NOTE: This post is mostly theoretical, but I hope potentially helpful and worthy of discussion.
Over the last few years, and particularly since the advent of Hummingbird, I’ve noticed Google becoming more nuanced about the content it ranks, even for queries where they don’t have lots of data on what users click, what they engage with, what they ignore, and their behavioral habits around a search (related searches, usage patterns before/after the query, etc.).
My theory is that this new intelligence presents a dramatic opportunity for marketers and content creators who can identify the patterns and spot queries where critical questions may lie unanswered.
Historically, much of what we’d see from Google’s rankings could be explained through a few big factors:
- Links
- Domain Authority
- Keyword Matching
- Relevance
- Freshness (and, certainly during Google’s partnership with Twitter, social signals)
We know Google’s become more complex, but even from 2010-2012, I’d say the vast majority of searches’ rank ordering could be explained with elements contained in these broad categories.
Today, I’m observing a lot of rankings that seem to connect with brand signals, user/usage data, and a far more nuanced consideration of links, authority, and relevance, but perhaps most uniquely, and especially in queries that have information-gathering intent, there seems to be a set of ranking signals related to what I’ll call “relevance to alternative searcher intents.”
I’ll try to illustrate this with an example. Here’s a query for ”
space pen” in Google US (non-personalized with geo-biasing removed):
There’s three potential popular “intents” that searchers have around this query.
- Those seeking Fisher’s branded Space Pen
- Those seeking to learn about the oft-repeated myth around the supposedly costly development of the Space Pen by NASA when Russian cosmonauts used pencils
- Those seeking the Spacepen framework for Coffeescript
And Google’s done a nice job recognizing those unique intents and populating the SERPs appropriately with results to answer all three. Historically I’d have called this “QDD” or “Query Deserves Diversity” (something I
first wrote about way back in 2008).
But actually, I think we’ve seen an evolution from the raw “diversity” inputs (which, in my opinion, mostly revolved around combinations of click behavior in the SERPs and search modification behavior, i.e. people searching for “space pen” then refining to search for “space pen coffeescript”) to a model that has more sophistication.
That more sophisticated model might be better illustrated with this query for ”
most flavorful steak” (also Google US, non-personalized, non-geo-biased):
There are multiple intents around this query, but they’re far more subtle than those for “Space Pen.” Searchers are likely seeking things like a description of the various types of cuts, information about what makes a steak taste better, perhaps some interesting types of steaks they haven’t heard of previously or why certain cuts are more expensive than others.
What’s remarkable is how Google has made shifts in queries like this in the last couple years. If I performed this query in 2012 (which I’m fairly sure I did, but sadly didn’t screenshot), I would have seen a lot more keyword-matching and a much more singular focus on articles that specifically mentioned “flavorful” (or fairly direct synonyms thereof) in the title and headline. Actually, it would look a lot more like
Bing’s results (no offense to them; these results are actually quite good, too, just far more keyword match-focused):
Today, from Google, I’m getting a broader interpretation of the true intent(s) behind the use of the adjective, “flavorful.”
There’s results that touch on expensive cuts of steak, of types of beef itself (like Wagyu & Kobe), on what makes a steak more flavorful, and there’s a site showing up (Niman Ranch) that seems totally out of place when you look at the link numbers, but makes a lot of sense as a highly co-cited brand name. For reference, here’s a
basic keyword difficulty report for the phrase:
My opinion (and this is pure, unvarnished, speculation) is that Google’s using inputs like:
- Relationships between words, phrases, concepts, and entities to get closer to an understanding of language and an evaluation of the content quality itself
- Patterns detected in how authoritative pieces write about/mention the keywords
- User and usage data signals that look at multiple sessions, multiple queries, and identify patterns of searcher satisfaction (possibly using machine learning)
- Topic modeling that tries to identify terms and phrases that are associated with diversity of opinion and topical focus so there’s an element of finding not just useful information, but potentially new and interesting information, too
I don’t believe these are overwhelming signals today. Links are still very powerful. Domain authority is still clearly influential. But for a lot of what I’m seeing in the end of the chunky middle and into the long tail of the keyword demand curve, I think there’s an opportunity for marketers to perform some content gap analysis and win rankings without needing the quantities of links & authority otherwise required.
Here’s my strawman concept for starting out with some Content-Gap SEO (and hopefully y’all can rip into and improve upon it in the comments):
Step One: Identify the keywords you’re targeting that fit in the backside of the chunky middle and long tail.
Step Two: Prioritize your list based on the terms/phrases you believe will be most valuable (and remember that doesn’t always mean highest search volume).
Step Three: Starting from the top, write down 4-6 types of intent and/or pieces of unique information that you believe searchers might have/want when performing each query.
Step Four: Perform the query in Google, and look through the top 10. Do you see results that answer all of the intent/info types you wrote down? Write down how many are missing (including 0 if everything’s already fulfilled).
Step Five: Use your number as a potential prioritizer for the creation of new content or the modification/addition of content to existing pages. Then watch and see if Google feels the same way and begins rewarding you for this.
While this process is speculative and my theories are, too, I will say that I have talked to and emailed with a lot of folks in the SEO field of late who’ve talked over and over about the surprise they’ve had from purely content-based rankings and rankings improvements. I might be wrong about a lot of the details, but I’d be willing to bet that there’s something new going on in how Google analyzes and rewards pages that provide the right kind of content.
For marketers who can identify the patterns, find the content gaps, and fulfill them, I believe there’s opportunity to rank without having to pound nearly the same levels of external links at your pages.
Looking forward to the discussion!
