How to send qualified leads back to Meta

Meta's delivery system is a pattern-matching machine. It observes which people produced the event you told it to value, then goes and finds more people like them. This is why the quality of your Meta leads is, to a large extent, a description of the signal you have been sending.

If the event is "submitted a lead form", the pattern it learns is "people who submit lead forms" — a real behavioural trait, and not one correlated with booking a ₹2 lakh holiday. Changing the signal changes the audience.

The three layers of signal

A complete travel setup sends Meta three progressively deeper signals:

1. Lead — the enquiry was submitted. Fast, high volume, shallow.
2. Qualified Lead — your sales team confirmed it is workable. Slower, lower volume, far more meaningful.
3. Purchase / Booking — money received, with the actual value attached. Slowest, lowest volume, most valuable.

Most accounts send only the first. The work is in adding the second and third without breaking the first.

Step 1 — Get the Conversions API working properly

Browser-based pixel events lose a meaningful share of data to tracking prevention, ad blockers and in-app browser behaviour. The Conversions API sends the same events server-to-server, which recovers much of that loss.

Two things must be right or you will make matters worse:

• Deduplication. Every event sent from both browser and server needs a matching event_id and event_name, or Meta counts it twice and your reporting becomes fiction.
• Event match quality. Send as many hashed identifiers as you legitimately hold — email, phone, name, city, and the fbc/fbp cookie values. Match quality directly determines how much of your data Meta can actually use.

Step 2 — Capture the Meta click identifiers

When someone clicks a Meta ad to your site, an fbclid parameter appears in the URL and the pixel converts it into an fbc cookie. Store both fbc and fbp against the lead in your CRM at the point of enquiry, exactly as you would store a GCLID for Google.

For leads that arrive through instant forms rather than your website there is no cookie. Those leads carry a lead_id instead, which is what you will use for offline attribution.

Step 3 — Send the qualification event

When your sales team marks a lead as qualified, fire a server-side event back to Meta. Use a custom event name — QualifiedLead works well — and include the stored identifiers plus the original event time.

For leads that came from instant forms, use an offline event set or the Conversions API for lead ads, keyed on the lead_id. Meta matches it back to the original ad.

The practical timing question is when to send. Sending too early — immediately on any sales contact — defeats the purpose. Sending too late means the optimisation window has closed. A useful rule is to send as soon as a human has made a judgement, typically within 24 to 72 hours.

Step 4 — Send the booking with its value

When a booking is confirmed, send a Purchase event with the actual value and currency. This is what makes value-based optimisation and ROAS bidding possible, and it is the point at which Meta can distinguish a ₹40,000 domestic booking from a ₹4,00,000 European tour.

Travel booking values vary by an order of magnitude within the same account. An algorithm optimising towards undifferentiated conversions will happily find you a hundred cheap bookings instead of ten profitable ones.

Step 5 — Change what the campaigns optimise for

None of the above changes anything until you point the campaigns at the new events, and that requires patience with volume.

• Meta needs roughly 50 conversion events per ad set per week to exit the learning phase reliably.
• If QualifiedLead produces less than that, keep optimising towards Lead and use the qualification data for reporting and audience building only.
• As volume grows, move the optimisation event deeper. Many travel accounts run prospecting on Lead and retargeting on QualifiedLead, which is a reasonable compromise.

The audience dividend

Even where volume never supports optimising towards qualified leads directly, the data has a second use that is often more valuable: building better source audiences.

A lookalike built from your qualified leads, or better still from your actual bookers segmented by value, behaves quite differently from one built from every lead you have ever received. The second list contains your bad half. The lookalike will faithfully reproduce it.

Segment further where you can — bookers by destination type, by value band, by group size. A lookalike from high-value honeymoon bookers is a genuinely different audience from one built on domestic weekend trips, and treating them as one list averages away the advantage.

What to watch afterwards

Track three things in the weeks after implementation: event match quality, the ratio of qualified leads to total leads by campaign, and whether cost per qualified lead is moving in the right direction. The middle number is the one that tells you whether the feedback loop is actually influencing delivery, and it usually takes four to six weeks before it does.