PROBLEM
Budget was leaking across generic campaigns and invisible phone calls
The dealership was spending across search campaigns that grouped dozens of models into broad ad sets, while most high-intent conversions happened by phone and never reached analytics. With 300 cars in stock changing weekly, ads often promoted outdated offers and the sales team could not tell which campaigns generated real buyers.
✕CPL had climbed to ₸34,000 with large swings week to week;
✕300 cars in stock were not connected to live ad feeds or model-level campaigns;
✕Over 50% of leads came by phone, but calls were not attributed to ads;
✕Sales saw leads in CRM, but had no source-level view of which ads sold cars.
SOLUTION · 4 MONTHS
Rebuild the account around inventory, intent and closed-loop attribution
We replaced campaign sprawl with a feed-driven structure for available cars, separated high-intent model queries from broad research traffic, and connected calls and CRM outcomes back to ad platforms. The goal was not just cheaper leads, but clear visibility into which campaigns produced booked test drives and sold vehicles.
WEEKS 1–2
Tracking and attribution base
Implemented call tracking, GA4 events, CRM source mapping and offline conversion imports so calls, forms, test-drive bookings and sales could be tied back to Google Ads and Yandex Direct.
WEEKS 3–4
Campaign rebuild by intent
Split campaigns by brand, model, used vs new, financing and dealer-intent queries; removed wasteful broad traffic and rebuilt ad copy around price, monthly payment and stock availability.
MONTHS 2–3
Feed-based inventory ads
Connected a live vehicle feed for 300 cars, launched inventory-driven ads with model, year, price and availability, and synchronized landing pages with in-stock units to reduce dead-click traffic.
MONTHS 3–4
Optimization by sales quality
Shifted bidding from raw leads to qualified calls and test-drive bookings, excluded low-value search terms, reallocated budget to top-selling models and built weekly dashboards for marketing and sales.
RESULT · 4 MONTHS
Paid traffic became measurable, cheaper and closer to actual sales
| METRIC | BEFORE | AFTER | CHANGE |
|---|---|---|---|
| Cost per lead | ₸34,000 | ₸18,000 | −47% |
| Test-drive bookings / month | 39 | 90 | ×2.3 |
| Qualified calls from ads / month | 74 | 119 | +61% |
| Cars promoted in live feeds | 0 | 300 | NEW SYSTEM |
| Share of leads with source-level attribution | 18% | 93% | +75 PP |
“Before Pifagor, we knew how many leads we bought, not which ads sold cars. In 4 months, we cut CPL by 47% and finally gave sales a dashboard they trust.”