What Google Search Isn’t Showing You

The search engine has made up so much of our online experience for so long that it can be hard to imagine something better.
Illustration of ballon with the google logo deflating
Illustration by Nicholas Konrad / The New Yorker

When I recently Googled “best toaster” on my phone, thinking about replacing the appliance in my apartment kitchen, the search immediately yielded a carrousel of images of products from various high-design brands: Balmuda, Hay, Smeg. (Guilty: I had definitely searched for the Japanese Balmuda’s steam-enabled toasters before.) Lower down on the results page were ads for online retailers such as Amazon and Wayfair, then another carrousel of “Popular Toasters” with user-review metrics, then a list of suggested queries under the heading “People also ask.” (“Is it worth buying an expensive toaster?” “You can’t gain much beyond the $100 models,” an answer pulled from CNET reads.) Swiping down further, I reached aggregated listicles clearly designed to exploit Google’s search algorithm and profit from affiliate marketing: toaster tips from Good Housekeeping, the “4 best toaster ovens of 2022” from Wirecutter. Further down still was a map of toasters that could be purchased in physical proximity to my apartment. I felt lost among the suggestions, awash in information and yet compelled by none of it.

This kind of cluttered onslaught of homogenous e-commerce options is what recently prompted Dmitri Brereton, a twenty-six-year-old engineer at a recruiting-software company in San Francisco, to publish a blog post titled “Google Search Is Dying.” When it comes to product reviews or recipes, Brereton argued, results from Google’s search engine “have gone to shit.” Rather than settling for the default, those who want to know what a “genuine real-life human being” thinks of a certain product have learned work-arounds, such as adding “Reddit” to their searches to bring up relevant threads on that platform. On Reddit’s “Buy It for Life” forum, for instance, they’ll find users showing off a Soviet-era toaster, a restored vintage Sunbeam, and other toasters to “grow old with,” as one put it. Brereton’s post–which ended “Google is dead. Long live Google + ‘site:reddit.com’ ”—became the No. 10 most upvoted link ever on the tech-industry discussion board Hacker News. No. 11 is a complaint about Google’s search results looking too similar to its ads, while No. 12 is a link to an alternative, indie search engine. Clearly, others share Brereton’s sense of search-engine discontentment.

Brereton told me recently that his frustration began in late 2020. “I was browsing the Internet one day, and I began to feel like something was just off,” he said. “A lot of the content doesn’t feel authentic—it doesn’t feel real.” He sounded bemused by the runaway popularity of his post, which was part of a personal research project on how information is organized online. Better information could be found on social media, discussion boards, and small-scale personal blogs, but Google Search was deprioritizing those platforms in favor of corporate Web sites, which could afford the money and effort it takes to optimize for Google’s search algorithm. “The authentic Web” seemed hidden, Brereton said. “The algorithms tell us what to read.”

Google Search accounts for around eighty-five per cent of the global search-engine market. It has made up so much of our online experience for so long that it can be hard to envision anything different. The Google Search page today looks largely the same as it did when it first launched, in 1998: blue links against an austere white background. From the beginning, the company’s founders, Sergey Brin and Larry Page, recognized the tension between useful search results and profitable ones. “The goals of the advertising business model do not always correspond to providing quality search to users,” they wrote as Stanford students, in a 1998 paper. Yet ads were introduced in 2000 and have proliferated ever since. Links to Web sites have fallen down search-result pages, replaced by Google’s “Quick Answers,” which borrow bits of text from sites so that users don’t even need to click. Decades of search-engine optimization have resulted in content that is formulated not to inform readers but to rank prominently on Google pages. That might be one reason that my toaster results felt so redundant: each site is attempting to solve the same algorithmic equation.

Gabriel Weinberg, the C.E.O. of the privacy-focussed search-engine company DuckDuckGo, cited three other sources of dissatisfaction with Google Search. The first is the company’s practice of tracking user behavior, which drives the kind of creepy, chasing-you-around-the-Internet advertising that Google profits from. The second is Google prioritizing its own services in search results, by, for instance, answering a travel query with Quick Answers pulled from Google Places instead of from a richer, more social source such as Tripadvisor. Lastly, Weinberg argued, users are simply tired of Google’s dominance over their experience of the Internet. Google is reportedly paying Apple upward of fifteen billion dollars a year to remain the default search engine on iPhones. On Google’s own Android phone, changing one’s preferred search engine requires a cumbersome settings adjustment, and pop-up messages along the way urge the user to switch back to Google. “Most people have never chosen their search engine,” Weinberg said.

