Facebook ads for small business: what Meta is good at

Two numbers from one publisher, measured the same way in adjacent windows. The average cost per lead on Facebook is $27.66, and on Google it is $66.69. LocaliQ's WordStream team drew the Facebook number from US campaigns that ran between April 2024 and June 2025, and the Google number from 13,474 US search campaigns that ran between April 2025 and March 2026. A Google lead costs about 2.4 times a Facebook lead.

That gap moves budget for the wrong reason. Google sells intent: someone typed "emergency plumber near me" and you paid to be the answer. Meta sells attention, arriving while a person scrolls a friend's kitchen photos. The trade calls these demand capture and demand generation, and the 2.4x is the price of the difference.

Demand capture wins when people already search for what you sell. A roofer after a hailstorm should spend on Google first, because every one of those queries is a person with a wet ceiling. Demand generation wins when the need is real but no search happens yet: a new service line, a second location, or a $12,000 bathroom remodel that no homeowner types into a search box.

The follow-up work differs too. A search lead has chewed on the problem for days, while a Meta lead met you thirty seconds ago and needs more touches before you judge it. Owners who move budget on cost per lead alone, then count closed deals a month later, decide that Meta failed. They bought a different product and graded it on the old rubric. Give the two channels separate budget lines and separate scorecards, as we do in paid media.

The answer starts with what you are buying, because a traffic campaign and a lead campaign price differently, and most benchmark posts quote one without saying which. In the newest measured data, traffic campaigns averaged $0.70 per click at a 1.71% click-through rate. Lead campaigns averaged $1.92 per click, a 2.59% click-through rate, and a 7.72% conversion rate, down from 8.67%. Clicks got cheaper while leads got dearer, and the average cost per lead rose 20.94% to $27.66.

Legal shows the widest gap: $18.17 a lead on Facebook against $131.63 on Google. The spread measures the distance between the two leads, not a discount. Vertical averages also move faster than anyone plans for. Dental practices averaged $32.46 per lead in the data ending April 2024 and $76.71 in the set ending June 2025, same publisher, same method. Anyone who built a case on the lower number was wrong by half within a year. Read the tables as ranges and re-measure quarterly.

The cost curve nobody budgets for

Meta also gets dearer on a schedule. Its average price per ad rose 12% year over year in the first quarter of 2026 and 12% again in the second, while impressions across its apps rose 19% and then 14%. A budget set in 2024 buys much less reach today, and owners who never revisit it read the drop as weak performance. Seasonality sits on top: every online store crowds into the fourth quarter, so November costs rise for reasons unrelated to your business, and January is the cheapest reset you get.

Meta's own help pages name the floor. An ad set becomes learning limited, in Meta's words, when it is unlikely to get about 50 optimization events in the week after your last big edit. Under that line the system never gets enough data to work with, while your money still leaves the account.

Most small businesses spend a fraction of that, so they run hobbled without knowing why. You still pay for every click, and the system never sees enough of them to learn which were worth buying.

Four honest ways out, none of which pretends that $20 a day behaves like $200.

1. Aim at a cheaper event upstream. Landing page views and add-to-carts happen far more often than leads, so 50 a week is reachable.
2. Consolidate. One ad set at $60 a day learns. Four at $15 do not, and they bid against each other while failing.
3. Let Advantage+ pool the budget across placements and audiences instead of splitting it into pieces too small to teach anything.
4. Or change the scorecard. Judge Meta on booked revenue per quarter, since it buys demand over a longer horizon.

Meta has taken most of the targeting choices back. Advantage+, its automated campaign layer, passed a $75 billion annual revenue run rate by the second quarter of 2026. More than 9 million small businesses use at least one of Meta's AI creative tools. Under it sits Andromeda, the system that picks which ads a person sees, which Meta reported lifted recall by 6% and ad quality by 8% on selected segments.

Those are modest, disclosed gains rather than magic, and you control none of them. You control the creative. Nielsen's cross-media work put creative as the largest single driver of ad sales lift, at 65% in its 2006 Project Apollo study. The same work found media's share of the effect climbing from 15% to 36% over eleven years, across nearly 500 packaged-goods campaigns. The input Meta cannot automate is the ad itself.

So the account has one standing job: keep feeding different creative angles into a system that will find the audience for each. Build for the surface you are buying, a vertical video watched with the sound off and the thumb already moving. The organic content you already make is the cheapest source of angles. Meta's early tests on Advantage+ leads campaigns reported a 10% lower cost per qualified lead, a real number and a small one. Automation raises the floor. It does not write the ad.

