Tortoise’s Dmitry Shevelenko: Consumers Thought They Could Buy Something From the Robot

A wide-ranging interview on macro fundraising challenges, how to sell to Walmart, the vaunted "double pivot" and more

With the autonomous delivery industry in a momentous state of transition, I’m sitting down with execs from the leading PDD players to gather their insights into the challenges facing the sector. Summing up their thoughts on fundraising, sales, regulation, and more - we’re already starting to see some interesting trends emerge. (See these recent interviews with Cartken’s Christian Bersch and Serve’s Aduke Thelwell.)

Here I discuss the fate of the industry with Tortoise’s Co-Founder Dmitry Shevelenko. Dmitry’s unique background spanning companies like Facebook, Time Inc, Uber, LinkedIn and Superpedestrian made Tortoise a particularly fun company to watch: they started with self-driving scooters, before transitioning to tele-operated delivery bots, and finally they built a unique mobile vending robot. This interview with Dmitry is a long read (edited and condensed for clarity, italics added for emphasis) but I highly recommend you take the time to digest it - anyone else in the space would be quite well served by his lessons learned.

---

Jonah Bliss: Dmitry, great to catch up with you. Can you start by giving folks an overview of Tortoise, your founding story and pivots?

Dmitry Shevelenko: Yes, so the founding insight of Tortoise was when it comes to bringing automation to the real world: low-speed comes before high speed. Something that's going five miles an hour is a lot less dangerous than something going 50 miles an hour; remote control comes before autonomous. The very first incarnation of our technology launched in late 2019, early 2020 as a solution for shared scooters, to let a remote operator in Mexico City look out a camera and re-park a scooter in Minneapolis if it’s in front of somebody’s front doorstep. Or say, return it to a safe parking spot to get recharged and allow for the rebalancing of micromobility fleets. 

This was, of course, a fantastic time to be in the shared micromobility space because literally the weekend I was supposed to launch our first market in [Peachtree Corners] Georgia - all national air travel got shut down for Covid. And all movement of people in cities froze; so pivot number one was was realizing when people are frozen, the movement of goods becomes even more important. And we obviously all lived through this explosion in demand for last mile delivery. 

And so we'd already built this ability to remotely drive a light electric vehicle from A to B. We just needed to create a new form factor, so we built our own remote controlled delivery robot, with a focus on grocery. And we built a robotic platform that was modular and had different container types on it. We built a larger form factor, specifically to handle your typical online grocery order, which we didn’t think was being served by the existing players. Again the really key insight was this pure focus on on tele-ops and and keeping the bill of materials low by not trying to layer in lidar and uneconomic features, and just keep this something simple that we could launch right away.

We were able to get big grocers on board fairly quickly because there was a lot of interest in scaling last mile solutions with a surge in demand during Covid. So we got on board Shoprite and Meijer - national grocery chains - and they started doing pilots with us. But if your mandate is to deliver anywhere within a three mile radius of the store, you're running into a lot of edge cases on a day-to-day basis. Even when you have a 100% tele-operating solution, you have all the variances in sidewalk infrastructure, plus first generation hardware. So that made things not always the most reliable.

It became clear that this was not a business model that was only a few months away from profitable unit economic operations, at a price point that made sense for those grocers. But as we kept doing more pilots, we noticed a really interesting behavior which was that when the robots were parked in front of the grocery stores, they'd get a lot of attention. People would walk up to the robot and start talking to the robot and expect to interact with it. We just didn't know why or what they thought would happen. And so we started asking people, “Hey, you started talking to this robot that wasn't talking to you. What do you think is going to happen?”

We consistently heard the same thing which was “I thought I could buy something from the robot.” And that was pivot number two for us. It was this moment where we're like, “Hey, you know, maybe the killer use case here is selling stuff instead of delivering it.” We have this modular container system and each robot has two containers. Embed a tap-to-pay NFC reader in the container, which already has a Bluetooth controlled Internet connected locking mechanism, and within a few months we turned our delivery robots into the world's first mobile vending robots. As Curbivore fans might know, we debuted that at Curbivore ‘22. This product worked really well in dense environments where you have large crowds: shopping malls, sports venues, conference centers, anywhere you have a lot of people and long lines.

So we got interesting, commercial traction with NBA Arenas, NHL, MLB, large venue operators, Walmart, and Mall of America. So, there was a lot of interest in the solution but not not enough to overcome the macro environmental change which was that raising capital became much more difficult for unprofitable hardware startups. Unfortunately, we weren't able to see it through.

JB: I appreciate you being so candid about that. I’d like to touch on something that you seemed really successful at: you were able to make these pretty big pivots and pull them off quickly. But then just as quickly, you were able to sign up marquee brands. Do you have any advice or learnings about how you all were able to sell through to large companies so well?

DS:  I think one of our advantages was the design of the container; it was literally a white box with no Tortoise branding on it whatsoever. So, we offered the marketing value of getting your own large-scale, branded robot. In approaching a lot of these larger brands and retailers, we could very quickly generate a rendering of what the robot could look like, in a very photorealistic way. Generating those customized assets made it feel like something real and enticing. 

And I think the other piece that helped was our business model: no upfront costs, but we’d take a percentage of sales. That makes it easier to push through a procurement process from a budgeting point of view. And I think the other thing, particularly with the mobile vending, is there was no systems integration needed. We provide the payment processor, there's no app, there’s no website. All a retailer had to do was provide us with their marketing assets, so we could design the container wrap, and have a place to store the robot when it's not in use. The trick is to solve as many bottlenecks on the customer side as you can take on yourself. 

We were able to launch Walmart, where we had the mobile vending robots in their parking lots. We hired somebody on our side to restock the robot, so Walmart didn't have to train an associate. Because we knew that would be a deal breaker with Walmart operationally. So I think that those are the types of things you need to do: things that don't scale, but let you quickly bring the product to market, so you can you discover the consumer experience, see where things break, see what works and what doesn't. And if you have momentum, you can always figure out the more scalable answers later

JB: I'm curious if you've any sorts of you know statistics you can share in terms of your sales cycle - from first outreach to deployment. Additionally, what kind of revenue were you doing per unit?

DS: Because we had a sexy solution, because people love the idea of “a robot in a stadium” - sometimes we could go from our first conversation to a signed MSA in under two weeks. Especially if they had like specific games where they wanted to deploy the robot, which would add useful pressure on their side.

I think it's sales 101: finding a champion inside these large organizations, having empathy for that person, knowing who their stakeholders are, and just arming them with everything they could possibly need to push it through internally. It wasn't the case with most of these that we got a CEO level relationship; it was about finding a champion and and making them successful internally. 

In terms of the sales per robot, some days, at an event like Comic-Con, each robot was generating $3,000 a day in sales. And then we had other robots that would generate zero dollar sales days. It was hugely variable, depending on the volume of the crowds.

“Think of the potential strategic partners that might have upside in your success and try to get them on your cap table.”

That was actually one of the challenges in the business model. If we could do it all over again, I’d package the product less as the main benefit being direct sales and more as a marketing lever. Say, charge a fixed fee for the exposure that you get through the robot; I’d have the CPGs play a role in covering costs. Because we dealt with the whims of foot traffic, and of the venue operator or retailer having wisdom about what is the right product to sell in each location.

JB: With the whole industry going through a period of turmoil, are there any words of advice for your peers? Any North Stars to look to, or things to avoid?