WASHINGTON (AP) — Appearing in the biggest antitrust trial in a quarter century, DuckDuckGo founder Gabriel Weinberg testified Thursday that it was hard for his small search engine company to compete with Google because the powerhouse has deals with phone companies and equipment manufacturers to make its product the default search option on so many devices.
“We hit an obstacle with Google’s contracts,’’ Weinberg said in U.S. District Court in Washington.
The U.S. Department of Justice argues that Google has smothered competition by paying companies such as Apple and Verizon to lock in its search engine as the default choice — the first one users see — on many laptops and smartphones. Google counters that it dominates the internet search market because its product is better than the competition.
Even when it holds the default spot on smartphones and other devices, Google argues, users can switch to rival search engines with a couple of clicks.
But Weinberg testified that getting users to switch from Google was complicated, requiring as many as 30 to 50 steps to change defaults on all their devices, whereas the process could be shortened to just one click on each device.
“The search defaults are the primary barrier,’’ he said. “It’s too many steps.’’
The MIT graduate started DuckDuckGo in his basement in Pennsylvania in 2008, plucking its name from a children’s game. After a couple years, the company began positioning itself as a search engine that respects people’s privacy by promising not to track what users search for or where they have been. Such tracking results can be used to create detailed user profiles and “creepy ads,’’ Weinberg said.
“People don’t like ads that follow them around,’’ he said. DuckDuckGo’s internal surveys, he said, show privacy is the biggest concern among users, beating their desire for the best search results.
DuckDuckGo still sells ads, but bases them on what people are asking its search engine in the moment, a technique known as “contextual advertising.” That focus on privacy helped the company attract more users after the Edward Snowden saga raised awareness about the pervasiveness of online surveillance. It gained even more customers after Facebook’s Cambridge Analytica scandal opened a window into how personal information extracted from digital services can be passed around to other data brokers.
DuckDuckGo is privately held, so doesn’t disclose its finances. But it has said that it’s been profitable for several years and brings in more than $100 million in annual revenue. That’s loose change for Google’s parent company, Alphabet, which generated $283 billion in revenue last year.
DuckDuckGo still handles only 2.5% of U.S. search queries, Weinberg testified Thursday.
Under questioning from Google, Weinberg said much of DuckDuckGo’s search capability originated with Microsoft’s Bing search engine and was not homegrown. The company had allowed Microsoft to track some DuckDuckGo users until a security researcher flagged the practice in 2022.
Weinberg apologized at the time and DuckDuckGo tightened its tracker blocking to include Microsoft.
In earlier testimony, Eric Lehman, a former Google software engineer, seemed to question one of the Justice Department’s key arguments: that Google’s dominance is entrenched because of the massive amount of data it collects from user clicks, which the company in turn leverages to improve future searches faster than competitors can.
But Lehman said machine learning has improved rapidly in recent years, to the point that computers can evaluate text on their own without needing to analyze data from user clicks.
In a 2018 email produced in court, Lehman wrote that Google rivals such as Microsoft, Amazon, Apple, China’s Baidu, Russia’s Yandex or even startups could use machine learning to improve internet searches and challenge Google’s lead in the industry.
“Huge amounts of user feedback can be largely replaced by unsupervised learning of raw text,’’ he wrote.
In court Thursday, Lehman said his best guess is that search engines will shift largely from relying on user data to relying on machine learning.
During the exchange, U.S. District Judge Amit Mehta drew a laugh by asking how internet searches would answer one of pop culture’s most pressing questions this week: whether superstar singer Taylor Swift is dating NFL tight end Travis Kelce.
