Book a call
← All case studies
Case study · E-commerce analytics · Automotive

Where an international online shop
leaves revenue on the table.

For the global lifestyle and accessories shop of a premium car manufacturer, we analysed ten months of e-commerce data from more than 30 countries, with Google Analytics 360 and Adobe Analytics in parallel. The result: a seven-figure revenue potential, clearly located by market and product group, and a list of data errors that nobody had seen before.

30+country markets compared, with confidence intervals instead of gut feeling
2analytics systems reconciled: GA4 360 and Adobe Analytics
7 figuresrevenue potential per year, quantified with a 90 % interval
36 %of one month's revenue came from a single product drop
Starting point

Many markets, two tools, no clear answer.

The shop sells clothing, watches, lifestyle products and vehicle accessories in more than 30 countries. The figures were spread across two analytics systems and several properties, with different country assignments. The client's question was simple: where are we losing revenue, and what should we tackle first?

The standard reports were not enough for that. We combined the data, tested it statistically and quantified the potential in euros, with a range showing how certain each figure is.

Google Analytics 360Global roll-up property and the US team's property, BigQuery-ready, without sampling.
Adobe AnalyticsEvery checkout step recorded individually, countries by shop market instead of IP.
Data layerProduct categories from the data layer, grouped into 14 product groups.
The key findings

Five answers, each with a figure.

39 %
01 · USA

The US checkout is the biggest lever.

A quarter of all visits come from the USA. US customers add to the basket more often than anyone else, but only 39 % of started checkouts end in an order, compared with 46 to 52 % in Europe. Many click “Order” and still do not buy: an indication of payment and validation errors.

½
02 · Product range

Vehicle accessories hardly sell in the USA.

US customers buy an above-average share of watches, sunglasses and clothing, but only about half as many interior, luggage and charging accessories as comparable markets. The biggest product gap of all markets.

1.9×
03 · Switzerland

Switzerland is the most valuable market per visit.

The highest conversion rate and large baskets: every visit brings almost twice the average revenue. A benchmark for all other markets.

93 %
04 · Nordics

In Scandinavia, the shop is an accessories shop.

In Sweden, Denmark and Norway, 82 to 93 % of revenue comes from vehicle accessories. Lifestyle products hardly sell, which points to the local product range rather than a lack of demand.

36 %
05 · Product drops

A single product launch can turn a whole month.

A limited design collaboration brought in 36 % of the month's revenue in a single day. Without this drop, the month was average. Our recommendation: plan drops deliberately several times a year and always show monthly reports with and without the drop, so that normal business is not misread as growth or decline.

Funnel by country

Where buyers drop out.

We followed every session through four steps: viewed product, added to basket, started checkout, ordered. Choose a country and compare it with Germany. Hatched: clearly weaker than Germany.

USA compared with Germany

Viewed product
Added to basket
Started checkout
Ordered

The USA loses out in two places: only 46.5 % of sessions view a product at all (a lot of bot and non-shop traffic), and in the checkout. Those who do view a product, however, add to the basket more often than in any other market.

United Kingdom compared with Germany

Viewed product
Added to basket
Started checkout
Ordered

The United Kingdom follows the German pattern. The difference lies in the basket, not the funnel: orders are around a fifth smaller.

France compared with Germany

Viewed product
Added to basket
Started checkout
Ordered

France converts almost like Germany. The levers are cross-selling and bundles for larger baskets.

Switzerland compared with Germany

Viewed product
Added to basket
Started checkout
Ordered

Switzerland is at least at the German level in every step and also has large baskets: the most valuable market per visit.

Method

Statistics you can explain.

Small markets with few orders can easily look better or worse than they are. That is why we use models that account for chance and state every potential as a range.

Hierarchical beta-binomial modelEstimates each market's true conversion rate, pulling small markets towards the average. Target: the level of the best quarter, not of the single top market.p̂ᵢ = (yᵢ + α) / (nᵢ + α + β)
Monte Carlo simulation4,000 draws with bootstrapping give a 90 % interval for every revenue potential.ΔRevᵢ = nᵢ · max(0, pₜ − pᵢ) · AOVᵢ
Decomposing revenue per visitIs the gap down to conversion or to basket size? The log decomposition separates the two cleanly.ln(RPS) = ln(CR) + ln(AOV)
Market segmentationCentred log-ratio, Ward clustering and PCA divide the markets into “lifestyle” and “vehicle”. Each market is compared with its peers.clr(x)ⱼ = ln(xⱼ / g(x))
Significance testsχ² test and Cramér's V for differences between countries, Spearman correlation for basket and conversion.ρ = −0.49 · p = 0.03
ConcentrationGini coefficient and Herfindahl index show how much revenue depends on a few markets and categories.HHI = Σ sᵢ²
Recommendations

What the client is tackling now.

MeasureMarketWhy
Fix the US checkout, a “fits my vehicle” model filter, accessory bundlesUSA, CanadaThe biggest single lever. Customers buy rarely, but with large baskets, and vehicle accessories are under-represented.
Cross-selling and bundles for larger basketsUK, France, ItalyThey buy as often as Germany, but spend around a fifth less per order.
Review the lifestyle range and its visibilitySweden, Denmark, NorwayLifestyle revenue close to zero.
Plan product drops deliberately, reports with and without the dropAll marketsA drop can distort a month and is at the same time a real revenue driver.
Fix the data: reconcile properties, add checkout events, alert on missing product dataAnalyticsDecisions depend on these figures.
Found along the way

Data errors nobody else sees.

If you put two analytics systems side by side, you find contradictions. In Germany, GA4 counted only about half the visits that Adobe saw, but almost all the orders. This made the German conversion rate look too high, and the gap to the USA bigger than it is.

After a change in the A/B testing setup, almost all product data was missing from the purchase events for nine days. And GA4's standard checkout events were not being sent at all. We documented all of this and fixed it with the team.

For reasons of confidentiality, we do not name the client and do not show absolute revenue. All percentages come from the project.

Similar project?

Where is your revenue potential?

Tell us in 30 minutes where you stand, or send a short message. We will tell you honestly whether and how we can help.

Armin Sanjari
Armin SanjariFounder & CEO · Your contact

Prefer to call? +49 711 500 908-17

Book a 30-minute call →
or send a short message

Armin replies personallyWithin 24 hours on working days · no obligation · no sales pressure

Request a free tracking check →

Book a 30-minute call

Open at Calendly ↗ · Send a message instead · Privacy settings