Work/Decision systems
Building the reporting layer executives actually use
Different teams were calculating the same performance measures in different ways, so meetings often began by arguing about whose number was right. This work replaced ad hoc reporting with a standing view that every function drew from, first at McDonald's and later at Riot Games.
- Organisation
- Riot Games, McDonald's
- Region
- APAC and Pakistan
- Period
- 2018 to 2026
- Focus
- Data pipelines, executive reporting
Context
Performance data sat across several systems and was pulled by hand whenever somebody asked for it. Teams defined the same measures differently, which meant the numbers rarely agreed.
Reporting was also reactive. Analysts spent much of their time answering one-off requests, which left little room for the questions that mattered more.
Mandate
Replace ad hoc reporting with a standing view of performance that every function could rely on and that did not need an analyst to produce.
Approach
At McDonald's I built Power BI dashboards and OKR reporting covering more than 70 outlets, so that restaurant operators could see daily performance rather than waiting for a monthly pack.
At Riot Games I designed the reporting architecture for games across Asia Pacific on Tableau Cloud. Source data was consolidated and transformed in SQL into agreed tables, so that every dashboard drew from the same definitions.
In both cases the metric definitions were agreed with the teams that owned them before anything was built. That step took longer than the build itself and was the reason the reporting held up afterwards.
Outcome
Reporting became a routine that ran on a schedule rather than a queue of requests. Operators and leadership worked from the same figures.
Analyst time moved away from assembling numbers and towards explaining them.