B2B from Zero
5% of customers were driving 30% of revenue — and they were businesses. What followed was building a division, its team, and the tools to run it from scratch.
This business line didn't exist when I joined. It grew from a data observation into a structured division with its own pipeline, team, and infrastructure.
The Insight
Desertcart was B2C when I joined. Breaking apart order frequency data revealed a pattern: the top 5% of customers by volume were generating 30% of revenue, with basket behaviour that looked nothing like a consumer — multi-item weekly orders, company email addresses, repeat purchasing across categories. These were businesses being served exactly like everyone else.
"These customers already existed. Recognising them as a distinct segment and building infrastructure to serve them properly was the whole job."
Building the Division
Tag the segment
Scoring model built from four signals. About 5% of the customer base flagged as business accounts — the foundation for everything that followed.
Hire one AM, assign 50 accounts
Target: grow baseline revenue 30%. No promotions, no discounts — just dedicated attention. Orders managed correctly, someone to call with questions. It worked immediately.
External acquisition — tried and cut
Sales outreach agency (2 months) and Google Ads both failed. Wrong customer quality; acquisition cost far exceeded any realistic LTV. Both programs killed.
Build the infrastructure
No off-shelf CRM connected to our production order data. Built a custom tool with a 6-stage pipeline, automatic lead sync, and lifecycle alerts tied to real account behaviour.
The failed acquisition experiments pointed at the real constraint: each AM was already at the ceiling of what they could manage manually. More customers without better tooling would have just created more noise. The bottleneck was operational, not commercial.
The Productivity Shift
Before the CRM, account managers manually assembled their view each week — pulling order status, checking revenue trends, flagging at-risk accounts. The CRM automated that intelligence layer. Each AM gets the right signal at the right time instead of building it by hand.
Four times more accounts per person, same headcount. The division is still growing at 30% month-on-month.
"The technology was the easier part. Sitting with account managers and watching how they actually worked — that's where the final pipeline stages came from. The first ones I designed looked right on paper and were completely wrong."