Unfortunately for digital marketers, there is no yellow brick road that leads directly to good search rankings. However, at SMX Advanced recently, Rand Fishkin, the so-called Wizard of Moz, looked into his crystal ball to provide some insight into what may matter to search engines and impact rankings in the future.
Smart marketers who want to stay ahead of search algorithms and be found by consumers at critical moments know they need to watch where Google is going.
“It’s very possible these signals are purely speculative, but I think there is interesting evidence and/or patent applications from search engines suggesting it might happen now or in future,” Fishkin said.
Here’s a look at Fishkin’s take on some more speculative ranking signals, as well as his insight into what Google and Bing might use to rank results going forward.
1. Usage Data Of Pages & Sites
Fishkin uses the example of a query for “cloud computing” and points to Google Chrome and Android device usage as perhaps elements that are already informing Google of what consumers are searching for. In this particular example, a site with fewer visits per day ranks higher because Google is sending searchers to the page with greater visitor loyalty and engagement, he said.
And in another search for “silk flowers”, he notes this type of ranking input could be behind the strong performance of popular brand sites like Michaels and Amazon in queries where classic SEO elements are lacking, like poor keyword targeting and relevance, as well as few links, because the sites probably have stronger traffic and engagement than their competition.
2. Accuracy vs. Popularity Of Information
Fishkin pointed to a query for “do vaccines cause autism” and noted the website www.howdovaccinescauseautism.com ranks #2, which he said was initially troubling until he clicked on the link and discovered the site says, “They fucking don’t.”
“It makes you proud of the Internet,” Fishkin said. “By looking at multiple data sets, you realize an algorithm could determine the consistency of accuracy shown by a given website and potentially boost rankings based on that.”
3. Query Structure As An Anchor-Text-Like Signal
Noting this particular signal is “highly speculative,” Fishkin said how we construct search queries may be a factor that determines rankings in the future.
In other words, many searchers are using query structures in a particular fashion that could connect brands and modifiers to keywords, which, he said, opens an opportunity for marketers.
For example, a search for “sunglass” could bring up suggested results for the brand “Sunglass Hut.”
What’s more, popular searches around a brand could indicate associations that manifest in rankings input, he added.
4. Brands As Entities, Entities As Answers
Brands are increasingly becoming entities in Google’s Knowledge Graph, Fishkin said. And in many competitive SERPs, there seems to be a correlation between brand dropdowns and higher rankings.
“I see folks saying brands aren’t something Google talks about, but entities are,” Fishkin said. “Brands are becoming what Google classifies as entities.”
In other words, Google can see people also search for not just a given query, but a specific entity, which is “at least subtly showing some brand bias,” Fishkin said.
He also noted there is a strong correlation between brand dropdowns and higher rankings.
Using the example of a search for “Seattle real estate market,” he points out that real estate website Zillow is so tightly connected to the keywords and associated with the query that it appears in the suggestions dropdown.
“You just want to be what people associate with the query,” Fishkin said. “Now you own your traffic,” which, he added, is a potential SEO tactic for the future.
Google Suggest also shows brand queries that earn strong connections to URLs. Even some generic queries – like for “maps” or “toys” – bring up branded domain suggestions for Google Maps and ToysRUs.com.
Fishkin uses the example of asking a roomful of people – like, say, the SMX crowd – about what results they expect for a given query, like, “tech news,” or, “New York best restaurants,” and how it is fairly easy to predict which brands will rank highly and then saying, “Yup,” upon seeing expected results.
In fact, Fishkin said the best way for brands to rank in 2018 is to “find a way to be the first, ‘yup,’ that everyone yells out in the room.”
5. Tracing The Visit Path To An Answer
Google wants to disambiguate the query path to get consumers to completion faster, Fishkin said.
In other words, if Google sees that many people who perform these types of queries – like for “best ramen noodles,” “instant noodle brands,” “tastiest packaged noodle,” etc. – eventually end their queries on the topic after visiting instant noodle review site The Ramen Rater, Google might use the clickstream data to help rank that site higher, even if it doesn’t have traditional ranking signals.
6. Weighting Elements Of User Experience
Since launching its Panda algorithm update, Google has tried to surface not just quality content, but high-quality websites, which is why Panda can hurt a site with good content by having lots of bad content on it as well, Fishkin said.
And if Google isn’t already doing it, it is at least thinking about how to measure UX and rank sites that do it better higher, he added.
7. Replacing Flawed Humans With Deep Learning Machines
Fishkin also points to a shift from human-created algorithms to machine learning, which he said “changes the equation for how we are thinking about SEO.”
In machine learning, a computer is given images from across the Internet and trained to recognize and qualify said images and comparable images as pictures of, say, cats.
Deep learning takes it one more step, Fishkin said.
“We give the machine a bunch of pictures and we don’t have to give it categorization, but it figures out these things should be cats,” Fishkin said. “That’s super powerful and suggests if Google can do that for YouTube videos and images, why not for ranking elements? Replace YouTube with the Web and cats with any given search query, and it’s not hard to imagine Google creating a deep learning ranking algorithm.”
For example, Fishkin said that when he searches for “Paul Graham,” Google knows that based on his history, he is probably looking for Paul Graham the programmer and not the photographer.
And, in the future, even Google’s search quality engineers may have no idea why something ranks or whether they’re using a particular factor in the ranking algorithm, Fishkin said. The machine will simply ask, “What algorithm produces results that searchers engage with best?” and then make it, he said.
For Google, It’s All About User Experience
In addition, according to Fishkin, Google seems to be going down what he calls “a strange path” in which total searchers, number of searchers, and searches per searcher are going up, so he questions whether Google is sacrificing ad impressions to make searchers happy and whether Google is willing to take away queries that provide revenue.
In the end, Fishkin said he thinks Google is thinking long term.
“They want addicted searchers providing data about themselves so they can charge more per ad unit,” Fishkin said. “Facebook has shown Google that more data about users yields more dollars per impression and click. I think Google will chase better UX to almost any extent in order to keep searchers and get data, even at the cost of their existing model.”
In other words, Fishkin said Google cares about user experience almost more than money and could be sacrificing revenue directly to provide a better experience, which he calls “pretty innovative and brave.”
“Facebook has shown that by having all that data, it can charge more for impressions and clicks than Google can, which is scary to Google. That’s why there are so many algorithms around users, user experience, clickstream models,” Fishkin said. “It’s predicting people’s happiness. If you keep changing to better the user experience in order to keep searchers, you can keep collecting data. Google will chase better UX to almost any extent.”
What’s your take on Fishkin’s potential future search signals?