Instant forms live inside the feed and fill in from the person's profile, so a lead takes two taps. Landing pages ask for a click, a page load, and a form fill, and every step loses people. Cutting the friction also cuts the filtering it was doing.

You will see exact multipliers quoted for this, usually three times cheaper and five times worse. Every version I could trace states a different pair of numbers, and none publishes any data. The direction is sound; the precision is invented. Meta itself ships a higher-intent version of the instant form, which tells you the trade is real.

Your follow-up capacity decides it. If someone calls or texts every new lead within minutes, while the ad is still fresh, take the volume. If leads sit in a spreadsheet until Thursday, buy the filtering, because a cold lead answered three days late closes at close to zero. Automated routing fixes that gap cheaply, and it is the plumbing our AI automation work spends most of its time on.

If you send paid traffic to a page, that page is part of your cost per lead, and a slow one raises the price of every click you bought. Our free SEO report grades any URL on speed and page structure in about thirty seconds, which is a cheap check before you spend.

Retarget before you prospect. People who visited your site or bought from you convert at the lowest cost on Meta, and they are the one warm pool on a cold platform. They also ease the learning problem, throwing off events faster per dollar than any cold set.

Then fix the signal. Apple's App Tracking Transparency prompt broke click-level tracking in 2021, and Meta's then finance chief put the cost to Meta at roughly $10 billion for 2022. Opt-in has since climbed back to about 38% worldwide from roughly 16% at launch, and Meta rebuilt the rest with modeling. The durable fix on your side is the Conversions API, which sends sale and lead events from your server, not a browser.

Build lookalike audiences from a real customer list, not from everyone who visited a page. A seed of paying customers is one of the few places your own data beats Meta's model. Then spend twenty minutes in the Meta Ad Library before writing anything. That warm-first order drives local ads that convert and our paid work in the trades.

The tools, with honest one-liners

• Meta Ads Manager and Advantage+. Free. Advantage+ earns its keep when you cannot feed a manual ad set 50 events a week.
• Conversions API. Server-side events. Plumbing that raises the ceiling on every campaign after it.
• Canva and Meta's built-in creative tools. Static and short-video ads without a production budget.
• A CRM with a lead-ads sync, or a Zapier or Make connection. A form fill reaches a phone in seconds.
• GA4 with tagged links. The referee when Meta's reported conversions and your CRM disagree, which they will.

Half the pages ranking for this topic quote numbers that fall apart on inspection. The newest measured Facebook data ends in June 2025, so any table headed "2026 Facebook benchmarks by industry" is a forecast. One publisher, TheeDigital, says so and labels its rows as projections.

• "Advantage+ cuts cost per acquisition 32%." Repeated on dozens of agency pages, documented on no Meta property. Meta stands behind 10% lower cost per qualified lead.
• "83% of people discover products on Instagram." Traces through agency blogs to a study nobody can retrieve today.
• "Lead forms are 3x cheaper and 5x worse." Each version states different multipliers and none publishes a dataset. The decimals are decoration.
• "Run 15 to 25 creatives per ad set." Sourced to marketing blogs with no data behind them.

Quote any of these to a client and you set the bar against folklore, which is how accounts get lost at renewal.

Run it in this order and give it 90 days. Most accounts that fail on Meta failed at step two or four, long before anyone looked at the creative.

1. 1

   Name the job before you name the budget

   Meta generates demand rather than capturing it, so the metric is booked revenue over a quarter. Write it down before you spend.
2. 2

   Fix the signal first

   Install the pixel and the Conversions API before you spend a dollar. Retrofitting after three flat months wastes those months.
3. 3

   Start warm

   Run retargeting and a customer-list lookalike first, then add cold prospecting once the warm pool converts.
4. 4

   Pick an event you can afford 50 of

   Choose the event your budget can produce about 50 times a week, and consolidate to one or two ad sets so the money pools.
5. 5

   Ship angles, not sizes

   Put three to five different creative angles into the account rather than three sizes of one, and swap the weakest every fortnight.
6. 6

   Answer inside five minutes

   Route every instant-form lead straight to a phone, because a lead answered three days late closes at close to zero.
7. 7

   Judge it in the CRM, then budget for the curve

   Read closed revenue against total spend at 90 days, then set the next quarter about 12% higher, because Meta's price per ad rose that much in each of the past two quarters.

Ninety days in, the number that settles it sits in your CRM rather than Ads Manager: closed revenue against total spend, with Meta judged on the deals nobody would have searched for.