Weekly Pix transfers per active NuConta user fell 8% in Brazil. How would you find the cause?

Introduction

Pix is the heartbeat of NuConta. Brazilians use it to pay rent, split bills, buy from street vendors, receive salaries and pay online. Nubank is one of the largest Pix participants in the country, and features such as Pix no Crédito, tap-to-pay Pix and an AI-powered Pix assistant (reported at over 10 million monthly users) have made it central to the app. Pix transfers per active user is therefore a strong signal of engagement and of whether Nubank is the customer's primary account.

In this scenario, weekly Pix transfers per active NuConta user fell 8% in Brazil. I'll investigate systematically, starting with whether it is real, then where it comes from, and why.

Step 1

Clarifying Questions (3 minutes)

• Over what period? I'll assume this week versus the average of the previous four weeks.

• Does "transfers" include sent only, or sent and received? I'll assume sent Pix payments (outgoing).

• How is "active user" defined, and did that definition or the active base change? If many new, less engaged users joined, the ratio can fall without anyone using Pix less.

• Did the total number of Pix transfers fall, or only the per-user ratio? This tells us whether we have a numerator or denominator problem.