Ninja Delivery: 10 minutes or bust

Previously I talked about building tele-operated robots, running a food delivery service, and then scrapping the robots to launch Canada’s first 10-minute grocery delivery service using e-bikes. The 10-minute delivery service experiment we ran in Waterloo was a success, growing 20%+ WoW throughout the summer of 2021, so what did we do next? We raised another $2.8M to expand the service to Toronto (aka downtown Canada, the country’s largest and most densely populated city).

At the time of writing this, we are deep into a market downturn. Many of Ninja’s competitors have gone bankrupt, others have slashed their valuations dramatically, everyone is laying off employees, VC twitter is dunking on instant delivery startups left and right.

I usually write about projects chronologically and try to capture the exact order of events as they developed, but as of May 2nd 2022 Ninja is no longer taking orders, so let’s skip ahead to what happened: Almost exactly 1-year into its conception, after fulfilling 30,000+ orders, crossing $1M in revenue run rate, expanding to 3 dark stores, running successful campaigns with world-renowned CPG brands, Ninja was acquired. Won’t go into details about and deal and won’t be buying a lambo with the proceeds either. We started Ninja with the intention of building an empire and that did not pan out, this outcome was strategic as we looked to do what was best for our employees, investors and customers given the evolving market condition. In this post I’ll talk about some things I thought were cool about Ninja.

Let’s start with why it made sense to build Ninja, a 10-minute grocery delivery service, in Canada. The key insight here is that three companies hold 63% of the Canadian grocery market share, this translates into superior margins compared to US counterparts. The EBITDA margin for Canadian legacy grocers is 9% (5.6% EBIT) versus 4.9% for US based legacy grocers (2.7% EBIT). This alone makes Canada an appealing market for a model which struggles with its bottom line.

Canada’s smaller population means a smaller TAM with a $78B grocery market versus $757B in the US. Although Canada’s market size is comparable to Australia’s $89B, only a mere $7M has been invested in instant grocery startups in Canada (which basically all went to Ninja and Tiggy) compared to $78M in Australia and $4B in the US. This makes the case for Canada being the last blue ocean in this space, hinting the potential for a ‘land grab’ opportunity.

Last, Canada’s grocery e-commerce market has a lower penetration than the US, the exact figure varies depending on your source and how long since the last COVID lockdown but I read 1% vs 5%. Other markets love the convenience of ordering groceries online and Canadians will love it too, the market has just been underinvested for some time given the large market share of legacy grocers and them not having any real incentive to disrupt the industry.

So to sum it up, Canada’s rapid delivery market is comparatively underinvested, has less players and better margins. What Ninja did was simply apply a novel model with global traction and VC interest, into arguably one of the best suiting markets in the world.

I won’t try to sell you on the 10-minute delivery aspect of the business, at this point you either get it or you don’t. My take is that customers will always want to get things faster and not slower, you can take the limit on this all the way to teleportation. Some will question whether we really need to get stuff that fast, others will doubt the economics of the model. I strongly believe that if it was up to some of you, we would still be living in caves.

Ninja’s business and team could be split into Growth and Operations. Growth was responsible for bringing in new customers, making sure they are placing frequent and big orders, building a great customer experience that leads to high retention, and doing it all with capital efficiency. Operations ran the show, everything from running the dark stores, managing the supply chain, managing the delivery team, building processes, standardizing day-to-day operations, order batching, delivery route planning, and squeezing more efficiency out of everything that happens from order to delivery.

At first glance, a dark store is ‘just’ a convenience store that does delivery. Delivering thousands of SKUs in 10-minutes though? That requieres operational excellence and there’re opportunities to gain efficiency throughout the delivery process with measurable impact (I call it the Formula 1 of commerce). Where to place the dark stores in the city to minimize delivery times to customer clusters and maximize the addressable audience, indexing shelves for faster picking, packing orders live as customers add items to their carts, shelving products frequently bought together closely to reduce picking time, arranging the interior of the store so that you can fit as many SKUs as possible, optimizing shift schedules to have just the right amount of staff when needed, these were just some of the things our Operations team handled. No detail was too small, we would spend time ironing out simple things like making sure doors were easy to open with a bicycle, improving the design of the delivery backpack so that it’s easy to open and minimizes the chance of spilling, and even choosing the right chairs to use at the stores to promote an active workplace. Another huge part was building a team culture such that everyone in the team cares about the 10-minute promise, something that became increasingly more difficult as we scaled.

Our Operations team had world-class talent, including a Zomato early employee who launched their food delivery business, a Lime GM who lead their most profitable market, and an Uber head of operations who managed the largest Eats market in North America. The best part of building Ninja was getting to work with ridiculously talented people.

Back in the food delivery robot days I had been responsible for promoting the service both on social media and around town, this made me the de facto Growth lead as we started to expand Ninja. Mind you up to this point I had no direct experience in growth, marketing, or product. In fact, no one in the team had a marketing background, we only hired our first ‘real’ marketing person in Q1 2022. Our approach to Growth was the same approach we took to building robots: reason everything from first principles. Assume nothing, formulate hypothesis, run experiments to validate them, measure it all, and continue to place bigger bets on what the data shows is working (Highly recommend watching this presentation from Gustaf Alstromer if this is not how you run your growth team).

