Google Rolling Out Assistant Duplex Test to Four Cities This Month

At Google I/O a preview of Google’s Assistant Duplex, Google’s AI based human sounding  digital assistant that can auto create reservations, set off a maelstrom of both interest and criticism.

The Announcement from the Google My Business Forum

Hey everyone!

At our annual conference Google I/O, we showed a preview of a new feature in the Google Assistant powered by Google Duplex technology – which makes it easier for people to find information and get things done — and it’s also helping businesses connect to potential customers.

Later this month, we’ll expand public testing of this new feature with Google Pixel users at restaurants in New York, Atlanta, Phoenix and the San Francisco Bay Area! Our goal in testing this new capability within the Google Assistant is to provide another great resource (especially if you have a limited online presence) to help you reach a broader set of consumers, by making it easier for them to connect to your business.

There’s no learning curve for you and no extra step to get set up. You can keep operating your restaurant as you always have while also taking advantage of booking through the Google Assistant. And if you’re currently working with one of our online booking partners , you won’t receive calls from the Assistant for reservations.  

As we develop new technologies, we believe it’s critical that we’re responsible in our approach and are thinking through the impact on people and businesses. You’ll be given full transparency and control — so you’ll always know it is the Google Assistant calling you. And you can easily opt-out from receiving calls over the phone by telling the Google Assistant or going to the settings in your Google My Business account.

It’s important that we get the experience right both for people and for your business, and we’re taking a slow and measured approach as we incorporate learnings and feedback from our tests.



What’s Up with the Local Search Results…. Who the f knows?

I was sitting on my office, doing “research” and did the same local search (Custom Jewelry Design Williamsville) on my iPhone in three different Google mobile environments; Safari, Chrome and the Google App. All within seconds of each other.

Next time your boss comes and asks you what’s up with the local search results, throw up your hands, show him this post and reflect knowingly “who the f knows….”. Be sure to note that even though you were not at your desk when he first came by, you were using your time wisely.

Updated: Just for good luck we have added a mobile search from Stephen Davies of the UK. Why have 3 variations when four (or who knows how many more) can do?

Safari browser result. 1500 Pixels high. Includes category, 3 facetted search options, and website content.
Chrome browser result. The longest clocking at almost 1600 Pixels high. Like the Safari result Includes category, 3 facetted search options, and website content but also includes review content.
Google App (iPhone) result. The shortest pack result clocking at just under 1200 Pixels high. No category, only 1 faceted search result, no review stars, no web content, no review content, no images. Most importantly (as pointed out by Stephen Davies) it includes direct access to calling and driving directions. It also includes the more button. Go figure.
And one more variation…. As seen from the UK via Android Chrome. Thanks to Stephen Davies.

Local Mobile Image Thumbnail SERPS – It’s All About Context

Last December we started seeing thumbnails in the mobile local organic results. What was obviously a test then became more visible over the summer and starting last month, became widely visible.

It is a mobile only feature that made me and several clients curious. How was it triggered? Why & when does it show?

Why it matters?

These images create significant eye candy in the organic local mobile results. I have not yet seen click through rate studies but I have to think that having an image will serve an organic result well in terms of increasing click through rates. Even when these images are lower on the page.

How I tested

After struggling with viewing these on my phone,
I created a desktop test environment that included Chrome and a mobile browser emulator that allowed me to look at mobile, desktop and source for the page in a side by side way.

It’s all about file name and image alt tags, NOT

When I started this project in late summer, I assumed that image file names and image alt tags would play a big role in Google’s use of these images in search.

That turned out to not be true. In fact those factors were the least correlated with whether an image showed of all the factors I did look at.

Out of my sample of 20 only 4 sites were using meaningful file names AND a relevant image tag.

Clearly alt tags and file names had such a low correlation with whether the image showed, that it forced me to look further afield for stronger correlations. While these more technical features might impact the result they were not a clear and consistent driver.

I would even suggest that the low incidence of these image attributes across the web is what lead Google to start looking at alternative ways to understand images on the page.

What is it about? Context

While I did notice early on was that context seemed to be significantly more important than image name and alt tag. So I increased my examination of contextual signals to find better correlations.

