Trigger rate from 14% to 24% — solving a lead quality problem that dashboards couldn't see.

The Challenge

PaisaBazaar's campaigns were generating leads at Generation CPL consistently on target. On paper, the account was performing. But the downstream trigger rate — the percentage of leads that engaged further and showed genuine intent — sat at just 14%. For every 100 leads generated, 86 were form-fillers with no real purchase intent. Triggered CPL was ₹3,200 — more than 4× the Generation CPL of ₹760. The problem was invisible in the standard dashboard but costing significantly in downstream sales efficiency.

The Solution

Results

The combination of structural changes, AI-driven optimisation via Adtrafix OS, and weekly cadence produced measurable improvement starting in week 3, compounding through month 3:

Why Lead Quality Is Invisible in Standard Dashboards

PaisaBazaar's challenge is one of the most common — and most overlooked — problems in performance marketing. Standard Google Ads and Meta Ads dashboards show Generation CPL: the cost per form submission. This metric is visible, trackable, and optimisable. It is also incomplete.

Triggered CPL — the cost per lead that actually converts downstream — requires a second conversion event to be configured and tracked. Most agencies never set this up. The result is that accounts appear to be performing on paper while generating significant downstream inefficiency. A lead that costs ₹760 to acquire but never converts is not a ₹760 CPL — it is an infinite CPL from a business perspective.

The Keyword Intent Mapping Process

The first step was mapping every active keyword to a purchase intent stage. We categorised PaisaBazaar's keywords across four stages: Awareness (informational, high-volume, low-intent), Consideration (category comparison, mid-funnel), Intent (product-specific, high purchase signal), and Transactional (direct application queries).

Awareness keywords — "what is a personal loan", "credit score meaning" — were moved to a separate campaign with a much lower tCPA target and minimal budget. These terms serve brand awareness purposes but should never be optimised for leads. Intent and Transactional keywords received the majority of the budget and were structured into tightly themed ad groups with matching landing pages.

The Meta Custom Audience Rebuild

Meta campaigns had been running with manually defined demographic and interest-based audiences. We replaced these with three Advantage+ campaigns seeded with warm Custom Audiences: organic website visitors (30-day window), video viewers (50% completion on financial content videos), and lookalike audiences built from the top 10% of triggered leads (the highest-quality lead segment by downstream conversion data).

The algorithm's performance on these warm-seeded campaigns was noticeably stronger within the first two weeks — engagement rates higher, CPL lower, and critically, trigger rate significantly better than the cold-audience campaigns. By week 8, warm Custom Audience campaigns accounted for 60% of Meta budget (up from 20%) and produced 73% of triggered leads.

Measuring What Matters

The reporting shift was as important as the campaign changes. Adtrafix OS tracked both Generation CPL and Triggered CPL as primary KPIs on the same dashboard — making the quality gap visible in real time rather than in a monthly retrospective. Budget decisions were made weekly based on which campaigns, keywords, and audiences were producing the highest trigger rates, not just the lowest Generation CPL. This single change in measurement philosophy drove the sustained improvement seen across both channels.

Key Takeaway

This case demonstrates what changes when AI is embedded into the operating model rather than used as an occasional tool. The results compound weekly — each optimisation feeds the next — rather than spiking once and stabilising. Adtrafix OS running behind every decision is what separates a one-month win from a structural improvement.