Onlineshop-Geflüster - Der E-Commerce & Shop Podcast · Berend Heins

Die Zukunft des Trackings ist da! (🎙️Nils Jessen von Mable.ai)

·53 min·4 clips
Meta's ad algorithm behaves like a dog chasing a treat — it will target existing customers every time because they reliably convert, regardless of audience exclusions.
1. Onlineshop-Geflüster hosts Nils Jessen, co-founder and CEO of Mabel.ai, for a deep dive into e-commerce ad tracking and algorithm optimization. 2. Nils is a returning guest, having appeared approximately three years prior, and Mabel has evolved significantly since that first conversation. 3. The episode's central thesis is that most e-commerce brands unknowingly spend a large portion of their Meta and Google ad budgets targeting existing customers because the algorithms can't distinguish them from new ones. 4. Mabel's core differentiation from BI tools like Tracify and GetClar is that those tools inform the marketer, while Mabel directly improves the data fed to the advertising algorithm itself. 5. Nils explains that a Meta purchase event is like a package containing event parameters (what was bought) and attribution parameters (who bought it, via email, phone, IP, click ID). 6. Meta matches attribution parameters against its 3.5 billion user database with a 180-day lookback window, but mismatched emails, missing phone numbers, and cross-device behavior result in many returning customers being misidentified as new ones. 7. Mabel integrates directly into the Shopify backend and compares each new purchase against the complete order history in real time, creating two custom events: New Customer Purchase and Returning Customer Purchase. 8. Because Mabel runs on Shopify's own server — not a separate tracking server — it avoids the browser-level blocking that defeats pixel tracking and increasingly defeats server-side tracking as well. 9. Nils uses the dog-and-treat metaphor to explain the algorithm's behavior: it learns that existing customers are the 'easiest' conversions and gravitates toward them regardless of audience exclusions. 10. The marshmallow experiment metaphor is applied to the algorithm's response to audience exclusions: like a child left alone with a marshmallow, the algorithm eventually ignores restrictions to get the easy conversion. 11. Setting New Customer Purchase as the campaign conversion goal retrains the algorithm to treat existing customers as a negative signal rather than a positive one, achieving a structural shift in targeting. 12. Nils describes a gradual rollout approach: connect Mabel, observe the new/returning customer split baseline for two weeks, then switch campaigns from general purchase to new customer purchase as the conversion goal. 13. The HOLY brand, an early Mabel adopter, moved from 67% to 85% new customer acquisition rate over 12 months while tripling their Meta ad budget within unit economics targets. 14. A second unnamed case study went from 60% to approximately 80% new customer rate over three to four months. 15. A 25–30% drop in customer acquisition cost, as seen in these case studies, creates room to increase ad spend until CAC returns to the original level, enabling growth at the same economics. 16. Nils recommends Mabel from approximately 10,000 euros per month in ad spend, framing the tool fee (around 200 euros) as a reallocation from advertising budget rather than an additional cost. 17. For brands below the 10k threshold, Nils advises sticking with Shopify standard tracking and a clean cookie consent banner as sufficient. 18. The episode is structured as a technical deep-dive interview with Nils explaining both the conceptual framework and specific implementation steps for switching tracking setups. 19. E-commerce founders and performance marketers running Shopify stores spending at least 10,000 euros monthly on paid social or search will find the most direct value. 20. Brands not yet running paid ads, or those on non-Shopify platforms, will find limited immediately applicable guidance from this episode.
Listen to the show on