I looked at which image Google used, its position on the page, image relationship to the main body copy and the relationship of the search query to the page Title Tag and H1 tags.

What I found

  • Google is ignoring top of page images like logos, images in navigation and most carousels when choosing which image to show in the mobile SERPS.
  • The image that was used was most likely to be the first image in the main body copy of the site. On most of the pages there were multiple images embedded in the main content on the page beyond the first one.
  • 19 of the 20 images that showed up in the search results were above the fold on the web page.
  • While generally it was true that Google picked the first image in the body, in two of the 20 instances when there was a row of images, the image picked was the last of a row of images.
  • On mobile search results that showed the image there was a near 100% correlation  between the page Title Tag and the search query. In 15 of the 20 cases there was a H1 tag that closely matched the query as well.
  • I don’t think that it is the Title or H1 tags per se that is causing the image to appear, but I do believe that when the overall context of the page matches the search query and there was an eligible image, then Google is likely to show the image.

Most of my searches that did show the image in the search results were higher frequency head terms.  The pages are more  likely to be optimized for the search phrase and the concept.

I hypothesized that if these context clues of Title Tag and H1 were critical then there would be many fewer images on very long tail local queries.

That in fact, turned out to the be case. Searchng for “ob-gyn in Santa Monica” would return images while a search for “female ob-gyn Santa Monica” or “best female ob-gyn” would consistently not surface images. This was generally true across many long tail searches that I tried.

Google’s New Image Search Algo seems to be the driving force

This all makes sense given Google’s recent announcements about their Image search algo.

Cyrus Shepard summarized the above in a tweet to say:

I had come to the same conclusion several weeks ago and it was nice to have confirmation: relevance of the page is a driving force behind the images being shown in the local mobile search results.

Some Tips if you are trying to control what appears in the local mobile search results

For sure:

  • Make the image the first image in the main content area and make it central to the page content
  • Make sure that the page is topically relevant to your targeted search query with good architecture including title Tags & H1 tags.

Some possibilities that might have an impact:

  • I am not sure if Google is using machine learning to understand the image content but they might be so it doesn’t hurt from a user and long term Googly perspective to make the image contextually appropriate to the target phrases of the page.
  • While alt tags and file titles do not seem to drive these results, they couldn’t hurt and might allow Google to show different images to show for different search results. That is TBD.

Some style thoughts

  • Either center weight the content of the image so that the important things exist in the middle of the image or make the image square. Otherwise Google will crop the image with unintended consequences.
  • Make sure that the image scales down to small sizes well. Google will display the image at slightly more than 100 pixels wide. If the image is too complicated or shot too far away it will not scale down well.
  • Make sure that the image is one that you really want to highlight. When I saw this image of the empty desk for the lawyer, it made me laugh and my first thought was: “If he is not at his desk working, he might not be willing to work for me”.

There May be More than meets the eye*

Continue reading Local Mobile Image Thumbnail SERPS – It’s All About Context

Google Testing New Knowledge Panel Changes

Last week Google, in a disappointing action,  moved Google Posts from front in center of the Knowledge Panel to the bottom of the KP in organic and to the middle of the search results in mobile. Obviously, this is likely to cause a reduction on conversions as the Posts are less visible and lower on the page.  Nate Someson noted this decrease on Twitter.

This week we are seeing signs of a new test that is likely to further reduce searcher interactions with a brand’s Knowledge Panel. In tests first reported by Maulik Pnchal and confirmed by Dan Liebson we are seeing the “People also search for” being moved from the bottom of the Knowledge Panel to the absolute top of SERPS on brand searches:

This test was first spotted by Sergey Alakov who first noted it in May.  It is not visible for most searchers. Sergey has a mobile screen shot.  The fact that the test is still around 5 months later says that it was at least good enough to keep testing. Sergey noted on Twitter that  “the test is expanding, or the roll-out has begun”.

With these moves Google  seems to be minimizing brand traffic, attempting to increase the depth and diversity of searches that start with a brand search and possibly free up Knowledge Panel space for more ads.

