Well, it happened again: Google went and made new rules, and gave us a startling reminder of how deeply we had dug in on the old ones. On some level, we SEOs go to bed every night believing that the game will be the same tomorrow as it was today, even as we constantly attend and deliver talks about how much it has changed over the years. I don’t know where that disconnect comes from, but it’s probably just an impulse to protect our sanity; if we really took time at the end of each day to reflect on how many times the very definition of our job and our industry has changed since we first got started, who knows how many more tomorrows we could bring ourselves to face.
Featured snippets first arrived in 2014 and sent shockwaves through the content strategy world immediately. The fact that Google had created a special class of search results page that offered an organic position superior to #1, and that access to that position could be (in theory, anyway) attained on content alone, shook up a lot of our old talking points.
For one thing, long-tail keywords and conversational queries seized our attention like they never had before, sparking a wave of support for pure Q&A content as well as niche blogging. For another, our attitudes toward subheadings warmed considerably; suddenly, every basic, informational blog post contained a theoretical infinity of opportunities for featured snippet appearances. Just dig up a common question on the subject, stick it in an H2, and supply a punchy answer in a short paragraph just below.
How did featured snippets impact the overall importance of relevance signals to SEO?
Great question, H2! I would argue that relevance signals (that is, on-page SEO factors like “presence of keywords in page title”, or “proximity of related keywords to one another in page copy”) took on added importance for SEO in general when featured snippets were introduced, and that that importance only grew further as they became more common. Suddenly, you were less likely than ever to be stuck for eternity in a #4–10 ranking position behind a few authority powerhouses. As long as you were somewhere on Page One and the query triggered a snippet, you could theoretically leapfrog everyone else on relevance credentials alone.* This is why Dr. Pete at Moz called it “Position Zero”, and why that coinage stuck; being awarded a featured snippet didn’t mean you were seizing the #1 ranking with your relevance signals, but rather than you were climbing to a new place above it, a place that authority metrics wouldn’t help you reach.
[*This was largely confirmed by the big-data studies in the field (see Point #5 in this one from Ahrefs), and anecdotal evidence of my own backs up the idea that featured snippets were traditionally awarded on a second, from-scratch relevance signals computation that kicked in after the core algorithm determined the overall organic rankings. I saw some junk sites run away with featured snippets in my time, purely because whatever Google scraped from them was, in spite of everything, a good answer to the question as worded. I just now remembered a particularly egregious one and went back to the client document to read the note I wrote to myself: “Google Chrome gives the ‘Your connection is not safe’ warning upon attempts to visit this page, but Google Search gives it a featured snippet. Lol”]
Lastly: when the voice search hype died down and that discourse finally started cooling and settling around 2018, it was largely because everyone realized how strongly voice search answers correlated with featured snippets. Essentially, it turned out that voice search SEO was basically completely coterminous with a good featured snippet strategy (which makes perfect sense, given that the single most significant fact of voice search is that it returns one answer). Show up in a snippet and the next thing you know, a magical voice in someone’s living room is speaking your words as though they were God’s truth. (That points to the other reason why the voice search hype died down, though: the realization that that feeling — the feeling of satisfaction at having supplied the Great Official Answer — was to be the winner’s only reward. The people who own the ears that hear your holy message in that living room stand little chance of ever knowing it came from you, and less chance still of turning into your customers.)
My real point is that featured snippets disrupted content/relevance SEO when they arrived, but then stabilized into a new norm; we got used to them behaving a certain way and to optimizing for them in a certain way, and that’s because while the particulars of their appearance had changed a bit in the five-ish years that they’d been with us — to be brief, they got richer and started showing up more often — the basic rules surrounding them hadn’t really changed at all. Until January 22.
On that day, Danny Sullivan revealed in the course of a super-cazh Twitter convo with Dr. Pete (one that Dr. Pete started, it should be pointed out, after discovering something in his own data) that Google Search was going to stop “duplicating” featured snippet sources on Page One of the organic listings:
If a web page listing is elevated into the featured snippet position, we no longer repeat the listing in the search results. This declutters the results & helps users locate relevant information more easily. Featured snippets count as one of the ten web page listings we show.— Danny Sullivan (@dannysullivan) January 22, 2020
In other words, from now on, the appearance of a given URL in a featured snippet is going to constitute that URL’s only Page One appearance.
This is one of those funny cases where I support the idea as a net positive for the user, but I’m not crazy about what it implies for site owners and SEOs. It’s certainly true that “duplication” (though I think “redundancy” the clearer word) was an element of how it was before, and a SERP that doesn’t repeat any results is by definition more diverse — richer overall — than a SERP in which one of the results appears twice (or at the very least, it’s shorter; Dr. Pete confirms that the net effect of the change is a reduction in length on these kinds of SERPs from 11 results to 10). And numerous SEOs have taken to Twitter to bemoan the loss of the fun feeling we used to get from being able to show a client that they had both “Position Zero” and the traditional #1 ranking, stacked one right on top of the other. But I think there’s a sadder, subtler loss for SEOs in this change: the loss of the reassurance in the redundancy that we had before. To see the featured snippet result appear again among the organic listings reinforced the idea that the organic listings were the organic listings — ten of them, ranked by the core algo, same it ever was — but that the featured snippet was awarded by a different “mind”, making a separate and secondary judgment (one that was based more strongly on relevance signals). That’s the whole premise that made featured snippets seem like the amazing backdoor to SERP domination that they were: they would allow you, on certain kinds of queries, to jump over the high-authority sites who dominated you in the listings, by the power of relevance signals alone. Seeing yourself at #6, but also at #0, sent the clear message that you were not “ranked #1” for this keyword according to the complex arithmetic of the core algo, but that you were determined by a different calculus to supply the best simple answer to the question on the web. We would go on to argue, with backing from our private datasets, about the effect of appearing in a featured snippet on CTR (it improves it! it damages it!), but the idea that page quality could be your ticket to visual domination of a SERP — to designation as Google’s Favorite Answer — even if you were overmatched on authority by big corporate sites, had the ring of justice to it. Score one for the little site.
