Why Meta brings cheaper qualified leads than Google, and why Meta alone fails

LEV BROVTSEV · OCTOBER 20, 2026

Meta often wins on cost per qualified lead because you pay for attention, usually per thousand impressions, not for a contested high-intent click, and its algorithm finds people before they search. But Meta can't replace search. Meta creates demand; Google and Bing capture it. Meta's volume, policy and data limits cap how far you can scale on it alone.

Every owner who compares Facebook ads vs Google ads in a channel report eventually asks the same question. Meta's cost per lead is a fraction of Google's, so why not move everything there? It's a fair question, and in one account I run the answer on cost alone was even stronger than the cost-per-lead report suggested. Here is what the numbers showed, why Meta wins on that metric, and why I still keep search at the center of the budget.

The number from one account

In a luxury US healthcare clinic I run paid acquisition for, Meta was about 3× cheaper per verified patient than Google Ads over the same period. Indexed, with Google Ads at 100, Meta came in at about 33.

How it was counted matters more than the ratio. Verified patients came from the clinic's CRM, not from Meta Ads Manager or the Google Ads interface. Each verified patient was matched to the click ID or the Meta lead ID it arrived with, and spend and verified patients covered the same period. Platform-reported conversions played no part in the comparison. If your CRM doesn't record where each lead came from, that has to come first; the setup is in lead source tracking.

And I still wouldn't move the Google budget to Meta. The rest of this note is why.

Why Meta's qualified leads are cheaper: five reasons

1. You're buying attention, not a contested click

In Google Ads you bid on a search. Every clinic, aggregator and lead seller in your market wants the same query, and healthcare terms sit near the top of what anyone pays per click. Published tables are rough, and I explain why in Google Ads benchmarks for healthcare, but nobody disputes the direction. On Meta you usually pay per thousand impressions, in an auction shared with every consumer brand on the platform. The people you want aren't being fought over the way a high-intent search is.

2. You reach people before they compare

By the time someone types a treatment and "cost" into Google, they have usually opened several tabs. They're comparing you on price and convenience. Meta reaches them earlier, while the idea is still forming, and your practice is often the first one they take seriously. Fewer comparison shoppers means more of the leads that remain are worth a call.

3. The creative does the qualifying

A search ad has a few lines of text. A Meta ad can be a video of the founder explaining who the treatment is for, what it costs to start, and who shouldn't book. People who aren't a fit scroll past. That filtering happens before the click, which is why a lead from a well-made video ad tends to be further along than a lead from a short search ad.

4. Lead forms remove friction

Meta's instant forms let someone ask for a call without leaving the app or loading your site. That lowers cost per lead sharply. It also lets in people who tapped by accident or don't remember asking, which is the first limit below.

5. You can feed it outcomes, within limits

Meta's algorithm finds more of whatever you tell it counts. Send it form fills and it finds form fillers. Send it CRM outcomes through the Conversions API, a qualified or booked stage with a neutral event name, and it can learn from people who actually become patients. In this account that feedback is a large part of why the cost per verified patient held.

Health advertisers have less room here than other businesses, and it's worth checking your own status before you plan around it. Meta's Business Tools Terms don't allow data that includes or is based on health information, and event and audience names can't imply it. For websites it classifies under health and wellness, Meta may restrict specific mid- and lower-funnel events or fully restrict event sharing (Meta Business Help: data sharing restrictions). In practice that means: no condition or treatment in event names, parameters or URLs, a CRM ID instead of personal details, and a look at your dataset in Events Manager to see which events Meta still lets you optimize toward.

Why you can't run on Meta alone: six reasons

1. There's a volume ceiling

The first part of a Meta budget reaches the most responsive people in your audience. Each extra dollar reaches people who are a little less likely to act, and the same people see your ads more often. Frequency climbs, the creative wears out, and the cost of the next qualified lead rises even while the average still looks good.

That gap between average and marginal cost is the trap. The report shows the average, so the channel looks cheap right up to the point where the last part of the budget is buying leads at several times the headline rate.

