Yesterday I published an article on the evolution of Google Search. I wanted to write something about this topic because every few months some version of the same claim goes viral: Google Search is dead or dying. I spoke with a bunch of search-industry experts, folks at Google, and people who have labored in the content mines creating some of the keyword-riddled junk posts that tend to clog search results. I thought I’d use this newsletter to talk a bit more about search and the tensions that come with a maturing internet.

The evolution of Google mirrors the evolution of the internet in general. It follows, then, that some of our anxieties about Google Search may be tied to broader feelings and frustrations with the internet in general. For a specific type of internet user—one who can still remember analog and early digital life, but who was an early, eager, and savvy adopter of the technology—there is a deep frustration with the commercialization of the web, which appears to violate the radical, open spirit of the early promise of the internet. This feeling, or the stereotype of it, boils down to something like: We wanted an information utopia; we got intrusive spam, fakery, ads, walled gardens, disinformation, grifters, filter bubbles, and monopolies. Google Search, in some cases, is a great example of this failed promise. Are Google’s search results getting less helpful? Or is Google’s product evolving with an internet that is degrading, becoming filled with useless information?

As I argued in the piece, asking whether Google Search is dying is a fruitless question. Google Search is too massive and too opaque to adequately assess a claim like that. What was most interesting in the reporting of this piece was the way that a product like Search inspires all kinds of contradictions.

One reader had a great observation. He noticed that a Google spokesperson told me that, “on average over the past four years, 80 percent of searches on Google haven’t had any ads at the top of search results.” The reader rightly pointed out that, in their personal experience, terms that are difficult to monetize are less likely to be swamped with ads. But very commercial search terms become minefields for advertisements.

When it comes to Google Search, it’s difficult to make sweeping claims, in part because everyone’s experience using the platform is different, but also because every search term is different.

My inbox is currently awash in testimonies about how Search doesn’t work as it used to or is less helpful than it used to be. And yet the number of searches per day continues to increase (from 5.4 billion a day in 2020 to, according to one estimate, 8.5 billion in 2022). Is that because more people have come online in those two years? Or because search is simply more deeply embedded in our lives? Is it because Google Search is getting better and people are using it more? Or is it the result of Search delivering less useful results and forcing us to do follow-up queries? It is almost impossible to know. Google, rather predictably, argues that, over the past seven years, it has “decreased the number of irrelevant results by over 50%.”

That figure is derived from “internal metrics based on quality rater data,” so knowing how to evaluate the claim is difficult. The 50 percent reduction in irrelevant results is big enough that all of us ought to have felt this shift. If the stat is genuine, why aren’t we all talking about how great Google Search is?

This question came up in my chat with Danny Sullivan, Google’s search liaison. Sullivan is a former journalist who covered the search industry for years. He believes in the original promise of Google Search. and while he readily acknowledged the problem for Google if people feel like the product is getting worse, he argued that it definitely is not. From my story:

Sullivan believes that some of the recent frustrations with Google Search actually reflect just how good it’s become. “We search for things today we didn’t imagine we could search for 15 years ago, and we believe we’ll find exactly what we want,” he said. “Our expectations have continued to grow. So we demand more of the tool.” It’s an interesting, albeit convenient, response.

Sullivan’s explanation will probably make you roll your eyes. Ah yes, the problem is that we are too good! But it might contain some grain of truth as well. I put Sullivan’s answer to a friend of mine who also writes about technology, and they compared Google Search to a tool like autocorrect, which has evolved to be so much better over the years. Because we have become dependent on autocorrect, we are more annoyed when it fails. We primarily notice those failures and, as a result, develop this theory that the technology isn’t as good as it used to be.

Sullivan said that people frequently reach out to say that Google’s quote operators (putting a phrase in quotes in the search bar in order to return an exact word match) don’t work as well as they used to. He takes these concerns seriously and always tries to investigate himself … by searching. In one example, a person was looking for a specific piece of content about disassembling a blender. Sullivan told me that, after exhaustive internal searching, it was clear that the particular post the user was looking for just didn’t exist—at least not the way they thought. It’s possible, he argued, that they were simply misremembering a key phrase or term.. “In the earlier days of search, you might have said, I looked and couldn’t find it and so it must not be out there,” Sullivan said. “But now, people say, ‘I can’t find it and why didn’t you show it to me?’” Sullivan also acknowledged that this was not an excuse and that Google needs to get better at explaining why results do or don’t surface.


To read the rest, subscribe to The Atlantic.

Already a subscriber? Sign in