Ad management

The budget was going out,but which ad worked was invisible.

We built a Meta campaign for the same real estate team. The first ten days were expensive, and we saw that from the numbers rather than a hunch; because the measurement chain was already in place, the effect of every change could be measured.

Talk about a build like this Meta and Google Ads management

Sector

Real estate advertising

Built

Campaign, form and measurement chain

Period

29 August to 14 September 2026

Status

Campaign closed

Starting point

What the situation wasbefore the build.

The first setup opened with a high-intent form and ran expensively. Because measurement was in place, where the problem sat was visible in the first week.

In the first ten days 6,944 TL was spent and three records came in.
On the villa campaign, 271 form opens produced not a single record.
Meta's optimisation never learned: a lead objective wants fifty conversions a week, and there were three in total.

What we built

The pieces builtfor Kali ad campaign.

01
Measurement was built first
The pixel, server-side event sending and record matching were connected before the campaign started. The numbers in Meta, in the server log and in the client record matched exactly; no record was lost.
02
The form structure changed
The high-intent form was removed and a price list offer put in its place. The old forms produced no records across 139 opens; the new form produced three across 141.
03
Audience and placement were measured
The villa campaign ran on Reels and Story, aimed at ages 35 to 54. Narrowing to a single placement was tried and reversed: cost per thousand impressions rose 28 per cent and no records came in.
04
Records written into the sales system
Every record landed in the client database, tagged with the campaign, ad set and ad id it came from.

Measured result

Same budget,a third of the cost.

What saved the campaign was not a creative change; it was having the measurement in place. Because it was visible every week which form brought what from which audience, the result of every attempt could be measured.

1,767 → 587 TL

Cost per record

A 67 per cent improvement, in the same campaign

2,927 → 518 TL

On the villa campaign

After the form and audience change

Records lost

Meta, server and client record matched exactly

Measurement was built before the campaign, not added afterwards.
The failed attempts are on record too: the single-placement Reels test was reversed.
Every weekly budget decision came out of a number.

Built with