Performance Marketing Case Study
Multi-Market EU Dashboard: How CenturyPrint Turned 9 Countries Into a Growth System
How I combined Google Ads, Amazon, and Retif B2B data into one system that tells each European market whether to scale, localise, or expand — instead of ranking countries and stopping there.
Every European growth account I've worked on hits the same wall eventually. You launch in France, it works, so you copy the structure into Germany, Spain, Italy — and somewhere around country four, the dashboard fills up with numbers that all look reasonable and none of them tell you what to do next.
That's the problem I set out to solve with CenturyPrint's H1 2026 data. Not "build a dashboard" — build a system that looks at Google Ads, Amazon, Retif B2B, product performance, and geography together, and tells each market what it actually needs. Scale. Localise. Expand. Review. Not the same instruction dressed up in nine different flags.
Here's how I built it, and what it changed.
I Started With the Data, Not the Charts
The first mistake I've made — and seen other people make — is opening a BI tool before you've understood what the data can and can't answer. So before I touched a single chart, I mapped out what I actually had:
- Campaign, search-term, and location reports from Google Ads
- Order-level data across multiple sheets, tracking Dashboard, Campaign DB, and order history
- Amazon Sponsored Products data, split into paid vs. organic orders, spend, ROAS, and ACOS by country
- Retif B2B marketplace data, covering gross and net revenue, orders, AOV, and commission
Each of these sources answers a completely different business question. Google Ads tells you what's converting. Search terms tell you what customers actually typed. Location data tells you where it's working. Amazon and Retif tell you what's happening outside paid media entirely.
The question stopped being "how did Google Ads perform?" and became "which combination of market, channel, and product deserves the next euro?"
Reading the Account Past the Top-Line Number
On the surface, CenturyPrint's Google Ads account looked mediocre: roughly €21.2K in spend against €21.3K in attributed conversion value — a 1.01x ROAS. 1.68 million impressions, 30,211 clicks, but only 57 conversions.
If I'd stopped there, the takeaway would've been "Google Ads needs improvement." True, but useless — it doesn't tell anyone what to change.
Once I layered in campaign type, geography, and landing-page behaviour, a much sharper picture appeared. Performance Max was carrying the account: it made up 58% of spend but drove 87% of attributed revenue, running at roughly 1.50x ROAS versus Search's 0.38x. Display, meanwhile, produced no directly attributed revenue at all.
That's not "improve the account." That's "move budget away from the structurally weak formats and into the one already proving itself."
Building the Country Layer as a Set of Instructions, Not a Ranking
This is where most multi-market dashboards fall short — they rank countries by revenue and stop. I built the geography layer to do the opposite: pair each market's channel presence with its actual performance and turn that into an instruction.
Here's what that produced for CenturyPrint's seven active EU markets:
Notice none of those are the same instruction. That's deliberate. A dashboard that hands every country the same advice isn't a market strategy — it's a global average wearing a map.
The Nine Strategies Built From the Data
Scale Where the Signals Agree With Each Other
France was the clean case. It's active across all three channels, and French Performance Max is one of the strongest performers in the account — €6,535 in conversion value from €3,629 spend, about 1.80x ROAS, contributing 30.7% of Google Ads revenue from just 17.1% of the spend.
The instinct is to say "France is our biggest market, give it more budget." I built the logic to say something narrower: increase spend where France is already proving incremental efficiency, and watch for diminishing returns as you do it. Big markets don't automatically deserve more budget. Efficient markets with headroom do.
Let One Channel De-Risk Expansion Into Another
Spain and Italy gave me one of the most useful signals in the whole dataset. Neither had meaningful Google Ads presence. Both already had real Retif B2B revenue — roughly €4.2K in Spain, €3.7K in Italy — with zero paid media behind it.
I built the dashboard to treat that organic B2B revenue as demand validation, not as a coincidence. The question stops being "should we try Spain?" and becomes "organic buyers already exist in Spain, can paid B2C acquisition scale what's already there?" The recommended move: launch Spanish- and Italian-language Performance Max campaigns off the French structure, rebuilt with local-language assets, not translated copy.
Refuse to Copy-Paste a Winning Market
Germany is the cautionary case that made this principle non-negotiable for me. Google Ads was live there — €3,496 in spend, 11 conversions, a 0.61x ROAS — using largely the same French and English-language campaign structure.
