We Need to Talk About Google

Welcome new subscribers! If you like what you read, please consider forwarding this newsletter to someone else you think would enjoy it. And if someone forwarded this newsletter to you please consider subscribing!

I noticed the decline in Google search results the way you fall asleep; slowly, and then all at once.¹ I would demonstrate my tried-and-true google-fu techniques to students to show them what a difference they could make in their search results and… they didn’t always make that much of a difference. It seemed like the same kinds of sites appeared at the top of the search no matter what we were looking for. When searching on my own I found myself wading through a bunch of sites that were pretending to answer questions in order to get clicks. Predictably (and frustratingly) trying to search for stories about the declining quality of Google searches is not that easy. But within the last several months I started seeing more articles from tech reporters who were seeing the same things I was.

Before we go any further: Yes, I am aware there are search engines besides Google. But Google accounts for between 80% and 90% of searches and I don’t see that changing significantly anytime soon. And the kind of SEO behavior Google has incentivized has an impact regardless of the search engine we use.

How much can we blame SEO?

When talking with students about what influences the ranking of search results there were usually at least a couple of students who would ask if websites could pay to be at the top of Google search results. My answer used to be “no,” or to point out how sponsored posts are labeled. Now it’s usually “yes, but not in the way you think.” 

Search engine optimization has a long history, but if you’ve ever scrolled endlessly through a recipe blog trying to find the actual recipe you have seen what it’s done to the internet. I suspect many recipe blog writers have as much interest in writing about their relationship with beets as you do in reading it. But adding that story (and all of its associated search terms) can move your recipe up in the search results: the story isn’t for you, it’s for discoverability. 

Of course, Google doesn’t come out and say what will help your site show up near the top of search results; this is where the SEO industry comes in. The goal of SEO is, as Amanda Chicago Lewis of The Verge puts it, “to tell the algorithm whatever it needs to hear for a site to appear as high up as possible in search results, leveraging Google’s supposed objectivity to lure people in and then, usually, show them some kind of advertising. Voilà: a business model!” The primary audience for a website is no longer a person looking for information. The algorithm is the audience. 

This shift in the target audience has also made a lot of the internet look pretty homogenous. Mia Sato (also at The Verge -they’ve been doing some great reporting about Google) has an enlightening article about how webpage designers change their websites in order to appear higher in search rankings. I highly recommend reading the article; Richard Parry’s animations help make clear how Google’s incentive structure for discoverability has changed websites over time. Google doesn’t disclose what would move a website up in the search results (lest people try and game the system), so a lot of it is left to guesswork. This means if you want to show up in search results (and have people read your content) you need to be writing for the algorithm as much as you’re writing for your audience. It doesn’t matter how good your content is if your website is buried beneath dozens of other search results, so the needs of the algorithm end up outweighing the information needs of people. It is, to put it mildly, not great. 

Are search results getting worse?

Google has gotten better at natural language processing over the years, which can give people the impression that Google actually understands the questions we ask. For some questions that kind of quasi-understanding is fine, but a lot of times we end up in an “uncanny valley” of results: the results aren’t wrong, but they’re also not quite right, but they are close enough to being what you’re looking for that it’s hard to put your finger on what’s wrong. 

In the early days, Google’s goal was to show the most relevant results (based on how many other pages linked to it, amongst other signals of relevancy and quality) to your search query, and you would decide what to click, on what to read, and what information was most helpful to you. Now, likely in part because of the declining quality of a lot of search results, Google is often trying to just give you the answer to a question. These often show up as a “featured snippet” at the top of a results page² or in a box to the right of the search results. The problem is that Google doesn’t actually understand the questions we’re asking. 

