5 local SEO insights from Google’s API documentation leak

When news of the leaked Google Search API docs broke last week, our team quickly crawled them to look for anything relevant to local SEO. 

My first take was most of the “local” stuff was either fairly basic (“Hey, Google uses business addresses!”) to mostly irrelevant map nerdery (“This content describes a representation of a material used to define the appearance of geometry surfaces in a city data format, with attributes such as color, surface smoothness, reflectivity and transparency.”)

All I wanted to know is how this information will help our clients rank better in local SERPs.

Spoiler alert: I’m still not sure, but I have some thoughts/questions.

First, here are some common observations about the data that “may be a thing” that have been popping up around the web:

Chrome browser behavior.

“Toxic” backlinks.

Google may limit the number of different types of sites that appear in a given search result.

“Mentions” (a.k.a. “citations”) of your site on other sites.

Topic authority.

Clicks on your URL in search results affect rankings.


You may have to update a page 20 times before Google considers it a true refresh.

There are potentially thousands of other factors, so let’s get to the point. Is there anything in there that can help us improve our local SEO?

Below are some ill-formed thoughts for you all to chew on.

1. Video, video, video!

This isn’t really a “local SEO” tip, but thus far, it’s the most actionable thing I have found – or at least I think it is. Here’s what the docs said about video:

isVideoFocusedSite: Bit to determine whether the site has mostly video content but is not hosted on known video-hosting domains. The site is considered to be video-focused if it has > 50% of the URLs with watch pages.

People seem to like video, right? Anecdotally, we have noticed video results increasing in search results across virtually every vertical we work in. 

For example, here is the presence of video in SERPs according to Semrush for non-brand keywords for RotoRooter.com:

So, if you want to be considered for a video slot on a SERP, I’m guessing you’d want to be classified as an isVideoFocusedSite

Here’s the “local” part: most local SMB sites are small. It seems to me that you could quickly turn a five-page site into a video-classified site by adding five quick selfie videos on their own pages. 

Google seems to be picky about what it classifies as a video page.

For example, if you’re a veterinarian, you could do a selfie vid on how to brush your cat’s teeth, how often to schedule a wellness exam for your dog, etc. You’d have to self-host them or find an off-brand video hosting domain. Hit me up if you know of any.

By the way, “geolocation” is listed as an attribute in this document about meta information extracted from a video file, so it couldn’t hurt to make sure your video has your location in the meta information.

2. Are local bot clicks ‘CRAP’?

If you’ve been involved in local SEO for a while, you’re likely aware of various services that use bots to simulate location-based searches and click on your results, aiming to boost your local pack rankings by enhancing your “prominence” signal.

While much has been made of Rand Fishkin’s years-old proof that clicks matter for non-local SERPs, I don’t think I have seen anyone really talk about it publicly for local packs. 

That’s why QualityNavboostCrapsCrapsClickSignals.t, which is defined as “CRAPS signal for the locale,” caught my eye. 

CRAPs stands for, I think, “click and results prediction system.” This suggests that a specific location could have a specific score for how clicks on results affect rankings. If so, how could you tell what that score is? 

The first thing that comes to mind is to run a bunch of local bot clicks on competitor businesses for a number of different related queries (vet near me, animal hospital, dog vaccines in Pleasanton, etc.) and see how many clicks it takes to move the rankings. 

Test it periodically to see how it changes over time or on different days. Once you have hit on a formula that seems to work, apply it to your site.

Note that this is likely 100% against Google’s TOS, so I am not advocating you do this. I’m just reading the tea leaves.

3. Local authority vs. topic authority

We typically can boil down local SEO to a combination of proximity (are you near the searched area?), prominence (are you “good enough” to show up for this query?) and relevance (are you relevant to the query?). 

This API doc on NSR (likely “normalized site rank”) references: 

titlematchScore which is “score the site, a signal that tells how well titles are matching user queries.” 

localityScore which is a “component of the LocalAuthority signal.”

