Growing orders, growing stores and a retention chart that looked like a warning light. That was Perpule’s first big lesson.
Co-founder Abhinav Pathak joined The Builders Club to walk through a pivot from self-checkout to retail billing systems, the enterprise sales grind, and the decision to sell to Amazon.
What can founders learn from a retail tech pivot? Abhinav Pathak’s core thesis
Abhinav’s central point is that headline growth can hide a broken business. Perpule’s self-checkout app scaled by orders and stores, but customer retention was poor, and the cause was inaccurate store inventory, prices and promotions caused by legacy billing systems.
The team fixed the root cause. It moved from a B2B2C app to a pure B2B point-of-sale product that retailers could run on modern hardware, where retention was strong and switching costs were high.
- Watch retention early. Rising orders can mask churn, so check cohort retention from the start.
- Fix the layer underneath. The self-checkout idea needed accurate store data first.
- Zero to one is persistence. Landing a first large customer took months of repeated pitches.
- Do not over-customise. Building too much for one big client cost eight to nine months.
How did Perpule start with self-checkout?
Abhinav studied at NIT Surathkal, where meeting entrepreneur seniors made startups feel relatable, and later worked at Goldman Sachs. He met his co-founders there and through college, and they chose retail because it is large and technologically behind.
Perpule began with a self-checkout app for offline stores, as online giants threatened physical retail. Landing the first customer, Hypercity, took about three months of pitching to 30 to 40 brands, five or six times each. The pilot went live in late 2016 with a few stores.
Why did retention expose the real problem?
As the app reached around a hundred stores, roughly 10 to 12 percent of store transactions ran through it. Yet retention was weak: about 25 to 30 percent in month one and only a few percent after five or six months.
The team made some 500 customer calls and found customers liked the experience but lost trust when prices, promotions or inventory in the app were wrong. Stores updated items in their systems late, so the data was wrong.
- Customers will try a new experience, but grocery buyers expect price accuracy.
- Retailers already knew their systems were weak and encouraged the team to solve it.
- Abhinav says the team lost several months before checking retention.
How did the pivot to retail POS work?
The conclusion was that experiences like self-checkout or hyperlocal delivery would not scale until the store’s billing system was modernised. Many retailers ran billing software that was 10 to 20 years old and not built for cloud integration.
In 2018 Perpule became a B2B company building point-of-sale systems for large-format retailers, and raised a Series A.
- Build a billing app that runs on the Android card-payment machines that stores already own, which cut hardware costs.
- Win the same early customers back with the mobile version.
- Add a desktop solution when retailers asked for it, then scale to thousands of stores.
- Take the product to Southeast Asia and the Middle East.
Why are enterprise sales cycles so long?
Abhinav described sales cycles of about three months and deployment of three to four months. Store teams resisted change, so Perpule set up learning and customer-success teams, including messaging-based training for cashiers. The upside is retention: he said they never lost a customer in the POS business.
What mistakes should enterprise startups avoid?
Abhinav said that early-stage product and sales grow faster than technology, so technical debt is a sign of growth. The answer is a strong team and a willingness to launch before the product is perfect.
His biggest regret was over-customising for one very large retailer for eight to nine months, because the work was not reusable. The company was raising money and felt it could not say no.
- Keep a focused product team separate from sales.
- Talk to other founders, who have often faced the same demand from big clients.
- Do not get emotional about one customer: look at the business as a whole.
Why did Perpule sell to Amazon?
Abhinav said the team never set out to sell. Retail is a slow market, and the pandemic halved their volumes, so growth at venture pace was uncertain.
Acting on what was right for customers, employees and investors, and with relationships built with corporate development teams since 2018, they sold to Amazon. He stayed on to lead small and medium business technology there, citing Amazon’s process-driven culture as a learning opportunity.
Further reading and sources
Frequently asked questions
How do you know when to pivot a startup?
Look at the metrics that reflect real value, such as retention, not just growth in orders or stores. Perpule pivoted after customer calls showed the underlying problem was outside its control and would block scale.
How do startups win their first large enterprise customer?
Expect persistence to matter more than polish. Abhinav’s team pitched dozens of brands several times each before one agreed to a pilot, then built credibility from that proof.
Why should startups avoid heavy customisation for one client?
Software businesses depend on reuse. Abhinav spent months building features for a single large retailer that no other customer could use, so he now recommends a dedicated product team to decide what is built.
Conclusion
Abhinav Pathak’s journey shows that the right metric can save a company: retention told Perpule that a growing product was not yet a good one. The fix lay beneath the app, in the store’s core systems.
Check your cohort retention this week, and if it is weak, make customer calls before you add more growth spend.
Want more conversations like this? Join The Builders Club, our community of 35,000+ founders and CXOs, and explore more episodes of the Startup Founders Podcast.
