Context
An e-commerce platform from Zagreb — specialized in selling sports and recreation equipment. 8 years in the market, 45 employees, a catalog with 8,000 SKUs across 12 categories. Growth had plateaued — revenue was growing 4–6% per year, but conversion was below industry (0.9% vs. an industry average of 1.8%).
The owner had tried different approaches: more aggressive discounts (margins down, conversion didn't really improve), more ads (acquisition cost up without conversion gains), email marketing (a tired segment, weak responses).
The real problem was in the buying experience — the site was designed 5 years ago, the mobile UX was outdated, and there was no recommendation system. A visitor who came looking for one thing rarely discovered other products they might have wanted.
What didn't work first
Three attempts before this one. Price cuts — margin down, conversion flat. More ads — acquisition cost up, conversion flat. Email campaigns — tired segment, weak responses. The problem wasn't traffic or price; it was what happened on the site.
Approach
We combined Web & Digital Products with AI & Automation — a front-end redesign alongside the introduction of AI product recommendations.
Experience redesign — mobile-first approach, simplified navigation (12 categories collapsed to 6 main groups), checkout shortened from 7 steps to 3, transparent shipping costs from the start (instead of revealing them at the final step).
AI recommendations — a model trained on historical purchase and on-site behavior data. A visitor browsing hiking gear gets recommendations for complementary products — not at random, but following the patterns of real purchases. Personalized by category, season, and behavior during the visit.
Bonus — customer support: the on-site chat got an AI layer that answers 70% of routine questions (availability, shipping, returns) without a human stepping in. The support team was freed up for the more demanding conversations — equipment advice, custom orders.
Delivered
- Redesigned e-commerce platform with a mobile-first approach and simplified navigation.
- AI product recommendation system trained on historical behavior data.
- AI chat for customer support covering 70% of routine queries.
- Analytics dashboard for tracking conversion by segment.
- Team training for the new tools.
Result
- Conversion grew from 0.9% to 2.3% — a 156% increase. From the same number of visitors, almost three times as many people now buy.
- Cost of acquisition per sale dropped proportionally with the conversion increase.
- Average order value up 89% — the AI recommendations were working. A visitor who came looking for a cheaper product would often leave with a larger cart after seeing complementary items.
- Cart abandonment dropped 32% — the short checkout and transparent costs removed the main reasons for dropping off.
At a glance
+156%
Conversion (0.9% → 2.3%)
+89%
Average order value
−32%
Cart abandonment rate