If you’ve ever worked at a public service desk of any kind you know that the first question someone asks when they approach is rarely their actual question; it’s often the question that they think they can get an answer to. You usually have to ask follow-up and clarifying questions to find out what they’re looking for. Within librarianship this is referred to as the art of the reference interview, and it really can be an art. When a student asks me where the history books are they’re not generally looking to browse – they want to find something on a particular topic, or a specific book. When someone asks me if the printer is working, what they’re often asking is for help printing something. When a student asks for help finding sources on a particular topic I know that there are likely narrower parameters to the search that we need to consider. Google, however, takes your question at face value and then attempts to provide an answer. Google is trying to do our thinking for us, and it’s not particularly good at it. 

On February 14th (Frederick Douglass’s birthday) someone on Bluesky shared a pretty stark example of how trying to do our thinking for us can lead to questionable results. I don’t know what the original search was (though I was able to replicate the “People also ask” results on my own by searching [Frederick Douglass]) but there’s really no search in which “How many slaves did Frederick Douglass own?” should be included as a possible follow-up. You can also see how Google’s attempt to provide answers, rather than access to information, can go seriously awry; the snippet from that website is not referring to Douglass, but it’s been stripped of context (specifically, who ‘he’ is referring to). This has since been fixed, thankfully, but the conditions which made it possible in the first place persist. And while this was an obvious and egregious error, there are likely more subtle errors that go undetected for much longer. Just last week The Washington Post reported on fake contact information for airlines that shows up in search results; Google says they’re addressing the issue – which is also what they said the last time these stories were in the news. 

Can Google get better?

Google changes how its search algorithm functions and how it displays information regularly.³ They recently added a Perspectives filter to results, likely in response to people adding [Reddit] to their searches in order to find answers written by actual humans. The results under Perspectives are not great, in my experience, but I suspect they will get better over time. And then they will get worse again as folks figure out how to game the results. There is a “dog who caught the car” feeling about it, with Google and website designers/SEO professionals taking turns being the dog and car. Searchers are either holding the leash or driving the car; we’re caught up in all of this, but we don’t have much agency. 

The introduction of generative AI to search will have a significant impact on all of this, in ways I do not feel qualified to predict but am not particularly optimistic about. Google’s vice president and general manager of search, Liz Reid, was recently quoted in Wired saying that Google will be thinking about “which use cases are helpful” and “have the right balance of latency, quality, and factuality.” I don’t know what it means to “balance” quality and factuality, but it does not inspire confidence. 

So… what should we do?

Even though Google insists that its trademark only be used as an adjective, “to google” has become synonymous with search for lots of people. There are lots of alternatives to Google out there, which works on an individual level, but it doesn’t change how SEO has made webpages worse or prevent Google from showing inaccurate answers to other searchers. 

Search algorithms are comprised of editorial decisions made by their designers. Those decisions can be helpful! I like that Google shows me results for businesses that are in the same geographic area as I am. I like that Google often suggests I add [gluten free] to my search when I’m trying to figure out if some random grain is safe for me to eat. I don’t like that it’s not always clear how and why my search results might be different from someone else’s. I don’t like that personalization is an on/off switch, and not more granular. As much as we shouldn’t want Google trying to both figure out what it thinks our question is and then providing an answer, we do want curated information. But we need that curation to be more transparent. I know that may make it easier for SEO professionals to manipulate search results, but I wouldn’t say that the “throw spaghetti at the wall and see what sticks” approach they’re currently using is serving anyone particularly well. 

I don’t have a simple answer, but I think the first step is to shift how we talk about Google/”the algorithm” as being all-knowing. Just because Google wants to do our thinking for us it doesn’t mean we should let it. 

Thanks for reading! Want to work with me to help your community develop the skills they need to be savvy consumers and creators of information? Get in touch!

^{1 - With apologies to John Green.}

^{2- If you’ve ever copied a URL that looks absurdly long (or seen one in a student Works Cited page), this is why. The URL isn’t pointing to the page, it’s pointing to the “answer” on the page.}

^{3- Google recently experimented with removing the News filter for some users. It’s back now, but that change highlights how little control we have over how we sort our own search results.}