E-commerce · pharmacy

How we stabilized a fast-scaling BI team at Pilulka

A mid-level BI team drowning in a 560+ story backlog while the business expanded into four new countries. We stepped in as interim BI managers; within two months new-request volume dropped ~80%, and after 18 months the team was self-led.

−80%
new-request volume within two months
−92%
compute cost
560+
story backlog cleared to a sprint cadence
18 months
to a self-led team
The work

When we arrived at Pilulka, we saw what we often see when a company hires one or two mid-level analysts and then starts growing fast. Their workdays had stretched from eight hours to twelve as a matter of routine, and the backlog held 560+ stories, some of them more than two years overdue. Management had just announced expansion into Romania, Hungary, Germany and Austria. The team was on the verge of burning out.

Why? They didn’t have a seasoned BI manager. Data didn’t have a seat at the table.

We were asked to fix the situation, so we stepped in as interim BI managers and started from the fundamentals.

First, we explained the situation to the Board. Until that point they had not been made aware of it, and once they were, they granted additional resources. In parallel we set a clear process for gathering requests and feedback from the business — and, more importantly, established the channels to communicate results back.

Internally, we dropped the endless Kanban ticketing and set clear two-week sprints: plan the work, work the plan, and leave proper space for deep focus.

Then we analyzed the requests themselves — manually, there was no ChatGPT then — and discovered that roughly 70% of them were easy ones, the kind that proper self-service would solve. So we put together biweekly Data Fridays, where business users came with their day-to-day challenges and the BI team shared screens on a projector and walked through the steps live. Within about two months, the rate of new requests dropped by 80%, and the internal BI team could focus on the work that actually mattered.

We set weekly prioritization sessions with the business teams’ managers, where we collectively discussed how to allocate sprint capacity. That way every team had a say, and if one made way for another, the other would reciprocate by sharing capacity on the next sprint. It created an atmosphere of candid, open and trusting working relationships.

We helped design the concept of the data mart: a set of curated, mature datasets made internally available across the teams. We led the adoption of Tableau for data exploration, taught the business how to find and evaluate the data products held in those data marts, and fully supported the training of Data Heroes drawn from each team — marketing, procurement, finance, sales and HR.

We set regular retrospectives, organized in a team-building style, and moderated the discussion so that genuine self-reflection had room and rough edges were smoothed over. The team gradually started working as a single delivery unit.

Finally, we set daily 15-minute standups. They gave us clear operational feedback and reinforced a delivery-oriented mindset — ownership of the tasks you take on, and the habit of asking “what was my contribution to the sprint yesterday?”. Asking for help within the team became natural, and as a team they grew enormously.

Our curation lasted 18 months. By the end of it, one of the internal team’s own members had stepped up to become BI Manager, and is leading the team to this day.

The biweekly retrospectives are still held 3 years on — and we are still regularly invited to take part.

Prague from Letná
You know who does the work.

Let’s explore hidden opportunities in your data.

Tell us where it hurts. We will get back to you within one business day.

Book a consultation