As the Head of Rocket Internet's Business Analytics team, my task was to rapidly roll out data analytics infrastructure across Rocket Internet's portfolio companies.

Jabong was a different beast, though.

The energy in the Delhi / Gurgaon office was insane. You could feel it in every meeting, every conversation, every late-night planning session.

Those were some of the smartest and hardest working people I ever met.

The intensity was intoxicating - we were building something extraordinary, and we knew it.

The rollout of our data infrastructure was intense.

Here's our insane rollout plan:

→ Day 1: Server hardware provisioning (yes, we had our own servers, not Cloud servers) 

→ Day 2: Oracle database installation on primary and backup servers 

→ Day 3: Initial Core data infrastructure deployment, connecting sources 

→ Day 4-5: Business Objects reporting frontend setup 

→ Day 6: Web tracking data export implementation 

→ Day 7: Rollout of all data transformation logic, ETL and data model

Our initial assumption was that 80% of our core data warehouse would work out of the box, with 20% needing customization.

Reality flipped this completely - we ended up customizing 80% of the data infrastructure to meet India's unique demands.

What we discovered was a level of complexity that transformed our entire approach to data infrastructure.

Here's a few things that made India unique:

• Legal structure: Restrictions on foreign-owned multi-brand retailers created a company setup that was quite unique. Due to Rocket Internet’s majority stake, Jabong could not sell directly to customers and we needed to set up a complex legal construct where the customer would place the order with Jabong, Jabong then sold to a reseller who then sold the product to the customer.
• Interstate regulations in India: I will never forget how incredibly complex the rules of shipping goods between different states in terms of taxation and shipping costs were.
• Last mile delivery: Delivering goods in India was generally very different from other Rocket Internet ventures which is why Rocket built their own last-mile delivery infrastructure. This created additional complexities and massive customization needs for our data infrastructure team. It also added analytics needs that were non-existent in other markets (e.g. fleet management and route optimization, which is a nightmare in India).

The rise and fall of Jabong

Jabong's growth was astronomical:

• Only a few months after launch in early 2012, Jabong was among the most visited websites in India
• Jabong was in the Top 10 of the most searched terms in India within a year after launch
• Also within the first year, they were shipping 14,000 orders daily
• In the second year Jabong grew revenues by 50x

Sadly, after reaching a valuation of over $500 million and becoming one of India's most recognized brands, the company was eventually acquired by Myntra for a mere $70 million.

It was one of the rare markets where Rocket Internet's proven playbook of market domination didn't succeed.

Despite the eventual downfall of Jabong, every day brought new challenges and learning opportunities that no textbook could teach.

During our data project in India, we learned about things we never expected to encounter as a Berlin-based data team:

• How different Indian states handle taxation
• The metrics that matter when setting up a last mile delivery fleet
• How to calculate shipping costs and operational costs in an ultra-complex market
• Forecast demand for shoes and clothing in the world’s most populous country
• Driving customer lifetime market in such a vast country with such a diverse population

I watched our team grow from data specialists into business problem solvers who really understood e-commerce operations in India.

We weren't just building dashboards; we were helping shape how a modern e-commerce company operates in one of the world's toughest and most interesting markets.

The experience was truly transformative and shaped my career like no other project before.

It was sad to see that Jabong didn’t find the success we hoped for.

But what a ride it was!

Bottom Line

Building data teams in high-growth startups is the ultimate roller coaster ride.

You'll learn more in one year than you might in five years at an established company. The risk of failure is high, but the learning experience is unparalleled.

For future data leaders considering a role in a fast-growing scale-up, remember:

→ Be ready to learn about the business and not just the data 

→ Prepare to adapt EVERY day 

→ Value the knowledge gained, regardless of the outcome 

→ Build networks that will last beyond any single venture

In high-growth startups, data teams don't just build infrastructure - they become business architects who shape the direction of a (hopefully) future market leader.