Open any Meta Ads guide, watch any course, read any thread. Almost all of it is written by and for ecommerce advertisers. That is not a conspiracy — it is a data problem. Online stores generate huge volumes of instantly measurable conversions, which makes them the easiest environment to learn in and to teach from.
The trouble is that lead generation violates nearly every assumption underpinning that advice. If you run a property, health, finance, trades or hospitality business and your Meta results feel worse than the case studies promise, this is usually why.
Difference one: the conversion event is not the outcome
In ecommerce, a purchase is the outcome. Money changes hands, the value is known, and the platform can optimise directly toward it. In lead generation, the conversion event is a form submission — a stranger typing a name into a box. Its value ranges from zero to tens of thousands of dollars, and most of the time it is closer to zero.
Optimise for that event and Meta will do exactly what you asked: find the people most likely to fill in forms. Those are not the same people as the ones most likely to buy, and the gap grows the harder the algorithm works.
Meta is extremely good at getting you what you optimise for. In lead gen, that is the whole problem.
Difference two: signal volume is a fraction of ecommerce
Advice about broad targeting, consolidated campaigns and letting the algorithm learn assumes hundreds of weekly conversions per ad set. A services business generating forty leads a month is nowhere near that threshold. Learning phases never complete, results swing wildly, and every recommendation about rapid testing becomes statistically meaningless.
- Consolidate aggressively — fewer campaigns and ad sets so what signal exists is concentrated.
- Use upper-funnel optimisation events, such as landing page views or engaged sessions, only as a bridge while volume is thin.
- Judge tests over weeks, not days, and accept fewer simultaneous experiments.
- Do not restructure the account every fortnight; instability is more expensive than imperfect structure.
Difference three: the sales cycle is long and human
An ecommerce conversion happens in a session. A buyer's agency engagement, a mortgage application or a health service enrolment happens over weeks or months and involves phone calls, appointments and trust. Attribution windows of seven days simply cannot see it, and dashboards inside Meta will always understate the channel's contribution while overstating the value of the wrong leads.
Which means the truth about Meta performance lives in your CRM, not in Ads Manager. If those two systems disagree, believe the CRM.
Difference four: quality is the constraint, not volume
In ecommerce, more traffic at an acceptable ROAS is almost always good. In lead gen, a flood of poor leads has a real cost: reps waste hours, response time to good leads slows, morale drops, and the sales team stops trusting marketing. Cheaper leads can genuinely make the business worse.
What to do instead
Feed real outcomes back to Meta
Use the Conversions API to send server-side events for qualified lead, appointment booked and closed-won directly from your CRM, with values attached. This is the single highest-impact change available to a lead-gen advertiser. Once Meta optimises toward qualified leads rather than raw submissions, cost per lead usually rises and cost per customer falls — which is the correct trade.
Design the offer, not just the ad
Most lead-gen underperformance is an offer problem dressed up as a creative problem. Ask what a stranger three months from purchase actually wants. Often it is not a consultation: it is a suburb report, a rate comparison, a pricing guide, an eligibility check or a five-minute assessment. Match offer to readiness and both conversion rate and quality improve.
Make creative qualify
- State price ranges, service areas and eligibility in the ad rather than hiding them for the landing page.
- Show the actual service and real people, not stock optimism.
- Use founder and staff video — in service categories it consistently outperforms polished brand film.
- Test messaging angles, not micro-variations of the same claim.
Use instant forms carefully
Meta instant forms lower friction and lower intent simultaneously. They can work well when paired with review screens, a qualifying question or two, and a callback within minutes. Without fast follow-up they generate a list of people who do not remember enquiring. If your team cannot ring within the hour, send traffic to a landing page instead.
Build the follow-up before you build the ads
Lead-gen results are produced as much by the twenty minutes after the click as by the media. Instant routing, an immediate acknowledgement, multiple contact attempts over two weeks, and a nurture track for those not ready yet. Media buying cannot outrun a broken follow-up process.
Benchmarks worth using
Ignore cost-per-lead benchmarks pulled from other industries. The metrics that matter are cost per qualified lead, qualified-lead rate, cost per appointment, appointment-to-sale rate and cost per acquisition against lifetime value. Measure them in your CRM, segmented by campaign, and review monthly rather than daily.
Meta is an excellent lead generation channel for Australian service businesses. It just requires you to stop taking advice from people selling a fundamentally different kind of thing.
Written by
Erica Marthy Rivero
Performance Marketing Specialist
Erica runs day-to-day media buying across Meta and Google, obsessing over creative testing cadence, audience structure and the cost per qualified lead behind every dollar of spend.