Machine Learning and An Amazon Take Down?

Marketing - 11 Jun 2019

This years’ Google Marketing Live did not disappoint! New ad types, bidding strategies and a massive Google Shopping announcement were the highlights for launch later this year.

Google prides itself on building habit forming products. Take Gmail for example. Gmail fully rolled out “smart replies” to its 1.5 Billion users just under a year ago and they now account for 10% of all emails sent. Smart Replies scan your email content and write 3 short responses you may find appropriate to send, saving you time. This is hugely significant, showing how quick consumers are to adopt machine learned features. Machine learning was a key component of this years marketing updates.


Ailsa Duncan

About the author

Ailsa is a Search Manager integrating paid and organic search to improve performance in both. She develops and delivers client search strategies and ensures we are always ahead of the latest search trends.

Discovery Ads

The biggest Google Ads announcement and new product came in the form of Discovery Ads. A brand new placement for marketers to reach their audience. The mobile app remaned it’s Google Feed to Google Discover in September last year to better reflect its function. This section curates content it deems relevant to you based on your search activity, allowing you to further explore topics of interest.
Google found that 76% of consumers enjoy making unexpected discoveries when shopping, so ads in this space are expected to have a high interaction rate. Discovery Ads will be introduced later this year with a focus on rich and personalised ads with interactive formats. TechStyle were one of the first to test out Discovery Ads and found on average a 25% reduction in cost per lead!

Gallery Ads

75% of mobile users expect immediate information. This statistic led Google to develop Gallery Ads which combine intent of search with a more interactive and visual ad format. Appearing in the main Google search feed this is similar to the swipable Facebook carousel ad for those familiar. They can contain 4-8 images with a 70-character tagline under each image. Creation of the ad will allow you to provide 3 headlines, so you can test your success with different call to actions.
This may look like “just another ad format” but the impact of this is much more significant. These ads provide the customer with far more information in the initial part of their purchase journey, allowing them to make more decisions without exiting Google. Fewer migrations to sites will be necessary, reducing the number of touch points. Whilst this will decrease campaign traffic overall, the traffic you do get will be more likely to convert.
What was not discussed at the event was how each interaction will be reported on. It will be interesting to see if a swipe will be counted as an impression or if it will be monetised and how this will be measured in comparison to a click. We may even see new reporting metrics specifically designed for Gallery Ads based on how many images the customer swipes.

An enhanced feature coming to Google Ads is deeplinking. This should end the frustration of being directed to a retailer’s website from an ad click when you already have the app downloaded. This will also be fully reportable, so you can understand your customers journey across mobile web and apps to further optimise your campaigns. App events and conversions will also be trackable in Google Analytics, an exciting introduction which I’d expect will lead to increased Google Ads conversions.

Smart Bidding

Machine learning has been the big talk of recent years. It harnesses the power of data, finding trends in historic activity and becomes reactive to the decision-making process.
Smart bidding has been around for a while (actually as far as 2010!) The most recent Marketing Live update brings us the new bidding strategy of “Maximise Conversion Value”. This allows you to optimise a campaign towards the particular conversion which means the most to you. Most accounts and campaigns will track different goals such as calls, purchases and store visits. The functionality will now let you prioritise real world purchases from your stores for the first time ever!
The Maximise Conversion Value bidding strategy will also introduce conversion sets. You will be able to create rules to adjust the value of your conversions depending on the audience. A conversion from a returning customer may mean more to you than a new customer and this will all be trackable very shortly.
Smart bidding strategies will also be able to factor in seasonality changes such as summer sales or Black Friday. Simply add your dates for the bids and conversion rates to be adjusted in anticipation of higher than average traffic.


Bumper Machine

Video ads did not miss out when it came to the 2019 shake up. The 6 second uskippable Bumper Ads designed to complement TrueView campaigns were rolled out in April 2016. These can prove costly with extra production costs to condense existing promotional videos. Google’s answer? Bumper Machine. This new feature in Google Ads will take your current long form video ad and generate 3-4 different 6 second ads with no extra cost. Some editing functionality will still give you control over the final creative.
Another example of how Google is employing machine learning throughout its product offering. Bumper Machine will “watch” your existing video and pull your brand elements, faces and take-home messages to create the Bumper ads. Not only does this speed up your production time, it could also show you a new creative direction. For those who make heavy use of video ads, this will be a valuable tool.

Custom Audiences

Audiences have always been a useful way of finding new and relevant customers for your campaigns. The update to this feature is to merge Custom Affinity and Custom Intent audiences creating simply Custom Audiences. The new functionality will give you insights and impression forecasting to give you an idea of your audience’s worth. In addition to this merge, the Audience Expansion Tool is being created for you to find even more people, with early reports showing advertisers have seen up to 50% more conversions.

Finally, Google pretty much announced its intent to take on Amazon. You will soon be able to buy products straight from Google Shopping without going to the retailer’s website. This will then extend outwith Shopping where customers can purchase in Google Search and Images where relevant.


The natural progression is then to YouTube, allowing you to buy products featured in a video whilst it continues to play. I can see this proving popular with low cost items and particularly clothing brands where consumers can make one-click impulse buys.
Google is further strengthening its shopping capabilities with the introduction of “Shopping campaigns with Partners”. This allows brands to jointly fund their advertising using additional budget from retail partners. So, Adidas could accept budget from JD Sports to boost their campaigns (not that they would need it, but you get the picture). This gives smaller brands a bigger platform and sales opportunity.
Machine learning is a recurring theme throughout many of the updates from Google Marketing Live. Designed to save marketers time on manual tasks so we can spend more time on strategic and what we do best: marketing.
There is also a clear trend for Google to keep consumers on Google, minimising the number of touch points in the purchase journey with Gallery Ads and soon leading to the actual transaction completed on Google itself. It will be interesting to see how Google Shopping evolves over the next few years and if it really can battle Amazon in the shopping stakes.

Are you interested in finding out more about how we are implementing these new techniques?

Get in touch