Real estate · Lead quality · CRM
Real-estate lead quality: from MQL fluff to compound-walking buyers
Meta will happily deliver you 400 leads at EGP 80 CPL if you ask for raw form submissions. Most of those leads are 19-year-olds curious about prices, single-name junk entries, and serial form-fillers. By the time your call centre has tried each one, your real cost-per-qualified-lead is closer to EGP 600, and the campaign you congratulated yourself for is quietly losing money.
The fix is not "tell Meta to optimise harder." It's changing what Meta is optimising for — and that means feeding lead-quality signal from the CRM back to Meta as a custom conversion event. Here are the four signals that actually correlate with closed deals in real estate, ranked by how clean a signal each one is.
1. Call-back rate within 48 hours (cleanest signal)
This is binary, fast, and impossible to fake: did the lead pick up when sales called back? Most of your junk leads silently fail this check — wrong number, no answer, hung up. Real buyers answer.
Wire it: tag every lead in the CRM with answered_callback: true / false within 48 hours of submission. Push the answered cohort back to Meta as a "QualifiedLead" custom event. After ~50 events Meta starts optimising the auction toward people who actually answer their phones — typically a 30–45% improvement in qualified-lead-cost within 2–3 weeks.
2. Site-visit booking (highest-intent signal)
If the call-back was answered, did the lead agree to a site visit or sales-office meeting? This is the strongest leading indicator we've found for actual close. Less than 10% of raw leads typically pass this filter, but the 10% that do close at 4–6× the rate of the rest.
Wire it: push booked_visit: true to Meta as a separate, higher-value custom event. Use this as the optimisation target for top-of-funnel campaigns at scale — Meta gets a much cleaner training signal than raw "Lead" gives it.
3. Payment-plan inquiry depth
How specifically did the lead ask about payment? "What are your prices?" is generic. "Do you have a 7-year plan with 5% down for the corner units?" is a buyer doing diligence. Tag the latter in disposition notes; aggregate weekly into a quality-score column.
This one is more art than science — it requires the sales rep to populate a free-text disposition field every call — but on accounts where it's enforced, the correlation with close rate is strong enough to justify treating "depth_inquiry: true" as a Meta event in its own right.
4. Repeat name / phone appearance across forms
A lead that filled the same form twice in three days, or filled forms for two of your projects, is a buyer in the market — not a tire-kicker. Your CRM can flag this automatically with a simple UNIQUE check.
Repeat appearance is a weaker signal than the first three — some of these are spambots or affiliate scrapers — but it's free to compute and a useful tiebreaker when you're scoring borderline leads.
What NOT to feed back
- "Lead source" as a quality signal. Every source produces a mix of quality and junk; using source as a proxy will trap you in the cohort you already know.
- Email-open rate. Real-estate buyers don't open follow-up emails — they pick up the phone. Email engagement is a tech-savvy proxy, not a buyer proxy.
- Form-completion time. Counter-intuitively, our fastest form-completers convert at a higher rate than the careful 4-minute fill — sub-15-second completions are often returning buyers, not bots.
The 6-month picture
"You're not buying leads — you're buying a probability distribution. Cleaner signal in, cleaner distribution out."
Accounts that wire the call-back-rate + site-visit-booking feedback loop typically see, over 6 months: blended CPL up 20–30% (because Meta is narrowing targeting), qualified-lead cost down 35–55%, and close-rate per lead doubled. The CFO loves the second two numbers. The first one is the price.
If your dashboard can't track lead disposition end-to-end and feed the right cohort back to Meta as a conversion event, you're optimising on noise. Fix that before you fix anything else.