Reflections on Troav, my Delivery Startup

I started one of those 10 minute delivery startups during Covid - It didn’t work out. I learned a ton, and have a lot of lessons I’m taking forward into my next project.

Why Delivery

We were psyched about delivery when lockdowns began. It was clear that Covid was rapidly accelerating adoption of online delivery services. Volume across Instacart, Doordash etc. was up, but more interestingly, average customer age shot up. Delivery was transitioning from a luxury for young and single customers to families with higher AOV and repeat behavior. Covid was was a key answer to the “why now” question.

We also had a lot of experience with last mile delivery - it’s always been the least efficient and most expensive part of the delivery process. With higher volume, changing customer preferences, we saw (like many others did) an opportunity to turn last mile delivery into a core business.

Most models used microwarehouses - small hubs where ~2000 SKUs of product would be stocked, and e-bikes would complete 1-2 deliveries per batch. Because the warehouses were already close to the customer, delivery times would be short. This had the dual purpose of actually reducing labor per delivery, driving down costs to the point that even small batches of deliveries could remain profitable. We had a slightly different model (vans as microwarehouses), but it accomplished the same goal of starting products as close to the customer as possible.

The Customers

One question we got a lot was, “Who needs 10 minute delivery”. We looked at it a different way: Customers want to get things when they want them, and they’ve gotten used to taxis that come to them and a massive library of TV shows they can watch anytime. It’s inevitable that commerce was next. Delivery speed is a large point of friction to e-commerce whether customers knew it or not, and Amazon sales surged every time they shrunk their free delivery time.

And we saw the same thing with our customers - reorder rates were best-in-class, and we got rave reviews. Fast delivery is a product people want, and a behavior they became accustomed to. People would tell us time and time again that 2 day delivery and even 1 hour delivery started to feel slow, and they hated planning around it.

What didn’t work

So what didn’t work? We boiled it down to two things: Strategy and Capital.

Strategy

Intuitively, this kind of product works best in dense urban areas, right? It’s a volume game, and with a high volume of orders, you can batch orders more efficiently and cover your fixed costs. So we began with big dreams in Downtown Boston. We soon saw our numbers never really grow there - in fact, they were solid in all the outskirts of the city - the Peri-Urban regions. This was right out of the Doordash playbook, and something we were hesitant to fully commit to because of the allure of customers in extremely dense areas.

Urban areas are the most likely to have existing food/grocery infrastructure accessed with walkable behavior. Delivery didn’t represent a huge value add for them. Areas where cars are the norm use delivery services the most. This is a strategy decision that helped Doordash steal market share from Uber Eats. This wasn’t possible for most microwarehousing solutions. E-bikes can’t easily serve suburban distances and don’t scale as well as cars can across long distances.

Our unique van model could alleviate this, but we took too long in fully committing to it.

Capital

Capital is a moat in an asset-heavy business like this. High Opex and relatively high CAC meant economies of scale were only really reached after long periods of negative cash flow. This isn’t unheard of at all - but it wasn’t something I was keen to do. We weren’t seasoned entrepreneurs with large networks. We had the right background for a tech startup, which is what we wanted this to be - but it wasn’t.

We should have committed to raising large amounts of capital with the intention of growing quick and getting customers in volume, even with bad scale economics at the start. But with full hindsight, that realization should’ve prompted us to focus on an area more aligned with our strengths.

I wouldn’t trade this experience for anything else. I think we did a lot of things right, but there are a lot of things I’d do differently. I’m taking all those lessons and applying it to our next venture - a software product for renewable energy construction. If you’ve read this far, thanks :) I can be reached at harshal@fliteworks.com

Note that most of these takeaways are only applicable to the USA market. Labor costs, CAC, SKU selection etc. all differ greatly across markets and regions.