With this redundancy now gone, Position 0 is no longer Position 0; it’s just a different-looking Position 1. So what reason do we have any longer to believe that a #1 ranking with featured snippet is calculated differently from a #1 ranking without featured snippet? You could point to the later Sullivan Twitter reply threads revealing that the URLs in featured snippets are still showing up in the organic rankings… on Page Two. But his comment on that basically amounted to “Lol Page Two”, and that sums up how I feel about it; even if the re-appearance of the featured snippet URL on Page Two proves that it’s still the same two-algorithm process determining which page gets the snippet*, the harder fact is that Page Two is a wasteland and to appear there is as good as not to exist.
Sullivan explicitly said that the priority was to clean up Page One, and as everyone knows, to state that “Page One is the priority” for both Google and for SEOs would be a stunning understatement. I am not heartened in the slightest by the idea that a URL basking in the featured snippet spotlight also appears at, say, #17. This means either that Google has expanded snippet eligibility, and now scans the whole infinitude of its organic listings on every snippet-triggering query to determine which one deserves the spot (unlikely), or that they’ve rejiggered the layout of featured snippet SERPs to arbitrarily demote the second appearance of whichever URL earns the snippet to “Somewhere on Page Two”, for reasons that they refuse to make clear. Either way, here’s the takeaway for SEOs: the calculus that awards featured snippets is now inextricably knotted up with the core algo, making it as impenetrable as the core algo itself.
[*It doesn’t prove that the process is the same; the one thing we knew to be true about the previous process was that you had to appear on Page One of organic rankings for the keyword to get the snippet, and that isn’t true anymore, so it isn’t the same process.]
So, that covers the most frustrating aspect of the change philosophically. Now we come to its most frustrating aspect professionally, and there are two, a big one and a little one. Here’s the little one: we who do this for a living now need to attach another asterisk — our biggest one yet, at that — whenever we speak of a “#1” organic ranking. It’s already been a bit of a semantic maze to talk about #1 rankings in the post-ten-blue-links era, when there’s been no predicting how far down the page a #1 organic ranking might be pushed by top-of-page ads and newfangled SERP features, including featured snippets. But now, not only is there no predicting how far down a page the #1 organic result will appear, but there are two totally different ways in which it can appear, one of which is way, way more desirable than the other. (And indeed there is still a spectrum of appearances found among featured snippets; a featured snippet with a big eye-catching image inside is that much more powerful than with only text.)
So from now on, whenever we’re speaking to a client about a #1 ranking on a keyword — whether it’s their ranking or a competitor’s — we’re going to have clarify: “#1: featured snippet,” vs. “#1: no featured snippet”. At the very least, we’re going to need a shorthand for this. “OneFS” vs. “OneNoFS”? Boy does that sound awful. I’m glad I’m not an SEO software developer; can you imagine trying to communicate this new reality in numbers and cute little icons?
And here’s the big one: we will not know how to optimize for featured snippets anymore, because we will not know for certain which of our clients’ pages are eligible. UpBuild’s Featured Snippet Strategy deliverable has historically been oriented around the old norm which stated that a page was eligible for a featured snippet if a) the query triggered one, and b) the page ranked on Page One for that query. Item b has just flown out the window. The way I see it, that leaves us three options:
- we can expand the size of the document massively, to account for every page our client has published about every snippet-worthy topic, regardless of where the page currently ranks in the organic listings;
- we can carry on under the (now completely unfounded) assumption that if our client’s page is currently ranked on Page One, it is more likely to win the snippet than if it were ranked below (and that if it were to win the snippet, its new #1 ranking would replace its old #2–10 ranking);
- we can throw the deliverable in the trash.
This is how it so often goes in SEO: not only are you forever adapting to Google’s rule changes, but — and this is the part I didn’t want to say at the beginning — the effect of most of those changes is disempowering. Either they take away a control you used to have (remember when you could demote sitelinks?) or they decide to start withholding information that they used to share (I don’t even need to mention not provided — oops, I did). Well, so it is again; the less we know about how Google awards featured snippets, the less we can do to earn them.
So, given that we will always have clients asking, “how can we appear in that box?”, which path do we choose? Well, path #2 is the most obviously promising, so that’s the one it will be, but we will be forced to proceed with a new halo of uncertainty around the whole premise. It was nice when we could answer that question with, “we have a detailed strategy package with a proven track record that we will gladly execute for you”, and now that the rules have changed, that “proven track record” piece will take time to build back up. We (and the rest of the SEO community) have a whole new round of testing ahead of us to figure out what’s possible and what’s worth doing. But just as Google never sits still, neither will we.