We knew what had worked in Waterloo to kickstart Ninja, but as we landed in Toronto we had a bigger audience, more channels to reach them, and a bigger budget we needed to allocate effectively. Below I’ll expand on some early growth initiatives that I think were interesting. Won’t be mentioning the obvious SEO/Paid Acquisition/Emails but of course all of that is important and works too. For context, Ninja’s growth team was optimizing for: New Users, Retention, and AOV, these metrics drove decision making and defined the objectives of each initiative.

• The art of the flyer. Ninja’s first customer ever placed her order after receiving one of our flyers under her door. This was in Waterloo where you could walk into student housing and slide as many flyers as you wanted under every door, put them up on elevators, and even on the lobbies. This worked well and we spammed it to the point building managers were calling us and asking us to stop. This medium however would not work in Toronto, where buildings had concierges and security, even when you managed to sneak in, stuff would get taken down immediately. We worked around this in 3 ways:

  • Branded paper bags: Originally we were doing deliveries using plastic bags we would buy off Amazon. We upgraded to recyclable paper bags, branding these had a long lead time so we started by placing stickers on top. We then started dropping these bags in lobbies across the city. You see, a concierge will take down a poster, but a food delivery bag flies under their radar. The bags would sit in the lobbies where everyone could see them, and the branded stickers would catch their attention and incentivize an action through coupon codes and extra discounts for sharing or interacting with the bags through scanning a QR code (classic AIDA playbook).

  • Door Hangers: Our first successful flyer campaign in Toronto was printing door hanger shaped flyers and putting them on the front door of every house around our stores. It was often our own delivery team putting these out during their downtime, and we took the time to include a hand written personalized message in each flyer to emphasize we are a new small local business. We literally spammed these until we were getting calls from angry neighbours (There is something to be said about flyers wasting paper and polluting).

  • Direct Mail: Running mail campaigns through the postal service was relatively expensive but it allowed for a bigger distribution scale than anything we’d done before. It was also easy to measure and attribute orders, easy to A/B test different coupon structures and creatives, and also introduced new variables such as frequency of mailing, and pairing mail campaigns with other mediums such as OOH or digital. Targeting households based on demographics and census data was good as it allowed us to tune our message for specific customer profiles, but the most impactful feature of direct mail targeting for our business model was the ability to target every household within 5 minutes from our stores. These were households we knew we would deliver to in less than 10 minutes, ensuring a 5-star experience and allowing us to be more aggressive with our promotions (and ergo increase conversions) with campaigns such as “get it in 10-minutes or it’s free”. We had many tricks to maximize and even ‘hack’ 5-star customer experiences, these were great and would compound as these people would share their experiences with friends and leave favourable reviews.

• Weaponize every interaction. Every touchpoint a customer had with our brand was an opportunity to provide a great experience and get them to talk about us. Here’re two simple ideas which seem obvious but I see a lot of brands and competitors not implementing them:

  • Thank you cards: Originally we were trying to figure out how to get people to share us on socials, so we thought, why not just ask? We started to include thank you notes with every order, these had a written component to make them feel personalized, and a simple call to action: Impressed by our service? Share us on social media and tag us to get $5 off your next order! As simple as it sounds, this worked and people were regularly sharing us over Instagram and TikTok. We would then leverage their posts and compile them on our main page to create social proof.

  • Receipts: Some customers will opt out of your mail marketing but no one will opt out of email receipts, this made receipts and order confirmation emails a great place to promote stuff like our referrals program. We also made receipts shareable by highlighting how much money you had saved today by ordering Ninja, or how fast your order got delivered today in minutes and seconds. Imagine seeing a friend got something delivered in 5 minutes? That’s worth a double take.

• Feedback loops. In order to run growth experiments and validate hypothesis, you’ll need to have the tools to measure all of the relevant metrics. Some metrics are qualitative and harder to quantify, and at our scale we had the opportunity to simply call every single one of our customers (we had a person on the team whose entire job was dialling customers). Also, as good as your team might be, talking with customers and hearing what they have to say will uncover insights that the data could not have always told you. Every single person who ordered Ninja for the first time would get a phone call within 7 days of placing their order, similarly VIP / high LTV customers, and churned customers would also get follow up calls to gather insights around their experience and promote growth initiatives (listen, you can launch a referral program, but it won’t work unless everyone knows about it). These calls helped uncover which were the true value props of our service, hone in on customer acquisition attributions, expand our SKU selection (customers loved when we added products they specifically asked for), and get customers’ thoughts on changes we planned to roll out in the future. The phone calls themselves were also an experiment: Do customers who got the call retain better than those which did not? Was spending time calling them justified?

• Leverage existing platforms. While we were trying to funnel people into our app, we would still leverage existing platforms including UberEATS and DoorDash for top of funnel initiatives. We did this in two ways:

  • Ninja listed on UberEATS as a grocery store under its recently launched (and hence not very crowded) grocery section on their app. The timing was great as Uber was investing heavily into promoting grocery purchases on their app. Customers who ordered from Ninja through Uber’s app would then be encouraged to download our app for a wider selection, faster delivery, and a discount on their next order. Uber also allowed to run ads on their app, and indexed every product so customers searching for specific things would land on your store if no one else had it (which was the case for many alcohol and COVID-19 products).

  • We would offer popular restaurants on UberEATS money in exchange for them to include our flyers with every order they shipped. We would often word this so it looked like the restaurant was gifting their customers free dessert through Ninja (a win for the restaurants and for the customers).