May 4, 2015 at 6:16 am #7727Mike BlumenthalKeymaster
Google has long described the elements of the Google Local ranking algo as proximity, relevance and prominence.
Local search’s relevance factors, long having been categorical search, were driven by categories at Google, particularly custom categories. And to my simplistic way of thinking, they appeared to be somewhat binary. Either a business was properly categorized in a certain way or it wasn’t. Either a business within a geographic boundary fell into a given category or didn’t and then rank would be determined by location prominence factors.
But the recent experiments with Yelp review content as a driver of Local result clearly point to Google’s thinking about relevancy as a value along a scale not a yes or no choice. Like many things that Google does there is a valence to relevance and a given listing needs more of it rather than less. It is a value that will define whether a local result shows at all even within Maps and whether any given listing is trusted enough to be shown on the hallowed space of a Google.com front page result.
A Model for Relevancy
We have known many things that become categorical; Business name, Google categories, categories at the IYP, review content, authority (home or landing page) page content & links of the business. But it appears that Google has expanded the concept of category relevancy (or at least enough to drive a local result in the main SERPs) as any keyword phrase that occurs on authoritative local sites (most likely in conjunction with other signals) in the right context. For example the hardly used phrase “rings appraised Cheektowaga NY” generates a OneBox as does the phrase “Google Local Olean NY”. These are not targeted phrases for local on either website. They could be triggered by semantic relationships or business graph relationships but they are not triggered. In the case of 3rd Base the use of the words “dive bar” was used nowhere on their site or on other sites except Yelp. You can often pick odd two word phrases from Yelp combine them with the city of location and drive a OneBox at Google. Where Google came up with the town Cheektowaga for Barbara is not clear. It is one or two suburbs away from the actual store and not used ANYWHERE as far as I know.
Now take this expanded category model and add to it some priorities and assumptions (OK so this is total conjecture but stay with me). Let’s assume that the highest category value for any business is 10.
Business Name Match – category value 5
Category Match at Google – category value 2
Keyword pair occurrence in reviews – category value 1 1.5
Category Match at selected IYPs – category value 1
Home page content match with links – category value 1
And finally lets speculate that we need certain values for Google to move the listing up the food chain of search.
Relevance as a threshold
In the case of the “Sports bars Olean NY” example 3rd Base was second in Maps but was not showing in the main search results against the better branded “Dugout Sports Bar”. 3rd Base was not categorized as a sports bar at Google but was at Yelp (and probably elsewhere).
The addition of the 3 keyword pairs (there is nothing magical about 3 it just happened to be the number of times I could get the phrase into Yelp) left 3rd Base at second in Google Maps but broke the listing into the main search results for the phrase “Sports Bars Olean NY”.
Thus, on the above scoring system, 3rd Base started with a category value of 1 (category at Yelp) but added enough category juice to at least achieve main page display parity with the keyword branded & properly categories “Dugout Sports Bar”. In this model, it had enough relevance value to surpass the threshold value imposed by Google.
Yelp & Pixie Dust
I don’t think that there is anything magical about Yelp. There are other sites and ways to achieve lift off. I do not believe that relevancy can come from an infinite number of sites and I do think that Google needs to have classified the sites at Locally prominent but I have seen plenty of examples of One Boxes that show no Yelp correlation. They have achieved the relevance threshold somehow.
This is a great example of that – and its in the real estate arena which adds more intrigue. If I had to guess it was due to great local website optimization.
Rethinking Hummingbird & Pigeon
In Hummingbird we saw a massive increase in OneBox results and OneBox spam. We long thought that Google was identifying company name as a critical Local factor. What if what Google has done is just readjust the values assigned to categorical information and these many oneboxes could be countered with additional listings with increased relevance?
As for Pigeon, we saw a drop in Local Packs. Perhaps what has happened is that Google increased the relevance threshold needed to make it to the front page? Thus listings could repopulate the front page over time if they achieve a higher relevancy score? Maybe the lost of real estate local listings on the front page could be overcome with better focus on increasing relevancy of a given listing?
Obviously there is a lot we still don’t fully understand about the Local ranking system. Can a non keyword named business compete in the local results if it gets enough relevance elsewhere? Can we create packs in real estate with just a focus on increasing relevance? We do know that there are plenty of categories where Google will show results (sushi, shuffleboard) where they don’t have categories, which sources are most valuable for Google?September 12, 2015 at 10:37 am #9243Mike PedersenMember
What is the best way to get ahold of Google? I know they keep changing it.
I would like to get them on the phone again if possible.September 12, 2015 at 10:42 am #9245
- You must be logged in to reply to this topic.