The Sugat deal as a case study: "Data is a bonanza that can push competitors out"

When one company wants to acquire another, the Competition Authority is accustomed to examining familiar questions such as what is the market share of the companies, who are their competitors, whether after the deal they will have the power to raise prices, and whether other players will be able to continue competing with them. However, in recent years, another question has entered the equation, which was once at most a secondary matter: who holds the data on the customers.

Recently, the Authority published a draft of a new document for examining mergers, which replaces a document about 15 years old and relates, among other things, more broadly to the power derived from holding information. Tobi Harris, director of the new Markets Department at the Competition Authority, is the person who manages all merger checks in the economy, with the data issue falling under her department.

"Approval of a new merger between companies can raise concerns these days due to data issues," says Harris in an interview with Globes. "This is something we examine more deeply and try to understand its implications in some of the deals that come to our desk."

Harris, 39, immigrated alone from the USA at the age of 19 and arrived at the Competition Authority already during her internship as a lawyer. Over the years, she has filled a series of roles in the Authority, and in 2018, with the establishment of the Markets Department, she was appointed head of the finance team. In the summer of 2023, she was chosen as deputy director of the department, and last spring she entered the position of department director.

The department she is in charge of is one of the central junctions where the state meets the business sector. About 25 employees, two-thirds of them economists and the rest jurists, handle about 200–250 mergers a year and dozens of other collaborations and arrangements between companies. The department is divided into five teams: finance, transport and tourism, food and retail, energy and infrastructure, and communications and health — and at the same time, the department participates in government discussions on regulation.

Ability to take over the market

Harris suggests thinking of data as a component with business significance, just like raw material or infrastructure. "There is information that is an input," she says. "Facebook profits from data and it can be seen as an input and it can be seen as something of value. I can give myself, as a leading company in a certain field, preferential access or block others from touching this data."

The risk becomes greater when the information allows the company to understand not only the market, but also the specific customer in front of it. Harris brings as a hypothetical example the deal that was examined in the past between Harel and Isracard and was not approved by the Competition Commissioner. "If Harel, as a leader in the insurance field, is the only one that has data of Isracard customers, and knows with its help to locate customers who are 'suckers' willing to pay more, it can route the high prices to these customers intentionally," explains Harris. "It's like creating profiling on certain customers, knowing according to people's consumer behavior on the credit card how they pay for things."

---

Tobi Harris

• Personal: 39 years old. Serves as director of the Markets Department at the Competition Authority.

• Professional: Filled a series of roles in the Authority over the years, including head of the finance team in the Markets Department and deputy director of the department.

• Something else: Immigrated to Israel alone from the USA at the age of 19.

---

The example is not intended to be a claim that Harel would act this way in reality, but an illustration of the type of concern that the Authority seeks to examine. "This is really a topic that is the future of merger deals and in general," says Harris. "If there is someone who has access to this bonanza and others do not, then he can push them out of competition and effectively take over the market."

This concern is currently at the center of attention of competition authorities around the world. A new OECD report describes how the use of detailed data allows algorithms to identify customers who may move to a competitor and offer them a lower price, and at the same time identify customers with a higher willingness to pay and charge them a higher price.

Case study

Another example that Harris brings is the examination of the merger between Sugat and StoreNext. On the face of it, these are companies from different worlds: Sugat is a food supplier while StoreNext provides information and data on the retail market. However, it is precisely their connection that illustrates how information may become a competitive advantage. "We saw that Sugat has a status in all kinds of markets and the concern is that if it has preferential access to StoreNext, then its ability to identify which supermarket started buying more from another supplier — that is a great asset in the hands of a leading supplier," explains Harris.

From the Authority's point of view, the mere existence of a database is not enough to raise a red flag and the test itself is more complex. Harris explains that first one needs to examine what information the company holds and how unique it is. After checking that, they check what can be done with it and at the final stage, what its various uses might do to competition. "It does not automatically mean that a company that holds information about customers raises a competitive concern. Everything is examined deeply, on its own merits," she clarifies.

