Android 6 “Marshmallow” & SEO Series: Sponsored Google Now Cards


It is unclear exactly how the user tracking works, but it is likely just based on a user’s web browsing history. This Sponsored Google Now card could have been triggered in a more conventional way with cookies stored from Google Shopping, but that information would be harder to access across multiple devices.

(Something like this could theoretically even be triggered by product schema on the page, though it could not be found when I looked through the source code.)

From an SEO perspective, this Google Now card is interesting for a couple reasons. First, because it is one of the first times we have seen the ability for a company to pay to generate a Google Now card. While this isn’t a tit-for-tat payment (because the card was based on previous browsing history and was only triggered when the price dropped), it is still apparently part of the Google Shopping experience in which sellers pay to participate.

It is clear that Google is trying to build out this kind of predictive search functionality with Android Marshmallow and Now on Tap, so leveraging Google Shopping feeds and this Sponsored Price Drop notification could be great for driving site traffic and conversions — and important for SEOs to at least be aware of, even if it does not fit squarely into the “organic” side of digital marketing.

The inclusion of “Visually Similar Products” in the landing page (aka “related products”) is also very interesting.

GoogleNow Shopping Price Drop Notification
picture: cheap formal dresses online

What is noteworthy about this aspect of this interaction is how Google is determining the “Visually Similar Products” to include. It appears to be done with the same matching learning algorithm that is being used on Google Photos to categorize and group photos (like the example from the previous article in this series, which showed a search for “my pictures of dogs”).

Remember, the machine learning algorithm automatically groups the photos based on the date and time, as well as location — all of which it can get from the metadata in the image file. But it is also uses image recognition and machine learning to group the photos based on what is featured in the picture, as you can see in the image below.

With that in mind, looking at the two groups of “Visually Similar Products” below, you might notice a connection: Both searches started with a text query for “blue dress.”

The click in the first example was on a “blue dress” for which the main photo featured a mannequin, so all of the “Visually Similar Products” show dresses on mannequins. In the second example, the initial search was still for “blue dress,” but the click was on a product whose primary image featured a human model, so all of the “Visually Similar Products” are on human models.

In terms of SEO strategy, this could have some very real implications for the product images you are using, especially if the products you are selling look a lot like products or brands that get more searches than yours do.

Aligning your product photos with the product photos of trendy or high-demand items could move your products into a shopper’s consideration set much more quickly — especially in the case of soft goods, where fit and finish are harder to determine online and there are no regimented technical specifications or guarantees to compare.

It is also useful to note that the inclusion of “Visually Similar Products” could limit the marketing value of a Sponsored Google Now card, if competitive items are available and ranking in these results. A user who responds to a price drop notification would probably also be persuaded by a similar item for a lower price — even if it is from a different retailer.

For example, the $9.99 dress from ModDeals looks nearly identical to the $13.98 dress from Target, and it could easily be purchased instead with only one additional click in the interface.

To combat this kind of customer loss, free shipping deals with branding and loyalty programs could become increasingly important, since the only thing keeping a user from shopping with another retailer may be the final price or their level of confidence in your brand and services.

The future of Sponsored Google Now cards is still up in the air, but it seems very likely that Google will include some type of Sponsored result in the Google Now interface (if not one, then possibly many). Google’s basic business model relies on advertising dollars, and search traffic lost to Google Now, compounded by the threat of Amazon Prime, makes this an obvious next step.

E-Commerce SEOs who have never considered a Google Shopping feed may have to include it in their strategies or begin working with other teams to ensure proper integration and exposure in Google Now. Beyond that, with or without Google Now, optimizing for image recognition algorithms appears to be an entirely new but relatively cheap way to drive online sales — with concepts that are similar to SEO, but based on shapes, colors and patterns.

SEO is changing, and the mobile platform seems to be accelerating these changes, so it is exciting to see what new skills top-notch SEOs will need to compete in the future.

read more: