Ecosh Case Study – How MMM Increased Revenue by 56% | ePPC
Case Study Digitegu 2025 Winner Ecommerce · Supplements

How ePPC Solved Ecosh's Attribution Puzzle and Grew Revenue by 56%

When GA4 said Meta Ads drove 6% of revenue and Meta's own reporting claimed 54%, Ecosh needed a better answer. Marketing Mix Modeling gave them one — and the confidence to make budget decisions that led to record-breaking results.

+56%
Revenue growth (Sep+Q4 YoY)
12.8%
Media spend as % of revenue
+6%
ROAS improvement
The client

Ecosh Life

Ecosh Life is Estonia's first dietary supplement manufacturer and retailer, offering high-quality natural supplements made from the best exotic and local herbs. Their products are 100% GMO-free, made from the finest raw materials to support health and well-being.

The company sells primarily through its own ecommerce store and runs paid campaigns across Google Ads and Meta Ads to drive revenue.

Gertrud Rei
Gertrud Rei
Google Ads
Chinmay Kulkarni
Chinmay Kulkarni
Data Scientist · MMM
Andrus Kiisküla
Meta Ads ·
Ecosh products
The problem

Three sources, three different stories

In the first two quarters of 2024, Ecosh faced a measurement crisis that many ecommerce brands know well. Google Ads claimed it was responsible for 51% of all revenue. Meta Ads claimed 54%. And GA4 — their supposed source of truth — credited Meta with just 5.89%.

The numbers couldn't all be right. And the stakes were real: every budget decision was being made on data that contradicted itself.

51% 49%
Google Ads self-reports 51%
of total revenue attributed to Google
54% 46%
Meta Ads self-reports 54%
of total revenue attributed to Meta
The GA4 dilemma

Meanwhile, GA4 told a very different story

GA4's last-click attribution model heavily favors lower-funnel touchpoints — organic search, email, and direct visits. According to GA4, Meta Ads contributed just 5.89% of revenue, while email marketing via Klaviyo accounted for 13.7%.

This created a practical problem: if the team trusted GA4 at face value, they would have cut Meta spend significantly. But Meta Ads was potentially driving the brand awareness that fueled all other channels.

The question the team needed to answer: What is the true causal impact of each channel, independent of how platforms choose to count credit?

GA4 Revenue Split
Google Ads 45.7% Klaviyo 13.7% Organic 11.6% Other 11.3% Meta 5.89%
The approach

Identifying the real drivers with causal discovery

Rather than relying on any single platform's attribution, we built a Marketing Mix Model that uses Bayesian statistics and causal discovery algorithms to identify what actually drives revenue.

  • Used expert knowledge and a Causal Discovery Algorithm to map the relationships between channels, brand interest, and revenue
  • Applied ad stock transformations to capture the lingering impact of ads days and weeks after exposure
  • Modelled saturation effects so we could see exactly where diminishing returns began

Key finding: After adjusting for all factors, Meta Ads had the strongest positive impact on revenue — something neither GA4 nor platform-reported attribution could reveal.

Causal Graph · Key Revenue Drivers
Vitamins Interest Google Ads Clicks Meta Ads Impressions Brand Rolling Meta Ads Cost Google Ads Cost Revenue Daily Temp.
What the model revealed

Meta's marginal impact was 2.5× higher than Google's

The MMM showed that for every additional euro spent, Meta Ads generated significantly more incremental revenue than Google Ads — especially in the mid-spend range where Ecosh was operating.

This was invisible in GA4's reporting because Meta's impact worked indirectly: it drove brand awareness, which then converted through organic search, direct visits, and email — channels GA4 gave all the credit to.

  • Traditional attribution models failed to anticipate Meta's performance at scale
  • The model also showed diminishing returns, helping the team avoid overspending
Marginal Impact · € Revenue per €1 Spent
7.5 5.0 2.5 0 Daily Spend → Meta Ads Google Ads
Planning with confidence

Simulating outcomes before committing budget

With the model in place, the team built a forecasting system that could predict next month's revenue based on planned spend. This wasn't a black-box forecast — it was transparent, showing worst-case, most probable, and best-case scenarios.

  • Thousands of Monte Carlo simulations incorporating real-world uncertainty
  • Each month, the team reviewed risk/return trade-offs before allocating budget
  • The system used planned spend, dates, and historical patterns to estimate impressions, clicks, and revenue
Revenue Simulation · Daily Distribution
Daily Revenue Worst Case Most Probable Best Case
The results · September + Q4 2024

Record-breaking quarter driven by better decisions

With the MMM guiding budget allocation from September through Q4 2024, Ecosh achieved the strongest period in the company's history — surpassing their annual growth target in just four months.

+56%
Revenue growth
YoY (Sep + Q4)
+6%
ROAS improvement
vs prior year
12.8%
Media spend as % of revenue
Down from 13.6%
Net Revenue Index (Sep+Q4)
100.0
2023
156.2
2024
Media Spend % of Revenue
13.6%
2023
12.8%
2024

The period included two of the highest revenue months in the company's history, and the year closed with the highest annual revenue ever recorded — all while media spend as a percentage of revenue actually decreased.

What we learned

Key takeaways

GA4 systematically undervalues Meta Ads because its last-click model gives credit to lower-funnel touchpoints. The real contribution of upper-funnel channels is invisible in standard analytics.

Meta has a broader ecosystem impact than direct-response metrics suggest. It boosts brand interest, which then drives conversions through organic search, direct traffic, and email.

Standout creative is the catalyst. The brand interest increases were often driven by distinctive ad creative that sparked curiosity and tied all brand elements together.

Your turn

Facing a similar attribution challenge?

If your GA4 and ad platform data don't match — you're not alone. Book a free call and we'll assess whether Marketing Mix Modeling is the right fit for your business.

Karl – Founder
Karl Founder of ePPC — your first point of contact
Chinmay – MMM Expert
Chinmay Building custom marketing mix models to turn data into budget decisions

Or write to us: karl@ppc.ee