2. Search captures the demand Meta creates

Many people who see a Meta ad don't tap it. They search the clinic's name or the treatment a day later on Google or Bing, click a search ad or an organic result, and become a "search" patient. Turn search off and part of Meta's apparent performance goes with it, because those people now find a competitor's ad for your name or for the treatment.

You can see this effect in branded search: when Meta spend goes up, searches for the practice name tend to follow. The only clean way to measure it is a holdout. Pause Meta in a set of comparable regions for a few weeks, or run Meta's own lift study, and compare verified patients from all channels against the regions where it kept running.

3. Lead quality has a wider spread

Instant forms bring more junk: wrong numbers, people who don't remember the ad, people who wanted information and not a call. Meta's cost per lead holds only if intake calls back within minutes, because these leads cool faster than a searcher who was actively shopping. If your front desk calls back the next morning, Meta's advantage disappears. What every minute costs in high-ticket covers the intake side.

4. Policy and account risk

Healthcare ads on Meta run under its health and wellness ad standards. Ads can't imply that the viewer has a condition, and some categories need LegitScript certification before either Google or Meta will run them. Disapprovals and restricted accounts happen even when the ads are compliant. One channel means one point of failure: a single review can stop all new patients for weeks. In my own experience, account bans stopped once I became an official Meta partner. That's what happened on the accounts I run, not a guarantee.

5. Meta's own reporting flatters Meta

Meta Ads Manager counts conversions that happened after someone only saw an ad, within its attribution window. Google Ads counts by click. Put the two dashboards side by side and Meta looks better than it is, because it's claiming some patients who also clicked a search ad. The only fair comparison is the one above: verified patients in your CRM, each matched to the click ID or lead ID it came with.

6. The highest intent is still on search

Someone typing the treatment plus "near me", "cost" or "financing" has already decided they want it. They convert faster and with fewer follow-ups than almost anyone Meta can reach. Those searches happen on Google and Bing, and if you're not there, a competitor is. Microsoft Ads deserves a place too: an older, higher-income audience, cheaper clicks and smaller volume, as I describe in when Microsoft Ads outperforms Google for high-ticket.

How I split the budget instead

I don't split by channel preference or by average cost per lead. I split by the marginal cost per verified patient in each channel: what the last part of each budget is buying, measured in the CRM, re-checked every month.

1. Search first, for capture. Google and Bing get budget up to the point where the next increment costs noticeably more per verified patient than the rest. Beyond that point, extra search spend mostly buys the same people at a higher price.
2. Meta for demand creation and retargeting. Prospecting with creative that qualifies, plus retargeting people who visited but didn't book. Budget goes up while its marginal cost per verified patient stays below search's.
3. One holdout test a quarter. Pause one channel in comparable regions, or run a lift study, and see how many verified patients actually disappear. That number, not the dashboard, decides next quarter's split.

How I run the quarterly holdout. It doesn't need a data team, just discipline:

1. Pick two groups of comparable regions, by past verified patients per month, not by population.
2. Pause the channel you're testing in one group for four to six weeks. Leave everything else unchanged, including the other channels' budgets.
3. Count verified patients from all sources in both groups, in the CRM, for the test period and the same period before it.
4. The difference between the groups is roughly what that channel adds. If patients barely drop where it was paused, the channel was mostly collecting people who would have come anyway.

For a single-location practice, a geographic split isn't possible. Meta's own lift study, which holds out a random share of the audience, is the practical alternative there, where your account qualifies for one. Either way, the answer comes from the CRM, not from the platform that is being tested.

All of this depends on counting patients, not leads; why that matters is the subject of cost per patient vs cost per lead. The CRM outcomes that feed Meta's Conversions API are the same ones you send back to Google through offline conversion tracking, so one setup serves both platforms.

The account this comparison comes from is the same one in my flagship case: an inherited account rebuilt around verified patients. How I run each channel is on the Google Ads, Meta Ads and Microsoft Ads pages.

Diagnosis before treatment.

Written by Lev Brovtsev, healthcare performance marketing consultant. I do the work myself. Last updated: October 2026.