Google does not charge for Posts or the KP presence, there are lots of ways to increase income when you run the show.

What Percentage of Verified Businesses Use Google Posts?

TLDR: ~7% of businesses world wide that have verified their business listing are using Google Posts and roughly 26% of verified business in the US were doing so. On a percentage basis the use of posts was highest, around 30%, in Spain, Italy and Japan. The developed world has significantly more updake of this feature than Asia & South America.

Phil Rozak of Localvisibility did a nifty post a few weeks ago exploring business utilization levels of Google Posts. He manually examined a large sample of businesses in medium and large cities to determine that 17% of businesses in the US had used the feature.

In doing my Google Websites uptake analysis I realized that I might be able to use the Google site: search technique to examine this number both in the US and in countries around the world .  While it wouldn’t provide me with nuance like Phil ferreted out that 4% had done an update within the past 7 days, it would give me high level detail. And while Google Websites created is not a perfect proxy for verified businesses, it is a pretty good one.

% of Verified Businesses Using Google Posts
Click to view larger


As of today that stands at 26% of those businesses in the United States with Google Websites (and are thus verified) that are using posts.

As you might expect, utilization as a percentage of those that have created Google Websites is consistently higher in Europe, Canada, Japan and the US than in Asia, Turkey and South American countries:


Total Websites


Posts %

Espana Posts VER TODAS




日本 (Japan) Post すべて表示








United Kingdom








Polska (Poland) Posts WYŚWIETL WSZYSTKO




United States








South Africa
















Türkiye (turkey)








Viet Nam Posts XEM TT C








Україна (Ukraine) Posts ПЕРЕГЛЯНУТИ ВСІ
















Россия (Russia)




Using the searches: “United States” and “United States” + “View All”  you can get a rough idea of the use of Posts amongst businesses that are verified*.

Why is the number higher than Phil’s? Because my methodology is looking at usage of those that have created Google Webpages. That is a more engaged group and they have verified their listing. Both reasons would explain the higher than the overall usage number.

I am also making an assumption that the cohort of businesses that have created a Google Website is similar in behaviors to those that have just verified.

Some notes:

1- After a Twitter convo I realized that since Hotels and a few other categories don’t have posts, these numbers might be slightly high. That being said they are directionally accurate.

2- Also since we have no way of knowing how many businesses are verified in any country we have no way of knowing what the upside opportunity is for Google (or an agency). I did see Bill Hartzer noting that in a 10,000 listing sample of US SABs that 68% of them had been claimed. If that is the case then that would make my US numbers align perfectly with Phil’s. (.68 X .26 = .17)

* To see these numbers for any country you need to modiy the search with the name of the country (in its original language) and the phrase “View All” translated to that language. So a search on Poland would look like: Polska + “WYŚWIETL WSZYSTKO”

Google My Business Websites Zoom Past 10 Million as Asian Adoption Accelerates

In June of this past year, as Google My Business Website reached the age of one, I noted that businesses had created 2 million of these single page mobile ready websites.

The rate of uptake, particularly in Asia and Brazil, has accelerated and there are now almost 12 million of these sites world wide. 10 million of these sites have been created in the past 3 and half months.

What has happened?

While growth in the developed world was solid, growth in Asia was spectacular. Both Indonesia and India have each surpassed 4 million sites in their countries. Indonesia saw increases of 309% per month while India saw monthly increases of almost 200%. This incredible Asian growth story boosted world wide growth to 125% growth per month.

The top 10 spots for the leader board of the countries with the most Google websites saw 9 of 10 countries the same as June. Viet Nam, with monthly growth greater than 100% replaced Germany in the number 10 spot. And Indonesia’s outsized growth allowed it to take the number one position from India.


May ’18 total Sites

Sep ’18 Total Sites

Increase Per Month Since May









United States








Türkiye (turkey)








日本 (Japan)












Viet Nam








Some Factoids to Think About

  • Going from 2 million at the end of May to nearly 12 million today means that Google has been adding almost 93,000 sites per day.
  • India and Indonesia with 4.2 and 4.4 million sites respectively have created 72% of the total sites.
  • The other top 8 countries have created an additional 12% of the sites while the remaining 155 or so other countries in the world have created 12% of the total.
  • Indonesia has been adding sites at the rate of 38,500 new sites per day.
  • There are roughly 100-125 million businesses worldwide in Google Maps. The total uptake of GMB Sites represents roughly 10% of those businesses.

