App-first Quick Commerce Growth
Leading performance inside an app-first grocery business, then helping move campaign, conversion, LTV, and fraud decisions onto one attribution model.
Team outcomes reported in Adjust’s published istegelsin case study.
istegelsin was an app-first grocery and quick-commerce business. Acquisition, the first order, repeat behaviour, and retention belonged to one customer journey, even when the tools reported them separately.
As Head of Performance Marketing, I led the performance function across mobile acquisition, e-commerce growth, lifecycle measurement, and day-to-day budget decisions. The work was commercial and operational before it was technical.
The team brought together media buying, data engineering, and creative. Setting priorities, coaching the work, and keeping those disciplines connected were part of the day-to-day role.
One of the most useful changes was a shared attribution model. I worked on it with istegelsin, Unboxed, and Adjust. The implementation and results belong to those teams.
The measurement problem
Facebook, Google, and other platforms could each claim the same conversion. Different SDKs and dashboards made it harder to compare campaigns, first-time buyers, LTV, ROI, CPA, and conversion rates in one place.
The teams found duplicate attribution approaching 50% in conversion counts from some key performance ads. At that level, the measurement model affected day-to-day campaign decisions.
The system
The company moved campaign measurement into Adjust and used one attribution model across channels. Raw Data Export gave the teams access to more than 200 million monthly events and behavioral data points for deeper analysis.
Fraud prevention became part of the same workflow. Questionable sources could be reviewed quickly, while campaign and lifecycle performance stayed visible in one place.
The published result
Adjust reports that the teams improved ROAS by more than 35% after adopting its Fraud Prevention Suite. It also documents the duplicate-attribution finding and the scale of the raw event analysis.
These are published team outcomes. I do not present them as work completed alone.
Read Adjust’s istegelsin case study.
The next quick-commerce chapter
I carried that operating context into Getir as Digital Marketing Manager for Western Europe. The market changed, but the model remained app-first: acquisition, CRM, retention, and market growth had to be read together.
I do not attach unverified performance figures to that chapter here. Its value in this story is the operating range: moving from team leadership in one quick-commerce business into a multi-market role in another.
What I carried forward
Attribution changes the quality of a decision only when the team can use it. The shared operating model gave us a way to decide which source to trust, where to move budget, and when to stop questionable traffic.
That lesson now shapes how I build systems: understand the commercial loop, define the event, preserve its context, expose the source boundary, and keep the result available for the next cycle.