Google Search gets deeper into the ‘real-world’ with Busyness, Duplex and AR in Maps
Building the data bridge between online and offline.
At its virtual SearchOn event Thursday, Google made a range of announcements about how it’s making search more sophisticated and user friendly. The company also announced that it now uses its natural language understanding BERT algorithm on almost every query to better understand search intent.
Many of the announcements shared the common underlying theme that they help users better understand and navigate the physical world. Among them:
- Using Duplex to update local business information
- ‘Busyness’ data on physical places
- AR information about local businesses (in Google Maps’ Live View)
Duplex dials up local data
One of the historical challenges Google has had in local involves accurate and trustworthy business information. Previously,, Google addressed this by relying heavily on third party directory sites and things like NAP consistency. Over time, Google has shifted away from third party data and relies much more heavily on Google My Business and its own sources.
Google Duplex, a virtual assistant originally designed to help consumers make appointments or reservations at local businesses that lacked online scheduling, further accelerates this trend. Since last year Google has used the technology to confirm local business information with owners directly. Google said yesterday that Duplex has been instrumental in “over 3 million updates to businesses like pharmacies, restaurants and grocery stores that have been seen over 20 billion times in Maps and Search.”
While Google presents this matter-of-factly it’s radical stuff.
Somewhere between 50% and just over 60% of Google My Business profiles in the U.S. are owner-verified. However, even those that are claimed in many cases are still incomplete. Google is proactively reaching out with Duplex to fill in the critical blanks and may do even more over time. It’s also a not-so-subtle marketing tool for Google that the company could potentially use to push it’s own messages or to encourage businesses to try specific GMB features.
Google is also testing using Duplex to confirm “in-demand” inventory at local retailers. Google acquired Pointy to help get small business inventory online. This is another tactic to help provide more real-time local product information, which has huge interest among consumers.
Getting busy with ‘Busyness’
For some time Google has posted wait times/popular times on Google local profiles, especially restaurants, but increasingly other categories of business and local places (e.g., Post Office). These estimates are based on algorithms that use historical visitation information from anonymous mobile users’ location history.
Google said in a blog post that it has substantially discarded pre-COVID popular times data and is using “more recent data from the previous four to six weeks to quickly adapt to changing patterns for popular times.” Google also uses real-time mobile user data to show “live busyness.”
This is obviously very valuable information, as people decide whether and when to visit places locally. However, there’s evidence that some of the time estimates may be inaccurate. Spot checking of local restaurants by Mike Blumenthal, after a GMB forums complaint by BJ’s restaurants, indicated more than isolated inaccuracy on restaurant wait times. These problems have apparently been reported to Google.
Local business AR in Google Maps with Live View
Announced in 2018 and rolled out last year, Google introduced augmented reality (AR) walking directions to Google Maps. Called Live View, it uses Google Lens capabilities to overlay navigation on top of Street View data. It’s a great tool that doesn’t always work in practice.
During the SearchOn event Google spoke about a range of AR use cases and examples. The most practical one is the long-promised ability to point the smartphone camera at a business and see its GMB profile (e.g., reviews). Google has been working on this capability in various forms for a decade. It requires highly accurate location data — when your phone is pointed at a business across the street or down the block it has to know what specific business or place you’re looking at. Perhaps only Apple or Microsoft could also pull this off.
Many other local search apps, well ahead of their time, tried and failed at similar functionality (Yelp Monocle, Zagat, Layar). And before Google Lens Google Goggles promised this same capability more than a decade ago.
Because most people don’t make totally spontaneous decisions about where to go or eat — some amount of forethought or planning is typically involved — it will be interesting to see how this feature gets used and evolves. In some cases, like travel, I could imagine people using it to determine where to eat. But I don’t believe most users will regularly consult Lens as a way to make local purchase decisions.
Bringing more of the real world into Google
Back in 2010, Google’s head of search, Prabhakar Raghavan (then at Yahoo) spoke about the concept of the “web of things” vs. the “web of objects” (documents). He said at the time that 99% of search queries on Yahoo have a noun in them. According to Raghavan that reflected the search for real-world information. If you think about it, it only makes sense.
Google continues to refine its algorithms to do a better job of understanding websites and delivering content responsive to search queries. But in many ways the more impactful and interesting development is the way Google continues to build connections between the digital and real worlds.
Many of the tools discussed at SearchOn (Duplex, busyness, AR) are in one one way or another all directed toward making local search and Google Maps more dynamic and reflective of the real world and not just a presentation of data on websites. Indeed, Google increasingly has positioned itself as a gateway to offline activity and buying behavior. And that’s worth trillions, not billions of dollars.
Opinions expressed in this article are those of the guest author and not necessarily Search Engine Land. Staff authors are listed here.