Sahra-Josephine Hjorth
Back to blog

← Blog

AI in Practice

Your Competitor Gave Your Retailers a Free Tablet. Now It Sees Your Sales.

A free tablet that a supplier gives a small store does more than record its sales. It becomes the screen the shopkeeper looks at when deciding what to order next, so the supplier that provides it is present at the one moment every other supplier is trying to reach. For a manufacturer that sells through small retailers, that place in the reorder decision is worth more than the data, and winning it, or working around a competitor who already has it, is a strategy decision, not an IT project.

Sahra-Josephine Hjorth in a red floral suit standing on stepping stones over a pond in a tree-lined park

Where Software Projects Go to Die

I own an AI studio, and through it I work on AI projects for companies around the world. In this series, I write about some of the biggest mistakes I have seen in AI and software implementation projects, so that you can avoid making the same ones. This time I want to write about a move that worked. Unfortunately for my client, its competitor was the one who made it.

In the first article in this series, I wrote about a paint manufacturer¹ that tried to find out what was left on the shelves of the small hardware stores it sells through by asking the stores to count every can, every week. It failed, and the article ended on the day we visited five of those stores and found that three of them already rang up every sale on a free tablet from my client's competitor. The lesson of that article was about data: stop asking people to produce it for you, and look for the place where it is already being recorded.

This article is about something I only understood later. The data was not the most valuable thing the competitor got out of that tablet. In the stores we visited, the tablet was not only where sales were rung up. It was also where the person behind the counter checked what was running low and decided what to order next. My client's competitor had not just bought a view of the market. It had bought a place at the moment every supplier in that store is competing for.

The screen a shopkeeper reorders from gives its owner a place in the decision

A small hardware store carries several brands of the same thing, such as three or four kinds of white paint and several makers of brushes and rollers. When something runs low, someone has to decide what to order to replace it, and from whom. In the stores we visited, that decision started on the competitor's tablet. The shopkeeper opened it, saw what was running low and decided from there what to order. The order itself was then placed with each supplier separately.

Supermarkets have understood for a long time how much power sits at that moment. Many of them pick one manufacturer as the "category captain" and give it, in the words of two economists who studied the practice, "authority concerning retail shelf space allocation" for the whole category, including the space for its competitors' products. A supplier whose system shows a small store what is running low holds a quieter version of that position. Nobody appointed it, and the store still makes the decision, but it makes the decision while looking at that supplier's screen.

Even a completely honest system gives its owner two advantages. The first is the product list. Whoever maintains it decides how every brand is named and grouped, and which one comes up first when someone types "white paint". The second is time. The owner sees a product running low before the store has ordered anything, and long before any other supplier sees an order. A manufacturer with that information can make sure its own paint is the one in stock, on promotion or already on its way when the store reorders.

Figure 1

Whoever owns the screen decides which brand comes up first and sees the shelf empty before anyone else

How to read it. A simplified version of the reorder screen on a tablet like the one we saw, drawn for illustration. The shopkeeper types "white paint" and three products come up. The company that supplies the tablet keeps the product list, so it decides the order they appear in, and it sees its own paint running low before the store has placed any order. None of this requires the system to be dishonest. The brand names are placeholders.

As far as I know, my client's competitor never used the tablet to take the orders itself, and the store still ordered from whichever supplier it chose. That makes the point stronger, not weaker. The competitor did not need to handle the order to influence it. It only needed to be on the screen where the decision was made, and my client was not.

Companies have been paying for a place at the till for more than forty years

The idea is older than tablets. In 1980, a market research company called IRI gave scanners free of charge to 15 supermarkets in two small American towns, and the purchases of the local households that had signed up for its panel were sent to its computers in Chicago. That was a deal for data, and it became the start of a large market research business.

The version that goes all the way to the decision happened in China. Through a program called Ling Shou Tong, Alibaba gave small corner shops a free app for ordering stock, with next-day delivery, and selected shops also got a point-of-sale system and a full makeover of the store. Well over a million shops joined. According to one analysis, Alibaba used the data from the whole network to recommend which products each shop should stock for the people living within 300 meters of it, and the shop then ordered those products from Alibaba through the same app. The data told Alibaba what to recommend, and the recommendation turned into an order.

In Mexico, the Coca-Cola bottler Arca Continental bought a company called Brío in 2018. Brío puts a touchscreen system in corner shops that lets them take card payments and sell phone credit, and Arca was open about why it wanted it: to collect information about what is bought in traditional shops. By 2020 it had 10,500 of these systems installed.

Figure 2

Most deals for a place at the till bought data, and Alibaba's also bought the order

WhoWhat it gave the shopsWhat it got back
IRI1980, United StatesFree scanners for 15 supermarkets in two small townsPurchases of its panel households
Arca Continental2018, MexicoA touchscreen payments system for corner shops, 10,500 installed by 2020What sells in traditional shops
AlibabaLing Shou Tong, ChinaA free ordering app for over a million shops, plus checkout systems and makeovers for someSales dataThe order itself

How to read it. Three companies that gave shops equipment or software for free, and what each got in return. Orange is data about what sold. Pink is a place in the decision about what to buy next. IRI and Arca were open that they wanted the data. Alibaba used the data from its whole network to recommend what each shop should stock, and the shop then ordered those products from Alibaba through the same app.

