Where is Google Headed With Mobile?

Many have expected 2015 to be the year of mobile for Google. While Google has stated that there are now more searched on mobile than on desktop, the supposed Mobilepocalypse was something of a disappointment.

Some of this may have been because most websites were simply ready for the mobile algo changes. Alex Holmes of ThemeForest.net told me earlier this year that “pretty much every template now needs to be responsive or it won’t sell.”

On the other hand, before Mobilepocalypse, testing by Portent determined that 40% of websites were not mobile ready.

The fact that the reported impact has been minimal suggests that at the very least, Google oversold this algo change.

Now to blur the issue even further, Advent Communication has released the results of a study suggesting that Google really doesn’t know where you are located.

Advent asked respondents to compare where Google thought they were located with their actually location. They admit that their sample size was a bit limited, but found that the average error for desktop users was 117 miles, while it was 428 miles for mobile users.

I’m most surprised by mobile having a higher error. Granted, desktop users are by definition static, but I would think that between GPS data in smart phones and cell phone tower data, they should have a pretty good idea where mobile users are located.

While the average error numbers are probably biased by having some users thousands of miles away with a small sample set, the errors were still widespread. Only one-third of users were within 25 miles of where Google thought they were.

These errors have huge implications for AdWords advertisers who are attempting to control the geographic distribution of their ads. They both have people outside their target area seeing the ad (false positives) and people within the target area not seeing it (false negatives).

There are also impacts for searchers. Google is attempting to return local results for queries that don’t have a location (e.g. search on [coffee shops]). If Google has an incorrect location for you, the SERPs won’t be very useful.

I predict that Google will get better at handling this data, and over time we will see more localized results on non-geographic queries, but will see more and more hyper-local results as well.

Expect the radius of results to vary with the query. If I search for [coffee shops], I probably want something close to me. But if I search for [farm to table restaurant], I might be willing to drive a while.

The radius should vary by geography as well. If I search for [barbecue joint] in Philadelphia, I might expect to have to drive, but the same query here in Austin should show me things very close by.

It will be interesting to see what Google does with this over the next year or two.