How Klover Unlocked Millions in 30 Days by Measuring True Marketing Incrementality

Fast Facts

• Industry: Fintech / Consumer Mobile App (Cash Advance Service)

• Challenge: Unreliable attribution obscuring true channel contribution, risking misallocation across paid media

• Solution: Always-on MMM anchored with incrementality tests to quantify real lift and guide mix shifts

• Timeline: First insights in the first week; major optimizations within 30 days

• Results:

  • Meta iOS spend reduced ~50% with no measurable loss in conversions

  • Apple Search Ads scaled by ~10x while remaining incremental

  • Incremental CAC improved by 35%

  • Material improvement in efficiency and a path to 7‑figure annualized savings

The Situation

Klover, a cash-advance mobile app with a multi-million-dollar monthly paid media budget, lacked confidence in platform-reported metrics after privacy changes. The team needed causal, channel-level clarity to decide where additional dollars were truly incremental.

The Challenge

Platform dashboards looked healthy across channels, yet internal signals suggested diminishing returns at higher spend levels. Two core questions framed the pilot with BlueAlpha:

• Was Meta iOS still driving incremental conversions at prior scale, or had it hit diminishing returns?

• Could Apple Search Ads be scaled well beyond prior levels while maintaining incrementality?

BlueAlpha’s Approach

Phase 1: Rapid Integration and Baseline MMM (Week 1)

BlueAlpha ingested two years of historical spend and conversions across a dozen-plus channels and produced calibrated response curves. Early reads indicated diminishing returns on Meta iOS at higher spend and headroom on Apple Search Ads.

Phase 2: Incrementality Test Calibration (Weeks 2-3)

Platform dashboards looked healthy across channels, yet internal signals suggested diminishing returns at higher spend levels. Two core questions framed the pilot with BlueAlpha:

• Was Meta iOS still driving incremental conversions at prior scale, or had it hit diminishing returns?

• Could Apple Search Ads be scaled well beyond prior levels while maintaining incrementality?

Phase 3: Channel-Specific Optimization (Weeks 3-4)

BlueAlpha ingested two years of historical spend and conversions across a dozen-plus channels and produced calibrated response curves. Early reads indicated diminishing returns on Meta iOS at higher spend and headroom on Apple Search Ads.

Phase 4: Continuous Monitoring and Refinement (Weeks 5-6)

As the plan rolled out, BlueAlpha provided weekly readouts comparing realized outcomes to model predictions. The data confirmed stable conversion volume through the Meta reductions and incremental growth from the Apple scale-up. Apple was split into Brand vs Non-Brand to enable finer-grained control as volumes rose.

The Results

Meta iOS Optimization:

• Spend: reduced ~50%

• Volume: no material decline in weekly conversions

• Efficiency: double-digit improvement in cost per incremental conversion (iCPA)

• Impact: on a run-rate basis, these changes point to 7‑figure annualized savings

Apple Search Ads Scaling:

• Scale: grew by ~10x while remaining incremental

• Incrementality: positive lift maintained throughout scaling

• Structure: split into Brand vs Non-Brand to preserve efficiency at higher volumes

Overall Impact (Pilot Window):

• Spend quality improved: more dollars flowing to incremental media

• Conversions: increased by 8% compared with the pre-optimization baseline

• Economics: incremental CAC improved by 35% as waste was removed

• ROI on BlueAlpha investment achieved in first 14 days

Key Takeaways

• Platform-reported metrics can overstate contribution at high spend levels. MMM aligned with experiment readouts is a reliable way to detect and correct diminishing returns.

• Cutting back where curves flatten and doubling down where curves are steep can unlock millions in annualized savings while preserving growth.

• Integrating experiment priors into MMM provides an auditable bridge between causality and scaled media planning.

What's Next for Klover

• Launching a Google Android UAC holdout test to quantify incremental lift and set efficient spend bounds

• Continuing brand vs non-brand optimization on Apple Search Ads

• Expanding always-on testing to additional channels and creatives while MMM tracks marginal efficiency