There's a real difference between translating "Magnetic gift box" into German and actually localising: native copy, local search vocabulary, seasonal hooks tied to how Germans actually shop. I built the recommendation as a fork: fix the localisation foundation, or hold the spend until it's there. Scaling a mistranslated structure just makes the mistake more expensive.
Go Below the Country Level When It Lies to You
Belgium is where I learned country-level reporting can actively hide the answer. Wallonia's Google Ads CPA sat around €148 against an account average near €372. Flanders and Brussels, in the same "Belgium" line, were spending without converting at all.
I split Belgium into its own sub-regional view rather than reporting one blended number: budget behind Wallonia specifically, and out of the underperforming sub-regions instead of raising Belgium's spend as a whole. A blended country average would never have surfaced this.
Cross-Reference Product Data With Market Data
Countries alone don't tell you what to sell there. Magnetic Boxes turned out to be the through-line: the same product family led both Amazon and Retif, with Retif's top magnetic-box SKU driving B2B demand while premium Magnetic Box packs performed strongly on Amazon independently.
I built the dashboard to treat that as cross-channel validation, reframing the growth model from Market × Channel into Market × Channel × Product. A country can look mediocre overall while hiding one genuinely scalable SKU.
Separate Brand Protection From Real Acquisition
The search-term data forced an uncomfortable but important distinction. Of 56,276 unique search terms triggering the account, only 1,900 received spend — and of those, just 27 generated a conversion. 81.2% of tracked search spend, roughly €6,168, went to terms that converted zero times.
Worse: the terms that did convert were overwhelmingly branded — CenturyPrint, Century Box. I built that split explicitly into the dashboard, separating brand protection from new customer acquisition, because treating every conversion as proof the account can acquire new customers at scale is one of the most expensive misreadings in paid search.
Diagnose the Funnel, Not Just the Campaign
30,211 clicks produced 57 conversions — a 0.19% conversion rate. Display alone generated over 6,000 clicks with no directly attributed conversion. Meanwhile, one Test Landing Page campaign hit 1.18% CVR and 2.77x ROAS on its own.
That gap told me the problem wasn't purely a media problem. I pulled in the wider ecommerce funnel — 37,552 cart additions against 3,570 purchases — and built cart-recovery actions (dynamic remarketing, a structured recovery email sequence) directly into the dashboard's recommendations, not as a separate CRO conversation happening somewhere else.
Track Organic Momentum, Not Just Paid ROAS
Amazon revealed something a paid-only view would have completely missed: organic orders overtook ad-driven orders from March onward. Retif's revenue is overwhelmingly organic marketplace positioning, not paid acquisition, and it still generated €56.6K in gross B2B revenue against zero ad spend.
I built the dashboard to separate paid-attributed revenue from revenue the channel sustains without ongoing paid media — because if you're only watching ad ROAS, you'll never see the flywheel building underneath it.
Turn Seasonality Into a Budget Rule, Not an Observation
Monthly Amazon data showed the highest-spend month wasn't the most efficient one — April, with comparatively low spend, produced a noticeably stronger ROAS than January. Instead of a flat monthly budget, I built the recommendation as an active rule: pull spend back during the predictable post-holiday dip, and reallocate it toward the months where demand and efficiency both recover — as one unified seasonal calendar across Google Ads, Amazon, and Retif.
Why This Matters Beyond One Account
A multi-market EU dashboard isn't valuable because it fits France, Belgium, Germany, Spain, and Italy on one screen. It's valuable because it explains why those markets behave differently, using every signal available: Google Ads for acquisition intent, Amazon for paid-versus-organic momentum, Retif for demand validation, product data for what's actually scalable, and location data for where the unit economics genuinely work.
What I'd Tell Anyone Building a Multi-Market Dashboard
- Map your data sources before you open a BI tool. Each source answers a different question — know which one before you start charting.
- Give every market its own instruction, not a rank. A dashboard that treats every country the same isn't a strategy.
- Don't hide metric inconsistencies — govern them. A defined metric contract is what keeps stakeholders trusting the dashboard long-term.
Running paid media across multiple EU markets?
Get a market-by-market audit that tells you where to scale, localise, or pull back — not just a revenue ranking.
Talk to Us About Your Markets