How Mavlers cut CPA by 42% on 30% less spend for a Phillipines' bank

We stopped optimizing for the "Click" and started optimizing for the "Approval." By integrating internal bank data into our bidding strategies, we forced the algorithms to seek out high-propensity users rather than high-volume lead generators.

We paused expensive competitor campaigns (where CPAs reached $20) and redirected the budget into granular Brand-tier campaigns. By segmenting by card type (Blue, Gold, Platinum), we gained surgical control over budget allocation. 

Search architecture was rebuilt to reduce keyword overlap, improve relevance, and provide greater control across credit card categories. 

The team also refined ad messaging, extensions, and audience bid adjustments to better pre-qualify prospective applicants before they entered the application process.

Internal approval trends were analysed to identify demographic, geographic, and behavioural characteristics associated with stronger approval performance. 

Rather than relying on platform interest tags, we built Meta Lookalike audiences based on "Approved Applicants" only. We specifically isolated the "Carded Non-Depositor" segment, which historically showed the highest approval propensity.

We overhauled the creative assets to act as a filter. By clearly stating value propositions and eligibility cues in the copy, we discouraged unqualified clicks, improving the quality of traffic before the user even landed on the site.

Performance Max and Advantage+ campaigns were rebuilt using structured audience signals, refined asset groups, and historical performance benchmarks.

We applied a "test and cut" discipline. TikTok was introduced, evaluated against the approval benchmark, and promptly exited when lead quality lagged. Conversely, Bing Ads was scaled when it delivered a 10.5% approval rate, matching Google’s efficiency at a lower cost.