Meta for service businesses: cost per signed project, not ROAS

Meta's ads manager is built around ecommerce habits: purchases, average order value, return on ad spend. If you sell website builds, retainers, or other high-consideration services, those metrics will flatter the wrong behavior.

A cheap lead is not a cheap client. A high "conversion" rate on an instant form is not a pipeline. The number that matters is what you paid to get a signed project — and whether that project was one you actually wanted.

What usually goes wrong

Traffic lands on the homepage, or in a Meta instant form, instead of a page that exists to book one conversation.

The site asks for too many things at once: a contact form, a chat widget, a phone number, a "see our work" loop, and a vague "get a quote." A visitor who could have booked a planning session instead bounces between options and leaves.

Leads come in cheap and unqualified. You spend the next week discovering they have no timeline, no budget, or no real project — which means the cost per signed job is hiding inside all those polite discovery calls.

Proof is thin. One review is doing too much work, and it sits somewhere the ad click never sees.

And the account is being judged like a store. People look for ROAS on a service that does not check out in one session, then conclude Meta "does not work" — or they increase spend before the offer, the page, and the tracking are ready.

Get the economics straight before you spend

Before you ask Meta to find you clients, you should be able to say:

• What you sell, in packages a stranger can understand
• What a good-fit project is worth, and what you can afford to pay to win one
• What close rate you actually see from a qualified conversation to a signed yes

If the margin or the close rate cannot support paid acquisition yet, a smaller daily budget plus retargeting, referrals, and a strong Google Business Profile will beat pouring money into cold traffic. Spend does not fix an offer that is not ready.

One conversion, not five

Pick a single next step. For most service businesses I work with, that is a free planning session — not "contact us," not "download everything," not an instant form that asks for a name and a dream.

Put prices, or at least a clear starting range, on the page. Surprise quotes waste everyone's time and attract people who were never going to buy.

The form should qualify, not just collect. Useful fields:

• Type of business
• Timeline
• What they have today (no site, an old site, a store that is not converting)
• A budget band
• How they found you

You will get fewer leads. You will get better ones. That is the point.

Send people to a page that does one job

The homepage has other jobs: who you are, what you have made, why someone should trust you. An ad click should not have to go hunting.

A dedicated landing page should:

• Repeat the promise in the ad
• Show proof next to the call to action — recent before-and-afters, not a testimonial buried in a footer
• Lead to that one conversion
• Carry the same Pixel and Conversions API events you will optimize for (lead, schedule, or a custom event that means "qualified inquiry," not a generic page view)

How the page is built matters less than whether the click, the story, and the form all agree.

Creative that talks like a salesperson, not a banner

You do not need a dozen campaigns. You need a few clear angles — the problem you solve, the outcome you deliver, the proof you can show — and a couple of variants of each. Three angles times two variants is enough to learn without splitting the account into pieces.

If you are on camera, or your team is, script to those angles. Film what you can actually say with a straight face. Edit them short. The page behind the ad should finish the sentence the ad started.

Tell Meta what happened after the form

A form fill is an early event. The valuable ones come later: this lead was unqualified; this one was qualified; this one booked the call; this one signed.

When those stages feed back into Meta — through your CRM and the Conversions API — the system can stop optimizing for people who fill out forms and start looking for people who look like the ones who hire you.

Until that loop exists, treat cost per lead as a diagnostic, not a win.

The scoreboard

Watch:

• Cost per qualified conversation
• Cost per booked call
• Cost per signed project
• Close rate from call to contract

ROAS can still be a back-of-napkin check (revenue from signed work versus spend), but it is a trailing story, not the thing you ask the algorithm to maximize day to day. Service work closes on a delay. Optimizing as if it were a $40 checkout will send you the cheapest possible "yes" — which is usually a no in disguise.

If Meta is generating interest and not generating clients, do not start by making more ads. Start by asking whether the click had one place to go, whether the form told you enough, and whether you are measuring the signature or the inquiry.