DuckDuckGo, which doesn’t engage in any user tracking, has in the past year doubled both its estimated user count, from fifty million to a hundred million, and its search traffic, from 1.5 billion to three billion queries a month. It serves ads based only on users’ search terms rather than on their past behavioral patterns. According to Weinberg, each time you search on DuckDuckGo, “it’s as if you’re there for the first time.” In reality, though, choosing among existing search engines has only a modest impact on one’s results, because all of the major engines operate based on a template established by Google’s original web crawler and PageRank technologies. DuckDuckGo, for instance, largely leverages the search algorithm of Microsoft’s product Bing. A DuckDuckGo search for “best toaster” surfaces more or less the same results on my phone as Google, with ads, sale alerts, and affiliate-marketing listicles. (The Times recently reported that DuckDuckGo has become popular among conspiracy theorists who believe that Google’s results are censored, but DuckDuckGo disputes that it turns up any more controversial content than Google.)

Viktor Lofgren, a Swedish software developer and consultant who created his own indie search engine called Marginalia, told me, “One part of the sameyness is that recommendation and prediction algorithms often seem to work almost too well.” Marginalia, which Lofgren started working on a year ago, is a bare-bones Web site run entirely from a computer in his living room. The search engine’s stated mission is to “show you sites you perhaps weren’t aware of.” Its results, based on its own custom algorithm and data gathering, prioritize text-based Web sites that lack ads, mobile support, encryption, and other features that qualify as good S.E.O. “Google punishes sites that aren’t up to speed with modern Web technologies,” Lofgren said. “Legitimately old Web sites also deserve some attention.” A Marginalia search for “best toaster” brings up tech blogs from the nineties and vintage Internet jokes about technology companies of the day. (“If Apple made toasters . . . It would do everything Microsoft toaster does, but 5 years earlier.”) There are no images on the page, let alone carrousels or “Buy It Now” buttons. The Marginalia results would not help you choose a new appliance, but they are a fascinating glimpse into how much material the Internet contains, and how much never makes it to the top of current Google results.

Danny Sullivan, Google’s “public liaison for search,” told me that people using Google to find Reddit threads is actually evidence that search is working the way it should. Users on the whole have become passive, relying on Google to anticipate their desires. If they wanted, they could refine their queries, limiting results by, say, price point (“toaster $40..$100”) or by listing certain terms to exclude (“ ‘toaster’ NOT ‘oven’ ”). As machine-learning algorithms have grown more pervasive, we’ve lost some of the fluency with search that older Internet adopters may have learned in a high-school Boolean tutorial. “There’s a shift now where, if you don’t find what you’re looking for, you blame the search engine,” Sullivan said. At the same time, he admitted that many users have a desire for “more noncommercial information, more community-based information.”

If it wanted to, Google could adjust its search algorithm to prioritize Reddit or other social platforms by, say, pulling more of their content into its Quick Answers. Already, its algorithms are constantly changing. In 2020, for instance, I noticed that the results of my Google Image searches were often pictures from the mood-board site Pinterest, stripped of any identifying titles or source information. (I clearly wasn’t alone: a complaint about the phenomenon that I dashed off on Twitter got nearly a hundred thousand likes.) Pinterest had gamed the system, leveraging some quirk of the image-search algorithm to draw more visitors to its own site. Today, though, Pinterest doesn’t come up in my results with nearly the same frequency. When I asked Sullivan about the change, he cited the company’s recent work to “increase the diversity of domains” that they show.

However much certain users crave “authenticity” in their results, search engines are ultimately convenient, automated tools for finding answers. A fully non-automated search engine might resemble something like Are.na, which was founded in 2011 as a kind of hipster Pinterest popular among graphic designers and art directors. Are.na users can create their own thematic boards of images, video, and text: “Cool Photographers,” “Logos That Slap,” “Meaningful Garments.” Each board represents a person’s effort to curate a small batch of the world’s information by hand. “We talk about Are.na as a search index that is powered by humans instead of algorithmically,” Daniel Pianetti, its co-founder, said. There is no “best toaster” mood board, but one can find an annotated article about smart toasters, a stovetop D.I.Y. toasting device, and a beautiful mood board of kitchens. A list of guiding principles on the site’s rudimentary home page states, “Allow yourself to go down rabbit holes.”

For those who gravitate toward sites like Are.na, interesting results are perhaps preferable to useful ones. For plenty of others, a carrousel of Wayfair and Amazon toasters might be good enough. Sullivan noted that Brereton’s post, popular as it was, represents the feelings of only a small niche of users. “Hacker News does not represent the totality of everyone who searches,” he said. “Not all of them are going to be happy if all of our results are a hundred-per-cent Reddit.” For me, a superficial Google search is often useful only insofar as it surfaces subjects for further research, which is to say further Googling. I wasn’t going to buy a three-hundred-dollar Balmuda toaster just because it was given prominence on my search page. But once I dug around for reviews I admit that I was tempted. A writer at Wired determined that the “toaster makes [them] happy,” while a Reddit user offered this convincing testimony: “Before COVID, I would go to my next door neighbor's house just to use their BALMUDA.” It’s the sort of recommendation that no algorithm could give.