Google is a machine. It’s an incredibly sophisticated machine, with some of the most advanced artificial intelligence and learning algorithms used by a public audience available, but it’s a machine nonetheless. When we talk about the advanced semantic understanding of Hummingbird, Google’s goal of understanding user intent, and the iterative learning and improving processes of an update like RankBrain, it’s easy to think of Google as almost human, judging sites qualitatively the way a college professor might grade a paper, but at the end of the day, it’s still using analytical structures to rank sites for various queries.
According to recent evidence reported by Aaron Friedman on the Moz blog, Google may actually be using pre-constructed templates, or models for various industries, to determine where and how your company ranks for branded searches.
Over the years, Google’s algorithm has evolved from being very mechanical (ranking sites so predictably that spammers could easily manipulate the system) to being more qualitative (subverting the attempts of rank manipulators). For the noble, modern-day optimizer, SEO is a kind of hiring opportunity or presidential election. If you put the work in, improve yourself to the fullest, and adhere to all the basic expectations, you should be considered the best candidate for the job. It would seem unfair that other candidates would be selected over you because they had a certain height, or stature, or socioeconomic background.
Google using templates seems like an unfair breach of the search world we’ve been conditioned to anticipate (and let’s face it, we’re spoiled with this system anyway). Rather than purely judging sites based on quality or relevance, Google occasionally calls in pre-programmed patterns to map out the results. The questions you must be asking at this point are why? And how can we avoid it?
Let’s start with the fundamental problem that leads to templates seeming “wrong” or “unfair.” It all comes down to user intent. If a user searches for something like “emergency pet care near me,” it’s easy for Google to figure out that this user needs an animal hospital nearby, presumably for an injured or sick pet. Google doesn’t need a template to fall back on here, since the intent is decipherable and the relevance of possible pages can be easily evaluated.
Instead, let’s look at a more ambiguous, and much more common, type of query, like “Starbucks.” What is the user intent here? Does this user want to find a nearby Starbucks? Invest in Starbucks? Uncover Starbucks’s corporate history? Read news about the Starbucks CEO? There are too many potential meanings here to guess, so Google must have a safety net to deal with such queries—templates.
According to Friedman’s data, there are some clear patterns established for ambiguous branded searches like these—and they seem to be segregated along industry lines. For example, among hedge funds, the vast majority—74 percent—have their company homepage listed as a top result, with 72 percent getting Knowledge Graph entries, and most with Wikipedia entries seeing their wiki page ranked around the 4.5 mark.
On the other hand, among pharmaceutical companies, ambiguous branded searches return page one results that are only 20 percent corporate, compared to 37 percent for telecommunications companies. Engineering companies have far fewer media results than comparable industries, and food/drug stores rarely return stock quotes.
All of this seems like random, peculiar bits of information, but it’s important to note the underlying commonality here: Google uses specific ranking templates to help guide its allocation of results for intent-ambiguous queries.
You may be asking yourself, does this mean I don’t have to worry about optimizing my home page, since it’s going to rank no matter what? The short answer is no. Remember, these results are aggregated from 100 or more different companies in each industry; they aren’t a guarantee for every company involved.
Google must always strike a balance between the relevance of a page (how appropriate it is for a user’s intent) and its authority (how valuable a source it is, in general). These templates are a way of discovering relevance when no other contextual clues are given. The authority of a page must still be taken into consideration, which means if your homepage isn’t optimized, or if you don’t have a homepage at all, you could end up with no corporate page one rankings at all.
When it comes to optimizing pages for specific keyword phrases, long-tail phrases, and specific user intents, nothing should change. This is only about ambiguous branded queries and how Google tends to map them differently for different industries. Knowing what you now know, you can spin this to your advantage by optimizing the company presence you know to be most valuable.
For example, if your industry tends to have Wikipedia pages for companies ranking higher than social media content (like telecommunications companies or food/drug stores), make sure your Wikipedia page stays accurate and up-to-date. If media tends to be a popular result (as with pharmaceutical companies), work on publishing more images and videos. Think of it as a kind of reverse-optimization: instead of changing something so that it ranks higher, you’re changing something that already ranks high to give your brand a better reputation. Learn your industry well, and give your users the best information you can using the templates Google has already created.