Sources: Funding Universe; El CEO; El Financiero; RSM Digital Strategy.

A place at the counter is expensive to keep and easy to lose

Most manufacturers should not try to copy what my client's competitor did. They make paint, soft drinks or bread, and running a system in thousands of shops is a software business. Someone has to install it, train staff who change every few months, keep the product lists up to date for every brand in the store, and fix it on a Saturday morning when it crashes with a queue at the counter. A shop that loses a morning of sales because of your tablet will put the old cash register back, and your place at the counter is gone the same day. Alibaba could carry that cost because software and logistics were already its business. Most manufacturers cannot.

The place is also only as secure as the shop's trust. A shopkeeper who starts to feel that the tablet favors one supplier's products stops following its suggestions. They may keep ringing up sales on it for a while, because switching is a hassle, but they start looking for a way out, and the first supplier with a better offer will find them easy to win. The steering that lasts is the honest kind, based on what actually sells in that neighborhood, whatever the brand. Honest steering still leaves the owner with the product list and the head start, and in a market where most suppliers see nothing, that is enough.

There is a legal limit as well. In the United States, the tobacco company UST served as category manager for the retailers that sold its products, and its rival Conwood accused it of using that role to push Conwood's products out of stores. The jury's award, tripled under antitrust law to $1.05 billion, was upheld in 2002. In 2022, the European Commission accepted binding commitments from Amazon to stop using the non-public data of the sellers on its marketplace to compete with them. Neither case says that collecting sales data is illegal. Both show what happens when the company that runs a platform also competes on it and uses that position against its rivals. If you put a system in your retailers' stores, put the deal in a short agreement the shopkeeper can read in a few minutes, saying what data is collected, that the store's own records belong to the store, and what the store gets in return.

There are four ways to get closer to your retailers, and only some of them give you a say in what they reorder

If you do not have a place at the counter yet, there are four ways to get closer to it. They differ in what they cost, and more importantly in what they give you. Some give you only data about what sold. Others also give you a place in the decision about what to order next.

Fact box: four ways to get closer to your retailers

Build your own and give it away, as my client's competitor did. This only makes sense when your retailers' cash registers record nothing but the money, so nobody else is at the counter yet. It gives you both the data and a place in the reorder decision. You control the system, and you pay to run it for as long as you want that place.

Buy the company that makes the software your retailers already use. This works when most of them run on one or two systems. It gives you both the data and a place in the reorder decision, it is the fastest way to reach many retailers at once, and it costs the most up front.

Partner with the software provider, for example by paying to put your ordering, promotions or catalog inside the system the retailers use, in exchange for sales data. It always gives you data, and it gives you a place in the reorder decision only if your ordering sits inside the system. It is cheaper, but the provider can sell the same access to your competitors.

Pay for the data, from the provider or a market research firm. It is the cheapest route and gives you data only. It tells you what sold, usually late, but it gives you no place in the moment the retailer decides what to order.

Figure 3

The cheap routes buy you data, and only the expensive ones buy you a say in the reorder

How to read it. The four routes from the fact box, placed by what they cost you, to the right, and by how much say they give you in what the retailer orders next, upward. The dashed line under the partnership shows that its say depends on the deal: with your ordering inside the provider's system you get some, without it you only get data. The positions show the order of the routes, not measured values.

If a competitor got there first, you have to win the reorder somewhere else

So what do you do when you arrive second, as my paint client did? You cannot offer the same tablet a few months later, because the store already has one that works. These are the options I put on the table for my client.

The first is to give the store a better reason than the one it already has, such as faster deliveries, credit, or better prices on what sells best in its neighborhood. A competitor's system becomes much easier to replace when yours comes with something the shop can feel in its own pocket.

The second is to own a different moment. If you cannot own the till, you may still own the ordering. Suggestions for what to reorder, based on what the store bought before and built into the channel it already uses to order from you, such as your sales rep or a messaging app, put you back in the decision without asking the store to install anything new.

The third is to go one step up the chain, to the distributors and wholesalers that supply the stores. They see orders from many shops at once, and some of them will share that view if you make it worth their while.

The fourth is to protect the part of the market that is still open. If the stores in one region already run on a competitor's tablet, look at the regions, store types and channels it has not reached yet, and get there first.

I no longer work with that client, so I do not know which of these it chose. None of them is as easy as having been first, but every one of them is better than asking the stores to count.

My client asked the stores for numbers, and its competitor gave them a new cash register

When I first saw that tablet, I saw a clever way of collecting data. I now think the data was the smaller part of what the competitor got. My client was still asking the stores for numbers. Its competitor had become part of how they decided what to put on the shelf.

The best data strategy I have seen did not look like a data strategy at all. It looked like a gift to the shop, and what it bought was a place behind the counter.

¹ Every story in this series is real. To protect my clients, I change the details that could identify them, such as the industry, the products, the company's size and location, and the names of people and suppliers. What happened, the decisions that were made and the reasons they went wrong are kept as they were.

Sources

Link to this essay
Follow me