New Zealand markets open in 6 hours 26 minutes

Revealed: How Amazon uses third-party seller data to build a private label juggernaut

Krystal Hu

Amazon’s hybrid model of being a retailer and a marketplace may have served its customers well. But the dual role has rattled third-party sellers on its platform and drawn antitrust scrutiny in the U.S. and Europe.

One of Amazon’s businesses that regulators are looking into is the e-commerce giant's expanding private-label products, as third-party sellers — who are essentially business partners of Amazon's and generating over half of its sales — have long worried that Amazon uses their own sales data to compete against them. Two former employees at Amazon revealed to Yahoo Finance that the team used information from third-party sellers to develop Amazon private label products, including AmazonBasics, one of its most successful private labels. They also raised questions about Amazon’s lax control over employees’ access to marketplace data.

The former employees say the internal data are much more sophisticated than the information available in the public domain, or any third-party tools built for sellers. Employees can view aggregated data such as search interest and also pull reports from Amazon’s data warehouse on specific products to help them decide what Amazon should make, according to those employees, who asked for anonymity because they have signed non-disclosure agreements or fear retribution from Amazon.

By putting search inquiries in the data warehouse, the employees said they were able to analyze various datasets from the marketplaces, including the list of top-selling and trending products in certain categories, pricing points, return data and reviews.

One former employee, who worked on building AmazonBasics, said they would use the metrics to decide on an “inspiration product,” then order it from the seller and try to reverse engineer it. They worked with the sourcing team, who are usually based in China, to find a manufacturer.

Amazon Basics, a private label brand of Amazon providing low-cost, basic technology and other home products, has achieved major success. (Photo by Smith Collection/Gado/Getty Images)

In a statement to Yahoo Finance, an Amazon spokesperson said the company has policies in place and specifically forbids employees from the above behavior. Amazon wouldn't confirm if employees who work on developing its private label products can look up third-party sellers’ data. In June, Amazon's CEO Worldwide Consumer Jeff Wilke said the company doesn't allow anyone to have access to individual sellers’ data in order to build a private-label product.

“We train employees extensively on it, have technical controls in place, and audit compliance vigorously. Any employee in violation of this policy would face severe consequences,” the spokesperson said.

Wilke also said the data Amazon uses is “just the best-seller list” that is “fairly available to anybody.” “So the things that we end up making in private labels are the things that sell the most that are at the top of search results,” he said.

In a letter to the House Judiciary Committee in July, the company reiterated it prohibits the use of data related specifically to individual sellers in its private-label strategy. “Like other retailers, we use aggregated store data (e.g., total sales) and customer shopping behavior (e.g., search volume) to identify the categories and products with high customer demand over a given time period,” it wrote. Former employees said the use of data doesn’t target individual sellers—but they looked at specific products and categories.

‘Free-for-all’ access to data

The team at Amazon who works to develop private labels falls under the retail team, which is a separate department from Amazon Marketplace. Two teams operate individually, with different management and financial goals. But they share the same access to the data warehouse, which makes it possible for the retail team to use the data from marketplace sellers to develop private labels, according to former employees.

There aren’t enough restrictions on data access, former employees say. “Once you have data warehouse permissions, you can see just about anything in the company. Employees who have access to the data warehouse can pull individual orders, or sales for one particular seller, or all the different products from one particular vendor code or one particular seller ID,” a former employee said.

Another former manager at Amazon described it as a “free-for-all” situation once employees get access, as long as they have legitimate business reasons, which could be widely defined. This leaves room for employees to play around with the dataset.

“They know price points, they know trends, they know the value, they know the history,” the person said. “Third-party sellers are using tools that are estimating and guesstimating those numbers. The tools have gotten better and better, but none of them are nearly as good as Amazon's true data.”

Inside Amazon, some employees simply didn’t see how using third-party sellers’ data could be a problem, let alone create issues for the company. Instead, employees were focused on making data-driven decisions, something that Amazon encourages employees to do.

“I don't think people saw the seriousness of it,” a former employee said. “It was our data. We could look it up, right?”

Amazon says it regularly audits the systems and processes for compliance and trains employees on its data policy. Third-party sellers effectively give Amazon access to their data by signing an agreement with Amazon when they sign up to sell in the marketplace. Some feel the terms are one-sided in Amazon's favor. Whether clauses or provisions in these agreements, like handing over data, affect competition, has become a centerpiece of the EU antitrust investigation on Amazon.

Private labels aren’t always successful

Private label strategy is common among retailers. Traditional players from Walmart (WMT), Target (TGT), to wholesaler Costco (COST), all have developed their own private label brands. Amazon tends to draw an analogy to retail stores to show that what they are doing is not out of the norm. It also emphasizes that its private labels account for a tiny fraction of its retail sales.

Amazon private labels have expanded into various industries. (Data: Scrape Hero)

“Like other retailers, we seek to grow the private label selection in our store to satisfy that widespread customer expectation and demand,” Amazon wrote in a July letter to lawmakers. “Still, despite launching our private label business in 2005, Amazon’s private label products lag far behind those of our rivals; private label sales account for approximately 1% of our total sales, while retailers like Walmart, Costco, Kroger, Target, Trader Joes, and Walgreens have private label products that represent 18% to 85% of their sales.”

Critics, however, point to Amazon’s unparalleled scale and market position in the online marketplace. Amazon, known as the everything online store, carries about 120 million products, according to Scrape Hero data. In comparison, Walmart offers 75 million products online and most grocery stores keep fewer than 60,000 stock keeping units (SKUs) in-store. So Amazon may be able to gain more variable data into the retail business overall.

Amazon’s dominating position has also created more concerns for sellers who heavily rely on its marketplace. Amazon accounts for over one-third of the e-commerce sales in the U.S., followed by eBay (EBAY) and Walmart, which own single-digit market shares, according to eMarketer data.

“If you define Amazon as a monopoly in online retail, then you could argue that what they're doing, looking at the data on the platform, should not be done,” said Juozas Kaziukenas, founder of Marketplace Pulse. “So ultimately, it depends on your definition of does Amazon have monopoly power?”

Examples of Amazon private label brands. (Screenshot/Amazon)

Amazon has launched 144 private label brands and 638 Amazon exclusive “Our Brands,” some of which are namesake, while other brand names avoid affiliation with Amazon, according to TJI Research. Having access to data may give Amazon a jump start, but it doesn’t guarantee those products would be a hit. The most successful ones are in electronics, where Amazon owns 30%-40% of the top 100 products, Marketplace Pulse data shows. But in fashion, where Amazon has launched nearly 100 clothing brands, it hasn’t been able to achieve the same level of success.

This year, Marketplace Pulse research finds Amazon’s pace of launching new brands has slowed down, but it’s still introducing new products underneath existing brands like AmazonBasics. It has also made more efforts to attract outside brands to list exclusively on Amazon.

“Launching a brand and figuring out what consumers might want is one thing, and actually making that brand desirable is completely different,” said Kaziukenas.

Share your thoughts on Amazon with Krystal Hu via or follow her on Twitter.

Read more: