1300 490 147

What I Learned at the Google Machine Learning Masterclass

What happens when you send your sales guy into Google’s Head Office for a Search Engine Master Class on Machine Learning??

Let’s find out…

If you’re a newcomer to Digital Marketing you might be surprised to learn just quite how data-heavy and analytics-driven, the world of Digital Marketing is. The nice thing about machine learning; apart from the fact that their “deep minds” can compute far more data and analyse it much faster than multiple human brains can. Is the fact that the AI algorithms will perform many of these tasks for you; or as much of it as you would like.

Whether you’re new to Digital Marketing or an online veteran; Google’s focus, drive and energy, at least for the foreseeable future is AI and Machine Learning and if you take nothing more away from this post its this:

Machine Learning is here to stay.

As has become the standard, Google had broken everything down into easily digestible, bite-sized bits and chewable chunks of information that our mere human brains can handle.

The key objectives from the advanced training, and now my article are outlined below;

  • Simplify the Complex
  • Leapfrog the market with machine Learning
    Develop an action plan

According to the Googler’s if your Ad Account can master the implementation of Machine Learning Ad tools you could achieve a 27% increase in sales.

Who wouldn’t want a 27% increase in sales?

The big question kick-starting proceedings was; what is the difference between Google Advertisers accounts that perform vs those that don’t?

The answer as you may have guessed – Machine Learning!

Apparently, the Googler’s did a deep dive into Ad account performance globally, and the champion accounts had one stand out factor – They all made use of machine learning add-ons and products.

Google’s Deep Mind

Still not convinced about the capabilities of machine learning see the clip below where Google’s Deep Mind, successfully beats the world champion of the board game Go! Most impressive of all is that in an interview afterwards champion Ke Jie said that; the computer had earlier performed a move that seemed strange and potentially incorrect but later turned out to be a winning move and opened up new strategies and plays, that had previously never been used or thought of by human players.

(If you want to understand how to implement some of the machine learning tools this will be explained in detail in a separate post).

None of the AI Tool should be implemented until you have a solid base in your Google Ads account, so if you need a recap to check out our guide.

If your Google Ads account is compared to building a house your foundations are still conversions, ad extensions and your budget. You ideally want larger, wider and bigger where possible.

Your attribution model can be likened to bricks and paving; second phase building, a guideline or outline effectively. The roof would be the machine learning factor; an add on fitting in on top of an already solid structure but providing more stability, protection and versatility.

The two broad machine learning categories are called Smart Bidding and Smart Creative and I’ll dip into each one of them below.

What is Smart Bidding?

Basically, automated machine learning enhanced keyword bidding strategies, that are optimised for conversions or conversion value.

For those that just went, “huh” – It means there is a clever system at Google which helps you get the best bang for your buck when bidding on keywords for your advertisements. So clever in fact that it takes into account all the parameters you have outlined in your attribution model and a few others that you haven’t.

Smart Creative in the form of Responsive Search Ads is next up on the machine learning agenda. Read on to understand what they are and how they work…

In a traditional Google Ad campaign, one would have used one heading and descriptions section together creating one solid piece of content. With Responsive Search Ads you can use up to 15 different headlines and 4 different descriptions. The machine learning capabilities then test all the different combinations based on the best performance. Once that is worked out your Responsive Search Ads should overtime; be able to send the best performing ad copy, to the most relevant person, at the most appropriate time whilst considering factors such as device, operating system, location and history. Pretty Smart eh?

Add it to your existing ad groups and generate 10% more clicks and best of all you can still control the content.

Dynamic Display Ads

The other tool much talked up by our Googlers is the Smart Creative Ad tool – Dynamic display ads. This is a feature within the Google display network that allows a website to dynamically display advertisements to users, based on the products they have previously browsed online, or based on the content in pages on websites. The way to look at this feature is; that you’re effectively casting your net wider on customers searches, as well as pages on your site they have spent time on. With Dynamic Display ads turned on you could capture incremental traffic without changing or adding keywords and in time lowering your CPA by up to 10%.

In conclusion

my hope is that I have, ‘Simplified The Complex’ world of machine learning and broken down the two major facets that relate to Ad Account performance. You’ll now have a competitive edge over your competitors which allows you to ‘Leapfrog The Market’. It’s pretty clear that machine learning features should be implemented into an existing Ad account strategy. Whatever metric you’re trying to achieve or surpass, with 15% of all search being new these AI tools can only assist in helping you deliver results to your business or clients. The only caveat is that I’m leaving the ‘Action Plan’ up to you!

Ilan Shabad
Ilan Shabad
Ilan is the Co-Founder & Head of the SEO team at One Egg Digital. He has over ten years’ experience in delivering business growth across start-ups to multi-national enterprise-level clients.

Get advanced digital marketing tips we don't make public

Join subscribers from companies like Aventus, LegalVision, PsyQuation & Getac.

Related articles