Back to all articles
Payments
CheckoutTap to PayPricingGo-to-market

Ordered Online, Paid at the Counter: How Tap to Pay Reached $50M a Month

Janahan Thavendran4 min read

TL;DR

  • Customers finished the online checkout, then paid at the counter, so the money moved outside our rails.
  • Changing the checkout and printing a receipt QR code did not work well enough, so we put tap to pay at the counter instead.
  • Starting with the biggest merchants and pricing with a floor took it from zero to $50M monthly GMV and revenue capture from 13% to 28%.
In this article

Before this project, I managed consumer payments, mainly the online checkout experience. In certain merchant segments, the data showed a pattern I couldn't explain at first.

Customers went through the whole online checkout flow. They chose their items, confirmed the order and finished. Then they walked to the counter and paid there. The online payment never happened.

They used our interface to order. The money moved outside our rails.

This was our biggest blocker for transaction revenue, and the data showed exactly where the leak was.

We tried to move the customer first

The usual fix is to change what customers do, so we tried that. We worked on the checkout itself, but customers who paid earlier ended up with a lower total check size, and that hurt the merchant's business. Then we printed a QR code on the receipt to push people to pay online. It didn't work well enough.

Putting the terminal where the payment already was

Around that time, tap-to-pay technology was becoming viable. I proposed we use it as the wedge into the outlet. Instead of forcing people to pay online, we would put a counter terminal, tap to pay, at the point where they were already paying. That way we capture the payment from every place it was leaking.

Leadership approved it. We built it with our existing merchant base, rolled it out, then iterated and scaled from there.

There was one more choice to make: tap to pay on the devices merchants already had, or a dedicated payment device. We went with both, deciding by each merchant's volume and the stability they needed.

The team, and one rule

At Delivery Hero, I led a small cross-functional team of about twelve people. On the go-to-market side: me, one enterprise salesperson, an ops lead and a customer success manager for the account. On the build side: two Android developers, three backend engineers and two QAs.

I insisted on one rule: no engineers in merchant meetings. Part of that was geography. Our engineering team was based in India and Indonesia, and we had no engineers in Singapore. But it was also a choice. I went into every client meeting and did the translation myself. With an engineering background, I could sit with a merchant, work out what they actually wanted, and come back with a precise problem for the engineers instead of a vague request. The merchant got a clear conversation, and the team got a clear brief.

Starting with the biggest merchants

We didn't pitch everyone. We went after the merchants with the biggest GMV first, because those accounts move the volume. The trade-off was reach: we chose a few big adoptions over a long tail of small merchants.

Of the top ten, five said yes, four said no and one never confirmed. The four who said no taught us a lot about what enterprise merchants actually cared about.

Pricing the product

I designed the pricing model, sitting down with the sales team to settle what we could offer, which features would come when, and which target merchants to go after, based on the growth we were seeing.

The model was interchange++. The merchant pays the real card costs plus our margin, with our margin ranging by payment method, issuer, average check size and volume. Every range had a floor. That mattered for the sales team. They had a band to negotiate inside, so they weren't guessing on every deal.

What happened

We took the product from zero to $50M in monthly GMV in about six months. Around 70% of that came from our existing merchant base adopting it. The other 30% was new business. In total, about 300 merchants adopted it, existing and new.

The number I'm proudest of is revenue capture. It went from 13% to 28%. Company-wide, in six months, we more than doubled the share of payment volume we were actually monetizing.

What I would test now

Looking back, online payment failed for a simple reason. We gave customers nothing for paying online. The counter was the easier path, so that is where they paid.

If I ran this again, I would give them a reason: a loyalty program, or a subscription with a specific discount model, that only pays off when they pay through us. I believe customers would choose to pay online instead of walking to the counter if the value was clear. That is the first thing I would test.

The customers had already shown us where they wanted to pay. We spent our time trying to move them to our checkout. What worked was putting our product where the money was already moving.