Thus, for example, if similar information is found at a long line of companies, it is difficult to see it as a decisive advantage. In contrast, unique information that passes to a company with a strong status in the market may change the picture of the entire deal. "We ask in every merger the question of the merger's delta: how the world without the merger changes compared to the world with the merger," says Harris. "If it is data that many different companies have, then it is not such a big deal that everyone has it. If it is quality data that no one else has and it reaches someone with status — that is something else."

AI and non-competition

And if we thought the matter was complicated until now, the AI issue takes the matter a step further. The reason for this, from Harris's point of view, lies in one of the less discussed components of competition: uncertainty. "One of the things that creates competition is uncertainty," she says. "A company that considers a business move acts in uncertainty. In the end, it is a big part of the competitive game. I refrain from raising a price because I don't know if it will pay off for me or open a branch based on partial knowledge in the hope that it will work for me."

Data, according to her, can change this. "If there is someone who jumps over the lack of uncertainty with the help of data, it changes the rules of the game. Many times knowing the market data is a one-way game changer," emphasizes Harris.

A database, which in the past required teams of analysts and long research work, can become the basis for a system that makes decisions in real time, identifies patterns and reacts almost immediately to a change in the behavior of a customer or a competitor. "All AI is built on data," says Harris. "All the biggest companies in the end, whether they are a social network or a search engine — their data is like the oil, the gold or the engine of the new economy."

The algorithm question

Here begins the next problem that worries competition authorities in the world and it is no longer related only to mergers. According to the OECD report, one of the central concerns is a scenario in which several competitors use the same supplier of pricing software. In such a situation, the algorithm may become a kind of central link that receives information from several companies and recommends to each one how to price. Even when the final price is calculated separately for each company, using competitors' information for the purpose of training the algorithm may raise a concern.

This concern is especially relevant for markets where prices change quickly and most of them are visible on the internet. An algorithm can track competitors continuously, identify who lowered a price and react almost immediately. According to the OECD, such systems may also make it easier for manufacturers to identify retailers who deviate from a recommended price and react quickly, thereby making vertical price arrangements more effective.

At the far end of the discussion is a question that competition authorities are still struggling to answer: what happens if the algorithm itself learns that it is not worth competing? The OECD notes that some of the competition authorities in the G7 organization are examining a theoretical possibility of "tacit coordination" between learning pricing systems. Companies may program separately algorithms whose goal is to maximize profit, but the systems can, at least in theory, learn that the most profitable result is to avoid price wars.

However, the report emphasizes that this is at this stage a concern that has not yet been proven on a significant scale and that the number of enforcement cases in the world in the field of algorithmic pricing is still limited. The authorities themselves are currently investing more in learning the market, monitoring and building technological capabilities than in large-scale enforcement.

This gap also sharpens the Israeli challenge. A traditional cartel investigation may look for emails, messages or meetings where competitors agreed on a price. In a world of algorithms, the evidence may be found precisely in the code, in the way the model was trained or in the information that was fed into it.

"No need for a law amendment"

Despite the technological changes, Harris does not believe at this stage that new legislation is required in the competition laws themselves. "The Competition Law is written very broadly. It does not enter into specifics and it allows us to pour the content into it," she says. "We do not need an amendment to adapt ourselves. We need to study the business on its own merits."

According to her, the Authority is also involved in broader government discussions around AI and data, including in the financial field. "In general, there is a perception that information does not belong to the one who collected it but to the person the information concerns," notes Harris. From her point of view, the main challenge of the Competition Authority is not necessarily to catch more deals, but to know what to look for within the deals that already reach it.

"I think we are not missing out," says Harris. "Every merger between large companies works on turnovers and on the size of the deal. The problem is not in the net that catches the mergers, but that when the merger arrives — to know how to talk to the companies and understand also the things that are a bit behind the scenes."