3-pack ranking discrepancies in private vs normal browsing


Joined
Mar 27, 2013
Messages
37
Likes
7
Recently (probably since the Possum update) we are seeing some major discrepancies in the local 3-pack. We see great variation in 3-pack results when searching in private/incognito mode vs not.

The most concerning issue is that the local businesses that are now ranking well in private browsing modes (and in some cases non-private modes) really have no business being there...horrible websites, no reviews, un-matched GMB categories, etc. In many cases there seems to be no reason why many 3-pack listings should be ranking well.

Is this consistent to what others have seen since Possum?

Any feedback would be appreciated!
 
Joined
Jun 28, 2012
Messages
14,436
Likes
4,292
Thanks Eric for the heads up. Sucks that the results are not just different, but bad!

I never browse incognito so have never noticed, but guess I should.
I imagine lots of consumers do if they don't use Gmail.

Anyone else noticed or care to do some research/comparison screenshots?
 
Joined
Mar 27, 2013
Messages
37
Likes
7
Thanks Linda...it's very curious. I'll add a bit more info and detail along with some screenshots to help paint the picture.

We rarely search in incognito/private either, but we have noticed similarly different/bad results when using mobile phones in private mode (which produces identical 3-pack results as a mobile search with location services turned off...which is probably much more common among consumers).

We have noticed the same issues with clients, but results for our own business produce the most glaring example that we've seen. I've attached 4 screenshots that show 3-packs for the search query "web design"...these were all done from our office from the same WiFi network (we actually see the best and probably most accurate results with a mobile search on a mobile network but for this thread we'll use WiFi searches for comparison)

1. desktop search in normal browsing mode



2. desktop search in incognito browsing mode



3. mobile search (iOS) in normal browsing mode



4. mobile search (iOS) in private browsing mode



As you can see, there is great variation...and the 3-pack listings contain business in other categories (e.g. Graphic Design, Advertising Agency), while there are plenty of 'Website Designers' in our area. In addition to category issues, most of the ranking business have poor websites, little to no reviews, contain GMB violations, etc. There really is no reason why a lot of these places should be in any 3-pack whatsoever.

And, ironically, the mobile search in normal mode is the only example that shows us (Zee Creative) in the 3-pack, but the #1 listing is for a business that is not even located in our city.







 
Joined
Jun 28, 2012
Messages
14,436
Likes
4,292
Since this post came in over the Holiday, I wanted to bubble it back up to be sure you guys saw it.

What do you make of it? Do you see the same thing?
 

JoshuaMackens

Local Search Expert
Joined
Sep 12, 2012
Messages
1,502
Likes
314
If it's since Possum it would make sense that, based on the private browsing, Google is using your location outside of private browsing and not in private browsing. If that's true, we believe with Possum location now plays a much greater role in ranking. That might explain this massive variation.

Would want to hear the opinions of others as well though.
 

JoyHawkins

Administrator
Joined
Jul 18, 2012
Messages
2,323
Likes
1,351
So the city the business is in is irellevant because you didn't type in a city in your search (just "web design). According to the screenshot, PC-Solutions is closest to you (it says 1.3 miles, your own company is 2 miles). Mobile is super-hyper-local and the location trumps almost everything else, especially if you don't enter in a location name.

When I do a search for "web design billings mt" on a computer and set my location to "Billings", you guys are 2nd in the 3-pack and first organically.

As far as why you aren't there on mobile in icognito - notice how the distance away vanishes? Google is unable to determine your exact location but it looks like they have a general idea of where you are which is why they are still showing local businesses. However, your little blue dot (where they think you are) disappears so the businesses based on your location are bound to be different.
 

Eric Marshall

Forum Member
Joined
Mar 27, 2013
Messages
37
Likes
7
Thank's Joy. That helps bring some clarity to the mobile portion...I don't think I realized how sensitive the mobile algorithm was to location.

On the desktop, I still think those are really bad results and I can't make sense of it. In both scenarios, 2 of the 3 local results have "graphic designer" as their main category along with no or few reviews, poorly optimized sites, etc. It seems to be somewhat of a random pick from the top 20.

Do these seem out of place to you, or is that typical to what you'd usually see?
 

JoyHawkins

Administrator
Joined
Jul 18, 2012
Messages
2,323
Likes
1,351
Eric,

I can't recreate what you're seeing but I also don't know where Google thinks you're located. The map on the first screenshot definitely looks odd as it's zoomed out to show the entire US.

If you go to Maps.Google.com it will usually place a blue dot on the map for where Google thinks you're located. What zip code is it putting you in?
 

Eric Marshall

Forum Member
Joined
Mar 27, 2013
Messages
37
Likes
7
I'm actually not seeing a blue dot...but based on the lat/long it gives in the URL, they are placing us in 59102
 

JoyHawkins

Administrator
Joined
Jul 18, 2012
Messages
2,323
Likes
1,351
Okay, I still get something different but seeing something similar in the sense that you guys are really high organically but 10th locally.

One possible reason for this could be that it seems you guys have moved a lot? I found 3 different addresses listed online:


(Linda) Removed links to old incorrect addresses via Eric's request so we don't help propagate the bad data.

 

Eric Marshall

Forum Member
Joined
Mar 27, 2013
Messages
37
Likes
7
The 2 Billings addresses addresses are old addresses of ours that keep popping up (the WY address is not one of ours)...we're going to try and get those cleaned up and see if that makes any difference.