What does it Mean?

Google built Google My Business Websites for the “next billion” users or so they said. The reality is that in much of Asia, where the population never used the desktop web, Google had fallen way behind Facebook in SMB penetration & adoption. They had been marginalized. Google needed an “open web” so that they could leverage their existing technologies to understand what these many businesses were doing and where they were doing it.

When the open web never developed in that region as businesses and users jumped right to the Facebook app on phones, Google was confronted with an existential threat. The resultant Google website product, which has garnered them both basic business listing data and more nuanced information about the businesses there, was an “open web” created in Google’s image.

In this Googly version of the “open web” not only is Google able to control aspects like the speeds of these websites on mobile but is able to understand consumer behaviors when visiting them as well.

No Analytics needed here as Google already controls the whole enchilada. While Bing and Yandex can scrape these sites for inclusion in their indices the bulk of the data benefit accrues to Google and to a lesser extent, the business creating the site.

The impact of these sites on the current ecosystem of websites in the developed world has so far been minimal. These sites are not (yet?) really a replacement for a well done website and they don’t offer anywhere near enough “juice” to compete in the world of SEO.

But these sites do and can rank and they do in fact suffice for many low level uses. In the developed world, as currently designed, these sites function well for any business that doesn’t already have a site, or may have an old non mobile, that don’t have the resources in time or money to build a full blown  site or perceive that the gains are not worth a larger investment.

What is turning out to be a competitive Asian strategy for Google to attract the next billion users (and several millions of businesses) could very well end up being a disruptive force to web development at some point in the future in the developed world as well.

GatherUp Introduces a new website tool – the Conversion Pop-Up

When I helped start GatherUp (previously GetFiveStars), I did so with the idea that local businesses, small and big, deserved world class tools that helped them navigate and benefit from the world of reviews. And I also wanted to be sure that the agencies that helped them could leverage the more advanced features of these tools to help them on that journey.

It has been an exciting as I saw consulting ideas that I struggled to implement for 10s of businesses become a reality for tens of thousands of business locations. As part of that we, at GatherUp, always believed that if we could help local business develop the ability of listening to their customers that they would have achieved a certain “super power” towards becoming the great business that they wanted to be.Very early on we also recognized that these same businesses could use the reviews that they were gathering to share with potential new customers visiting their website or finding them on Google by embedding their own reviews directly onto their website. This helped with social proof, and improved their their seo. It dramatically simplified review schema for them and provided on-going source of fresh content for their website. This feature has gone through several cool revisions with a few more to come. Our Review Widgets  have worked well for businesses with one location or thousands as they communicated their strengths to the world.

Today GatherUp is releasing a new feature, the Conversion Pop-Up, that takes the idea of leveraging your reviews for marketing your business  one step further. The Pop-Up helps businesses share that greatness with the world more easily via a small pop up that can appear on the pages of their website, making hard earned reviews readily apparent as it rotates through them.

This new tool can be easily configured to show on one or many pages, link to a specific page and show first and third party review content. All with a small javascript snippet that works across almost every website type and host.

You can read more about the feature on our blog, watch it at the bottom of this screen or just ping me ( and I will be glad to show it to you.

You can see it in action on the following sites:

Barbara Oliver Jewelry
The Flooring Gallery (white label)

Google Adds Posts to Local Branded Search Results

Google has been expanding the branded local search results. Most recently (last November timeframe) they added an About tab that included corporate information, often a link to wikipedia or the main corporate about us page and social links Today I started seeing the inclusion of a Posts Tab in these results.

Interestingly this is NOT the long awaited multi-location posting product but rather a way to show the Posts that originated from the local agents.

In this spotted example (mobile only) the Post shown under the tab is the post from the #1 listing. None of the other listings have posts so it is still unclear what would show if other locations had posted.

