Our team's campaigns · Social apps

Social chat app · US/UK/AU

Our team combined Meta purchase-event optimisation, Google Ads and Mintegral for a subscription social chat app. Blended CPA purchase reached $73.39, inside the $65–$80 target.

Business model
Subscription
Markets
United States, United Kingdom, Australia
Period
September – December 2024
Channels
Meta Ads; Google Ads; Mintegral
Services
User acquisition / Creative studio / Measurement & attribution / Traffic quality

The challenge

The growth constraint.

The subscription chat app needed to scale a budget of $80,000+ and bring CPA purchase into a $65–$80 target range. Tier-1 competition, strict creative moderation and high Google keyword CPC shaped the tests.

The approach

Decisions behind the delivery.

  1. Optimise towards purchase

    Our team implemented deep post-install event tracking for Meta App Event Optimization on purchase. Google bidding moved from tCPI to tCPA.

  2. Qualify the audience

    Interactive quiz pre-landers filtered non-target traffic. Weekly Mintegral publisher pruning used high uninstall rates as a quality signal.

  3. Keep testing the conversation promise

    40+ video A/B tests on Meta explored live-conversation features while moderation constraints shaped the text and visuals.

Attribution

Our team’s campaign work.

Channels

Where the plan ran.

Channel scope and rationale
Meta Ads was the main source for scaling paying users through purchase-event optimisation. Google Ads combined competitive-keyword search with App campaigns. Mintegral added install volume at low CPI.
Delivery controls
The team pruned Mintegral publishers weekly where uninstall rates were high. Purchase-event signals and tCPA bidding connected source decisions to the paid outcome.
  • Meta Ads
  • Google Ads
  • Mintegral

Creative

The product promise, tested.

Formats and tests

40+ video A/B tests on Meta focused on live-conversation features. Interactive quiz pre-landers helped filter non-target traffic. Copy and visuals were adapted to strict moderation.

Readout

Meta recorded the largest purchase volume and lowest CPA purchase of the three sources. This is a source-level observation, not an isolated uplift from creative or pre-landers.

Measurement

The basis behind the results.

Source
Team campaign data.
Campaign period
September – December 2024.
Platform
Android.
Event basis
Trials and purchases are reported separately by source. The paid-event comparison uses reported purchases; uninstall rates informed publisher decisions but no aggregate uninstall reduction is supplied.
Cost basis
USD. Spend $88,800; 1,210 purchases; blended CPA purchase $73.39. Compare with the campaign’s stated $65–$80 purchase-CPA target, not with other apps’ targets.
Reporting scope
Campaign totals and source-reported metrics. Attribution windows and cohort maturity are not supplied; no causal lift or measured payback is inferred.

Results

What the campaign reported.

  • Blended CPA purchase

    $73.39

    September – December 2024

    $88,800 total spend ÷ 1,210 reported purchases; source-reported blend across three sources.

    Source: Team campaign data

  • Purchases

    1,210

    September – December 2024

    Reported purchases across Meta Ads, Google Ads and Mintegral.

    Source: Team campaign data

  • Installs

    46,490

    September – December 2024

    Reported installs across all three sources.

    Source: Team campaign data

CPA purchase by source

USD per purchase

Campaign target: $65.00–$80.00
Meta Ads $70.80Meta Ads · CPA purchase: $70.80 · inside the target band
Google Ads $76.82Google Ads · CPA purchase: $76.82 · inside the target band
Mintegral $76.66Mintegral · CPA purchase: $76.66 · inside the target band

All three reported source CPAs were inside the $65–$80 target; blended CPA purchase was $73.39 across 1,210 purchases.

Team campaign data · September – December 2024

View data table
CPA purchase by source
ItemCPA purchase (USD per purchase)
Meta Ads$70.80
Google Ads$76.82
Mintegral$76.66
Campaign target$65.00–$80.00

Interpretation

The blended CPA purchase of $73.39 fell within the stated target. Meta supplied 685 purchases at $70.80 CPA purchase; the campaign does not report renewal, ROAS or payback outcomes.

Source table

Spend and event counts share each row’s source scope. CTR = clicks ÷ impressions; CPI = spend ÷ installs; CPA purchase = spend ÷ reported purchases. “—” means unreported, never zero.

Source Spend Impressions Clicks CTR CPC Installs CPI Trials Purchases CPA purchase
Meta Ads $48,500 4,210,000 151,560 3.60% $0.32 24,870 $1.95 2,310 685 $70.80
Google Ads $24,200 1,610,000 46,690 2.90% $0.52 8,420 $2.87 910 315 $76.82
Mintegral $16,100 2,480,000 69,440 2.80% $0.23 13,200 $1.22 840 210 $76.66

Period: September – December 2024 · Source: Team campaign data

Additional reported figures

Campaign totals
  • $88,800 spend
  • 46,490 installs
  • 1,210 purchases
  • blended CPA purchase $73.39

Period: September – December 2024 · Source: Team campaign data

Metric basis explorer

Blended CPA purchase: $73.39. Period: September – December 2024. Basis: $88,800 total spend ÷ 1,210 reported purchases; source-reported blend across three sources. Source: Team campaign data.

What comes next

The next question.

Which combination of purchase-event signals, audience qualification and fresh conversation creative can sustain the target as spend grows? The next test should retain the source and purchase definitions used here.

Detailed case studies, including programmes above $1M per month, are shared under NDA.

Working through a related growth question?

Tell us about your app, markets and the paid event you need to understand.