On desktop...Google seems to be placing our search location in 59102 (our physical location when searching is 59105) which then skews the local results to businesses in 59102 and 59101...our business is in 59105. That seems to be a factor as to why we aren't showing as prominently.

Does anybody know how Google determines searcher location on a desktop search?
 

Eric Marshall

Forum Member
Joined
Mar 27, 2013
Messages
37
Likes
7
The 2 Billings addresses are old addresses of ours that keep popping up (the WY address is not one of ours)...we're going to try and get those cleaned up and see if that makes any difference.

On desktop...Google seems to be placing our search location in 59102 (our physical location when searching is 59105) which then skews the local results to businesses in 59102 and 59101...our business is in 59105. That seems to be a factor as to why we aren't showing as prominently.

Does anybody know how Google determines searcher location on a desktop search?
 
Joined
Jun 28, 2012
Messages
14,436
Likes
4,292
On desktop...Google seems to be placing our search location in 59102 (our physical location when searching is 59105) which then skews the local results to businesses in 59102 and 59101...our business is in 59105. That seems to be a factor as to why we aren't showing as prominently.

Does anybody know how Google determines searcher location on a desktop search?
It's largely based on your ISP and IP.

My IP from my ISP puts me in Carlsbad, 4 cities over, which I think is where my ISP is located.

Your IP shows you are in Billings at 45.7833 -108.5007. Plug that into maps and see if it's near your physical location or if its in the 59101/2 area.
 

JoyHawkins

Administrator
Joined
Jul 18, 2012
Messages
2,323
Likes
1,351
Google's MyLocation first debuted on cell phones and other mobile devices. In that case it would use 3 sources of information to determine location.

GPS - Many phones have GPS built in, but may not want to use it for battery reasons or because it takes too long to get a lock.Cellular - Google Maps can use nearby cellular towers to triangulate an approximate position. This is more accurate in urban areas due to the higher density of towers.WiFi - Finally it can use WiFi in the same way that cellular towers are used, as well as with IP information. Based on nearby WiFi networks and by doing a trace on the IP address through the ISP this can sometimes be as accurate as Cellular, but normally just ends up being city level accuracy, or worse in some rural areas.

Most phones end up using Cellular, since it is fastest and doesn't add to the battery drain.

Most computers use WiFi/IP information, since no other options are available
(Taken from http://superuser.com/questions/12495/how-does-google-my-location-work)
 
Joined
Mar 27, 2013
Messages
37
Likes
7
This is all very helpful...thanks for all the feedback.

What we're really trying to do is to get an accurate idea of what our potential clients are seeing when searching for our services in our area (and also how our local search clients are showing up in the local results).

Most of the commercial and residential population in our city is in 59101 and 59102 and for some reason Google is placing our location, when searching from our office in 59102, rather than 59105. We're trying and testing some things here to see if it's a firewall or network issue, or something else.

At any rate, we rank very well in the 3-pack if someone is searching outside of the city (web design billings mt) but it seems like once you're in the city, proximity as a signal increases drastically...that same search would yield very different results based on your actual location.

Again...thanks for the help getting some clarity. I'll let you know if we're able to make any more sense of it.
 

JoyHawkins

Administrator
Joined
Jul 18, 2012
Messages
2,323
Likes
1,351
I personally would rely on Search Console data to give you an idea of how much you are showing for "web design" searches - either that or use a tool like Bright Local that can scan from different zip codes and eliminate personalized history from influencing the SERPs.
 

JoshuaMackens

Local Search Expert
Joined
Sep 12, 2012
Messages
1,502
Likes
314
This is all very helpful...thanks for all the feedback.

What we're really trying to do is to get an accurate idea of what our potential clients are seeing when searching for our services in our area (and also how our local search clients are showing up in the local results).

Most of the commercial and residential population in our city is in 59101 and 59102 and for some reason Google is placing our location, when searching from our office in 59102, rather than 59105. We're trying and testing some things here to see if it's a firewall or network issue, or something else.

At any rate, we rank very well in the 3-pack if someone is searching outside of the city (web design billings mt) but it seems like once you're in the city, proximity as a signal increases drastically...that same search would yield very different results based on your actual location.

Again...thanks for the help getting some clarity. I'll let you know if we're able to make any more sense of it.
Hey Eric,

You're actually feeling the effects of the natural evolution of Local SEO. Rankings are a wonderful indicator of how you are performing as an SEO. However, as Google evolves, with personalization and location becoming bigger and bigger factors, we're heading into a time where accurately pinpointing how we look to searchers on Google Search is going to become next to impossible. In fact, I would argue that with Possum, it's already here.

The best you can do is...do your best. Optimize the way you know you should and benchmark that against your ranking reports with your static location. This will show you how you are doing in a general sense so you can adjust and optimize more. And really, that's the best you can do.

The only other option is to run 10 different location IP's in 10 different areas of your location to see how people in the north, south, east, west, etc. of your city are seeing you. But even then you probably will want to be more granular at some point and have 20-30 different IP locations in the city to really flesh that out. A) that's currently impossible from what I see and B) that's exhausting. Again, we're at a point where we just have to continue to optimize, benchmark, and optimize. That's about all you can do and about all any of us can do.

Unfortunately, that's just the lay of the land moving forward. And it's not all doom and gloom. Optimizing for Local SEO isn't changing as much as our ability to benchmark. We just have to adjust.
 

Local Search Forum


Weekly Digest
Subscribe/Unsubscribe


Google Product Exert

@LocalSearchLink

Join Our Facebook Group

Top