Would it just show posts from the top 3?

Would it dig deeper and show posts from the top 10 listings?

Did the local listing rank #1 because of the post?

Not yet sure.

Note the Posts tab between the Locations and About tabs ABOVE the map.

Heck and who knows maybe this is a precursor the long promised, naught delivered multi-location Post product.

If you see this in the wild and can share examples with more than one post I would love to see them.

Here is what the Post looks like in the search result: Continue reading Google Adds Posts to Local Branded Search Results

Bed & Breakfasts Forced to “Suck Hind Teat” at Google

Tim Capper pointed out a few weeks ago How Google Exploits the Hospitality Industry for Profit.  Bed and Breakfasts are effectively treated as hotels on Google so they have to put up with all of that and then some.

Growing up my father raised dogs. When there was a new litter, the runt was often forced to the rear and maybe not able to get any “love” at all. My father would say: “They are sucking hind teat“. As Wiktionary points out it is a colloquialism that variously means:

  • To feed from an inferior source of food.
  • To be the youngest or most neglected child.
  • To be last in line.

Bed & Breakfast operations are certainly all of that when it comes to Google and the hotel industry.

What got me thinking about this was a forum post that Joy Hawkins pointed out to me where the owners asks:

My GMB page has a blue Book a Room button that takes viewers to 3rd party booking engines. I want people to go to MY website and book, so I don’t have to pay the 15% commission. My website booking engine doesn’t even show up on the list!

How can I have my booking engine show up as top choice – or even better as the only choice?

The short answer; you can’t. At least not easily and certainly not without consequences. The long answer is even worse.

From Tim Capper:

Firstly you will need to contact all the OTAs ( Online Travel Sites ) and ask them to remove your hotel from their platforms.

Then for your own bookings to appear, you will either need to integrate your own booking system with Google Hotels API or use one of the parner booking sites.

Once integrated, you can then use Google Hotel Ads to Bid on your own rooms >> Yes you still have to pay Google to have your own bookings appear.

The thing you need to consider before removing yourself from all the OTA sites, is can you afford to do this – how visible is your website and GMB listing in organic searches. In other words, will people find you without using an OTA ?

What does Tim mean about visibility? Well it turns out that if you manage to get the booking button removed, your listing will literally sink to the bottom of the heap and show much less as the algo is tuned to show those listings with OTA availability.

Here is what Lisa Kolb of Acorn Internet Services, who specializes in the digital marketing for the B & B space, told me:

Regarding this sentence: “The problem is that if you stop working with OTAs your listing shows a whole bunch less in search as the algo is tuned to show availability.”

When a property chooses to NOT provide their availability to any OTA, their competitors who do provide availability typically show up HIGHER, based on the dates selected by the surfer. Thus the Non-Participating OTA property doesn’t show up less, they just show up lower in the Local Listings (typically below those that do have availability for the search dates).

This is such a complex issue for hotels and B&B’s. Their OTA strategy must be highly refined because that OTA reach isn’t just to Google (which is a biggie), it’s to other products such as some B&B directories, and also Trip Advisor (another biggie). And how you make (or don’t make) your property available ultimately filters down to each of these other systems, and can seriously affect your bookings.

Starting a B & B’s is often an act of love and not a totally rational business decisions. They often are shy of resources and have difficulty navigating the vagaries of the online world. Worse, they can’t very well afford to pay the highway ransoms that OTAs charge (25-50% of the booking fee). And yet they are stuck in this unenvious position by Google.

I think that you would agree that in affect they are forced to suck hind teat.

Google Insights – Queries used to find your business

I just saw this Google My Business Insight feature: Queries used to find your businessSend feedback. Per Google these are “The most popular queries for your business by unique users”

I have been on a trip for the past week and less than observant of Insights so this could have shown up at any point and I wouldn’t have known.

That being said it appears that it is a useful upgrade to Insights. They currently are only showing for 1 month and 1 week but they seem to surface the kind of detail that has been harder and harder to find from Google:

Let me know how long you have been seeing this, if you have been seeing this, whether you find the data useful and how you might use it.

Developing Knowledge about Local Search