Felipe Bonzanini

In this project, my team and I created a cube for the performance marketing team to control their spending and to develop budgets for the more performant campaigns. I extracted data from many sources like Google Analytics, AppsFlyer, AdWords, Criteo, Bing, and Facebook and combined them with sales data to be able to calculate metrics like Conversion Rate, Churn, Cost per Sale, LTV, CPA, CAC in many dimensions.

Together with the marketing team, we defined transformation rules for the UTMs in the web, and deep links in mobile, to aggregate it by the group channels (i.e., SEO, retargeting, social media, and so on).

Most of my work consisted of extensive and interactive collaboration with the marketing team to validate this data, which ended up with several changes to the UTM classification for proper control enablement. I had the opportunity to learn a lot of digital marketing and deeply understand how is the data analysis related to performance marketing.

Our new product replaced all of the worksheets and manual work of obtaining data from the source; it also sent all this data to an automatized dashboard and a cube to slice and dice the data.

Unfortunately, I cannot share a link as this is a private project.

At PlayKids, one of the companies of the Movile group, I had the opportunity to restructure all indicators and KPIs so that high-level management could have a more accurate view of the business.

I suggested a new methodology that is specific to the subscription business model that creates market KPIs without the noise that you'd have if you used a traditional data model like the ones for eCommerce. We also recalculated the active user base, churn, conversion, and trial success rate, along with other indicators.

I was the data engineer re-creating a planning builder tool for retail businesses. The plan involved a lot of understanding about annual and monthly retail sales, markups, turn, and inventory. The project took three months, and we were able to enhance its speed from 30 hours to 2.5 minutes and add a lot of new features.