Why Your Google Ads Cost Per Lead Keeps Climbing

Rising cost per lead is the most common complaint we hear about Google Ads, and the most common explanation offered for it — everything is more competitive now — is usually wrong.

The benchmark data argues against it. Across 13,474 US search campaigns running between April 2025 and March 2026, the median cost per lead was $66.69, the median cost per click $5.42, and for the first time in five years the all-industry cost per lead went down rather than up. Legal remains the most expensive category at $131.63 per lead, arts and entertainment the cheapest at $26.84. If your account is moving the other way against that backdrop, something inside the account is doing it.

Here are the six causes, roughly in order of how often they turn out to be the answer.

1. You are optimising towards the wrong conversion

This is the big one, and it is nearly always present in accounts where cost per lead has drifted upward over months rather than jumping.

Smart Bidding gets you more of whatever you tell it a conversion is. If your conversion action is a form submission, the algorithm will find people who fill in forms — including the ones who never answer the phone afterwards. If you count newsletter signups and quote requests as the same event, you have asked Google to treat them as equally valuable, and it will obligingly buy the cheaper one.

The fix is unglamorous and effective: define conversion actions that correspond to money. Feed qualified-lead and closed-won status back into the account through offline conversion imports or enhanced conversions, and set values that reflect reality. An account bidding towards revenue behaves differently within weeks from one bidding towards form fills.

2. Broad match has quietly widened your audience

Broad match with Smart Bidding works — when the conversion signal is good. When it is not, it is an efficient way to spend a budget on adjacent intent. The symptom is a cost per lead that rises slowly while impressions rise quickly.

Check the search terms report against spend, not against clicks. You are looking for the queries that consume budget without producing qualified outcomes, and for the drift that happens when a campaign starts matching a category rather than a need — searches for jobs, DIY instructions, free templates and competitor research all convert at some rate, and none of them are your customer.

3. Performance Max is taking credit for demand you already had

Performance Max is genuinely good at finding incremental demand and genuinely capable of cannibalising your brand traffic while reporting it as a triumph. The 2026 updates help here: Google added first-party audience exclusions so you can push a campaign towards new customers rather than existing ones, network segmentation in placement reporting so you can see where ads actually served, and budget reporting with end-of-month spend projections.

Use the exclusions. Exclude your customer list where the goal is acquisition, keep brand terms in a dedicated search campaign so their cheap conversions are not inflating a mixed campaign's apparent efficiency, and check placement reporting for the display inventory that eats impressions without producing anything.

4. Your landing page decayed while your ads improved

Cost per lead has two halves and most teams only manage one. If click cost holds steady and cost per lead rises, the conversion rate fell — and the page is the usual culprit.

Landing pages decay in predictable ways. A form grows two fields because someone in operations wanted them. A consent banner starts covering the call to action on smaller screens. A hero image gets replaced with a heavier one and mobile load time slips. Any of these can cost a fifth of your conversion rate without a single alert firing, which is why we treat conversion rate optimisation and paid media as one budget line rather than two.

5. Your ad-to-page match broke

Every account accumulates ads pointing at pages that no longer say what the ad promised. A product got renamed. A campaign was cloned for a new service and the URL was not changed. An offer expired and the page was updated but the headline was not.

Audit it directly. Sort by spend, open the top twenty ads and their landing pages side by side, and ask whether the page delivers the promise within the first screen. It is the least sophisticated diagnostic in the account and it routinely finds five figures of annual waste.

6. You are counting leads that were never new

Repeat submissions from existing customers, staff testing the form, spam from bots, and the same person filling in three different forms all count as conversions. Bot traffic in particular has grown enough to distort small accounts entirely.

Deduplicate by email or phone number before you report cost per lead, add a honeypot field and rate limiting to your forms, and exclude your own network at the analytics level. Cleaning this up sometimes raises the reported CPL — the number gets worse and truer at the same time, which is the point.

The diagnostic order

Work through it in this sequence, because each step changes the data the next one depends on.

1. Clean the conversion definition. Deduplicate, remove soft conversions from bidding, get qualified status flowing back in.
2. Re-baseline. Give the account two to three weeks before judging anything; bidding models need time to relearn against new signals.
3. Audit search terms by spend. Add negatives where intent is genuinely wrong, not merely unfamiliar.
4. Separate brand from non-brand, and exclude existing customers from acquisition campaigns.
5. Fix the page. Field count, mobile layout, load time, promise match.
6. Only then adjust budgets and bids. Bidding changes made on top of bad data produce fast, confident movement in the wrong direction.

Questions we get asked

Is a rising cost per lead always bad? No. If you have deliberately moved up-market, a higher cost per lead against a higher close rate and a larger contract value is a better business, not a worse account. Judge the campaign on cost per qualified opportunity and on revenue, never on CPL alone.

How long should I wait before judging a change? Two to three weeks for a bidding or conversion-signal change, and long enough to accumulate a meaningful number of conversions — a fortnight with nine conversions tells you almost nothing.

Should I move budget to another channel instead? Sometimes, but not before the account is measuring properly. Moving a budget away from a channel you are mismeasuring simply relocates the problem, and paid search remains the highest-intent inventory available for most B2B and considered purchases.

Does a higher Quality Score fix this? It helps click cost, not lead quality. Relevance work is worth doing, and it will not rescue an account bidding towards the wrong conversion.

What this looks like when it is working

The report that ends the argument

Most disputes about paid search are disputes about which number is real. Replace the standard platform report with four lines and the conversation changes.

• Cost per qualified opportunity, sourced from the CRM rather than the ad platform.
• Qualified rate by campaign, which exposes the campaigns buying volume at the expense of fit.
• Revenue and pipeline by campaign, on the sales cycle rather than the reporting month — a 90-day cycle reported monthly always looks like a failure.
• Spend against non-brand only, so brand conversions are not flattering the account.

Getting there requires the CRM and the ad account to be connected, which is a week of unglamorous plumbing and the highest-return week available in most paid media programmes.

A healthy account has a boring shape: brand and non-brand separated, one clearly defined primary conversion that corresponds to revenue, a negative keyword list that grows every month, landing pages that are checked when campaigns change, and a report that answers cost per qualified opportunity rather than cost per form fill.

That is most of what our pay-per-click management work consists of, and the reason it usually pairs with conversion rate optimisation and lead nurturing — because the cheapest lead in the account is the one you already paid for and never followed up.