1.63 to nearly $10, conversion rate from about 2.6% to over 16%) that the overall average describes very few real accounts.

The stronger yardstick is your own history, built carefully: fix your conversion definitions first, segment brand from non-brand and campaign type from campaign type, use roughly a 90-day window when volume allows, and record the normal spread and not only the mean. Use the external benchmark to spot a large gap, and use cost per lead or break-even ROAS, tied to what a customer is worth to you, to decide whether the campaign is doing its job.

If you take one thing from this, make it the two-lever view: cost per lead is CPC divided by conversion rate, so any gap is a click-cost problem, a conversion problem, or both. Working out which one you have will get you further than knowing what the average account does.