In theory, if you have titles throughout your site that are relevant to the user queries (a.k.a. “relevance”) + a strong localityScore (a.k.a. “proximity”), you should have a decent shot at good local rankings, assuming you are prominent enough. 

Typically, you can’t do much about proximity other than perhaps create location pages and get reviews that mention the location.

For the purposes of this thought exercise, let’s assume your locationScore is fixed. That means playing around with the title tags across the site could yield results. What do I mean by “playing around”?

Let’s assume you rank third for “SEO Company Pleasanton” in the local pack. If you crawl the sites of the first two businesses, you can see that about 10% of the title tags on each site target some version of this query. 

So, what if we updated the titles on our site so that 20% of the titles hit the titlematchScore? Could that improve our local rankings? Maybe it’s just for the organic results and not the local pack? Seems really easy to test. I think I know what I’m doing tomorrow.

Get the daily newsletter search marketers rely on.

4. LSAs vs. Google Ads?

Fishkin recently pondered,

“If Chrome click stream data is used for rankings, does that mean paid clicks could boost organic rankings?”

Let’s assume for the moment this is, in fact, how it works. If that’s the case, another question would be which ad unit is most effective for moving the rankings needle.

In local search, we have your standard PPC ads, but we also have Google Local Service Ads (LSAs), which show up above local packs and other local “surfaces.” 

It seems to me it would be pretty straightforward to test different ad units against rankings and organic clicks. 

Actually, that might cost you some money, so better yet, how about tracking some competitors who are spending a lot of money on these and seeing how their rankings change/don’t change?

5. Twiddle with local results

Mike King has this to say about Twiddlers:

“Twiddlers can offer category constraints, meaning diversity can be promoted by specifically limiting the type of results. For instance the author may decide to only allow 3 blog posts in a given SERP. This can clarify when ranking is a lost cause based on your page format.”

SEOs have long focused on the intent of a query by examining the types of results on the SERPs. So perhaps this whole Twiddlers thing aligns with the chorus of the “nothing new here”-niks. 

That said, this turned a lightbulb on for me on our approach. A common result of a local intent check is that an organic SERP has a few local businesses and business directories (e.g., Yelp, Angi, Forbes, etc.). 

Instead of worrying about how hard it is to compete with sites like Yelp, I now think, “There are only three directories in this search result. How can my local business become one of them?”

This is no knock against Yelp or any of its ilk, but what is that site at its core besides a list of businesses and content about them? 

If I were a local accountant, it wouldn’t be hard to put together a page on my domain or a new one about great local accountants in my city, and I am – for some reason – listed number one. (Not sure how that happened, but I’ll take it. )

This has been a go-to B2B play for years. No reason why a local business couldn’t do the same.

Some attributes I want to know more about

This IndexingDocjoinerDataVersion doc has some fairly intriguing Attribute names listed. I have no idea what they are, but it seems like several of these may play a big part in local SERPs:

localyp (Perhaps how they classify local business directory sites?)

localsearchAuthoritySiteAnnotation (As in “this site is an authority for this location”?)

qualityGeoBrainlocGoldmineBrainlocAnnotation (I am pretty sure GeoBrain is Google’s list of popular locations. Goldmine seems like it would be a list of advertisers?

indexingDupsLocalizedLocalizedCluster (Dedupes results based on the searched geo?)

imageRepositoryGeolocation (It has been a long time since we have seen geotagging images have any effect on local rankings, but it still makes sense for Google to store these in order to show them for specific types of queries.)

knowledgeMiningFactsLocalizedFact (If a fact has a local source/application, show the local version of it instead of the “national” version?)

tofu (Defined as the “URL-level tofu prediction,” this may be the key to the entire Google algorithm. It probably isn’t. I just couldn’t resist mentioning it.)

I’m trying really hard not to end on a “nothing actionable but happy testing” note. Ultimately, I guess I am no different than any other SEO guru wannabe.

Oh, yeah. I forgot to mention that the whole “mentions” thing could mean local business citations for “third-tier” directories still may be helpful for rankings. But I’ll let a listings management company jump into that can of worms.

Hopefully, I have given you some ideas to mess around with. Have fun!