Interview with Dylan LETIERCE, co-founder of KNAP, who raised 3 million Euros of seed funding to implement connected trolleys in our supermarkets.
How did you get the idea for KNAP?
We were business school students and we wanted to create a company, a startup. It seemed clear to us that we had to find a problem and try to solve it. We searched and what we found most interesting, what we liked, was the queue in the supermarket. We started doing research and we saw at the time (in 2018) that solutions existed that were interesting, but very oriented to convenience stores, very small surfaces. It was very difficult, even impossible or not profitable to implement them in supermarkets, supermarkets or hypermarkets. This was the beginning of Amazon Go and it led us to our idea, which was to apply connected convenience store technology to connected carts for big box stores.
What exactly is the proposition you are proposing?
We provide an autonomous payment solution for supermarkets. For supermarkets, it is an automated checkout system that can be compared to a handheld machine or an automated checkout, except that we provide other services.
Above all, our device is secure, which is to say that it does not allow for fraud, which is what you often have with handheld showers. Today, with scanners, if you forget to scan an item, no one will see it, hence the loss of margin for the point of sale. So it is very difficult to make this solution profitable. This type of technology, even if it is cheaper at the base, eventually becomes unprofitable because it takes fewer employees to run it.
Our goal was also to solve a “trust” problem, which is to say that with these solutions, customers are randomly checked because it is not known whether they have committed fraud or not. However, this method of control is as unpleasant for the customer as it is for the point of sale. We solve this crisis of trust and we control the actions users perform on the trolley in an automated way thanks to weight sensors and cameras connected to various algorithms especially with machine learning etc.
This makes it possible to check that all the work done is consistent. Usually, an inconsistent verb is: “I offer one product and submit another” Or “I offer a bottle of wine at €5, I deposit at €50”, If we detect this type of action, we refer it to a store associate who can handle customers. They present “suspicion” and are therefore prone to fraudulent actions. Another advantage of the solution is that it is not a question of controlling the entire basket. This does not waste their time and those who do not cheat will not be investigated.
Our tool will be with the customer during his/her shopping session. This allows him to have access to the volume of his basket, his promotions, “product” information such as Nutri-score, allergens, etc. We can also offer recipes or suggestions. The implementation of this type of additional service depends a lot on the brand.
The idea came to you in 2018. You had to work on R&D for a long time?
R&D was mainly done when the company was launched. Until 2019, we formed our own technical team to do R&D, because we didn’t have all the skills in-house. Since then, we have been developing our prototypes which have been placed in stores. We wanted to incorporate them from the start so that they were in a real context, used by real consumers. Today we are releasing our commercial model, which is going into production in several hundred units.
What were the big steps from the moment you had your prototype?
Originally, we had prototypes on the table that we didn’t put in stores, but quickly we got points of sale to test it out. We have worked with the Monoprix brand since inception even though they are not our target customers. The advantage was that this brand at that time had a software infrastructure that could easily accommodate our technology. For its part, the sign wanted to innovate and communicate upon the opening of a new point of sale. It allowed us to gain visibility. After this test, we began to test them with Intermarche, in points of sale. In terms of date, the trial with Monoprix was just after the first confinement, so mid-2020. The second, testing with Intermarche, was due for early 2021.
Do you believe that imprisonment has been a source of acceleration or deceleration for your business?
Honestly, a little bit of both. It slowed us down but right after the lockdown we faced a surge in interest so I’ll put the ball down the middle. The lockdown held us back but we focused on R&D during this time when the market was not receptive at all. Sines had other jobs during his quarantine besides placing connected carts in stores. But, later, brands had a little money and were able to invest in innovation to solve problems related to COVID such as the risks associated with exposure of employees to the virus. That’s the interest in connected carts, even if it’s not their primary goal. After confinement, we faced a surplus of interest.
What happened after the implementation in Intermarche?
We tested them and faced a lot of problems as we faced them realistically. The ecosystem of a supermarket is quite hostile and it is very difficult to implement technology there. There are too many item types, so setting one up is a chaos. We were lucky that we had enough points of sale to report problems in real time so that we could resolve them. This has been the majority of our work since 2021. In short, we tested our carts, we improved them until we had the version we have today.
You just raised three million euros, for what?
The first reason was to fund our R&D. Today, we have significant costs in terms of R&D and we should be able to continue investing. The second reason was to produce the vehicles that we are selling. We’ve already produced about 400 of them and it takes cash to produce, manufacture and sell them. Later, the goal was to recruit a sales team. We had almost no one on the commercial, only me. Earlier, we focused on relationships with retailers and in-store testing, but didn’t have anyone to sell solutions to us. Now that we have a product, we have to sell it to the best possible customers and hence we need salespeople and a successful sales strategy.
What has been the biggest difficulty you have faced so far?
There were two real difficulties. First of all, in in-store tests, that is to say that our technology needs to be connected to the software with a system of signals and that connection can be quite complex. Apart from this, there were problems related to hardware, mechanic or electronics in the trolley. The first months of the trial were difficult.
Then, perhaps the point that has been very difficult is the post-Covid cash flow. The fundraising we did this year should have happened almost a year ago, but the lockdown slowed down the fundraising process. So living with a cash crunch hasn’t always been easy, especially when we had a lot of expenses related to pilots in stores etc. we had to play in the middle “save money” And “Keep Pilots in Stores That Spend Money”, It turned out to be difficult.
What are the major challenges in the future?
I don’t think it is the production of the cart as we are well established. They were very well designed. I believe the biggest challenge before us will be “A Computer on Wheels” Which is given in the hands of people who can misbehave with it so that it can be tested. As I was telling you, bullock carts are used in quite hostile environment where people are in a hurry, do not have time to waste or think about its proper functioning, etc. Customers are often less alert when they are in the store than usual, operating in a somewhat mechanical manner. The challenge lies in managing a fleet of several hundred vehicles as there will be a lot of work to do with vehicle repairs, software maintenance, managing those many computers online, etc. If there are any components that malfunction or break, they should be replaced quickly at low cost to the point of sale so that the trolley is operational again at the earliest. This is going to be quite a challenge!
Do you have a more approachable strategy by brand or is it geographic?
Essentially, it is by brand as it needs to be integrated but not everything is the same. There are integrated brands such as Carrefour or Casino, where there are very few exchanges with points of sale. It is enough to discuss with the group so that it helps us to deploy our carts. They generally have a very long term and well defined strategy.
Then there are independent points of sale such as, for example, Sistem U, Intermarche, Leclerc. There, it is a strategy guided by points of sale. It’s about finding a point of sale that interests us and that is interested in our solution, that is willing to move lines within their brand to adopt our solution. Generally, we work by sector as it works by buying groups etc.
“It slowed us down but we faced a surge in interest right after the lockdown, so I’ll put the ball down the middle. The lockdown held us back but during this time we focused on R&D. ,
dylan letiersco-founder of KNAP
3 Tips from Dylan Letiers
- To be well surrounded. I think it really does matter. Even if I don’t use all the advice, I have a lot of people who give me advice. Of course, you should be able to sort it out, but most of my decisions are based on other people’s advice.
- You have to be flexible and you have to be careful with cash. I think we have to think carefully about whether we need funding or not. You have to prepare this part well so as not to cost everything from the start. Our first fundraiser, we didn’t necessarily prepare it very well and as a result, cash flow was hard to manage afterwards. Technically it was not a fundraiser but funding that came in bit by bit.
- Compare your product with the market. We have some competitors who are very tech-driven, offering or showcasing their products but we don’t know whether it is working or not. So sure these technologies look pretty sexy on paper and can be interesting, but they haven’t been thought out from the ground up. Me, I’m always facing the market and then offering its product.