MACHINE LEARNING · 10-MINUTE READ · By Tina Arnoldi on March 26 2018
Artificial intelligence (AI) and machine learning are new-ish buzzwords. AI is the science, the ability for computers to get smart on their own, doing things that require human intelligence. Machine learning is taking available data and learning through algorithms. And it’s incredible how fast things are moving in this space. Computational power has increased exponentially; presenting new opportunities we could not have imagined years ago. Our phones today do more than our computers once did.
If you are not one of the 39 million who now own a smart speaker, you are at least familiar with devices, such as Amazon Echo or Google Home, that exist because of AI and machine learning. From virtual assistance to online recommendations, we are impacted by machine learning every day. And it is not only our day-to-day consumer lives that are affected, but also how we work as marketers.
What this means for marketers
We have tools such as phrasee to generate the subject lines of emails and Acrolinx to help with content creation. Email providers tell us the best time to send an email campaign. These tools return much quicker results than we could do on our own, which does make some people nervous. (I’ll talk later about what these changes mean for our marketing jobs.)
What this means in the land of Google
Machine learning improves consumer experiences across the Google product line, including Google Clips, Photos, Maps, and YouTube. Google Clips demonstrates machine learning with a camera, “deciding” what to pay attention to based on how interesting an image may be perceived. With Google Photos, there is a Shared Libraries feature, for sharing family or team photos. It learns to save photos of people to a library based on recognition technology. And we’ve all seen recommended videos based on our past behavior when we visit YouTube.
What this means for PPC
Look at what we can already do in AdWords that alleviates some of that manual work. For example, portfolio bid strategies automate bid adjustments without the need to log into AdWords for constant manual changes. What is learned in one ad group with a bid strategy can apply to others in the campaign.
With automated rules, we have options to better manage our accounts: decreasing bids when the Quality Score falls below a certain number, adjusting bids if the CTR is below a KPI, or pausing ads when they do not provide a return within our benchmarks. When you set your success parameters before you start a PPC campaign, these rules can automate some of your ‘to-dos’, allowing more time for strategic work around a campaign. Below are other things Google offers that already utilize machine learning.
These are people Google finds who share interests with the existing audience on your remarketing list using a machine-learning algorithm. And it finds the people when they are actively searching for what you offer, helping you nurture new customers in close to real-time.
Smart bidding decides what to pay for each step along the purchase path with different strategies for conversions versus awareness. It identifies moment by moment trends throughout the conversion funnel in each auction.
These bid adjustments go beyond using device, location, and time signals. That no longer seems to be enough as even more signals are available.
A more advanced attribution model, it uses machine learning to determine how much credit to give each click in the user journal. This attribution model in AdWords is not available to everyone; only those with at least 15,000 clicks and 600 conversions over 30 days. However, if it is available in your account, Google strongly recommends using it.
In-Market for Search Ads
We have been waiting for In-Market for Search Ads ever since the announcement in May of 2017. It’s now expected for the end of 2018. With it, advertisers can segment users whose queries and online behavior show they are looking to buy in the near future while they are on the Search Network. It separates intent to buy versus those who are browsing.
AI Driven Ad units
Google recently announced “Auto Ads” for AdSense. “Auto ads use machine learning to make smart placement and monetization decisions on your behalf, saving you time. Place one piece of code just once to all of your pages, and let Google take care of the rest.” Publishers decide the general types of ads they will allow on their websites for AdSense and Google does the rest.
Universal App Campaigns
Universal App Campaigns (UAP) run across Google properties with automated targeted and bidding; simplifying the process. Only a few elements are needed for a UAP to run. Using machine learning, Google looks where to serve these campaigns based on a predetermined cost-per-install.
What this means for our jobs
Yes, they will change. And yes, we will still have jobs. But they will look different in the future.
The days of manual bids may soon be over with the advancements offered by machine learning. To secure our jobs, we need to set ourselves apart as strategists and develop deeper understanding of human behavior.
What humans are capable of doing better than machines is interpreting emotions and logic in another person – assuming an individual has soft skills and the emotional intelligence to qualify as capable. Since machine learning depends on reliable and predictable patterns and humans are not always reliable and predictable, that’s where machines are ahead of us. We can teach computers to do the mundane tasks that they can do better that we can, but we still need to analyze and apply data in our specific business. Computers also don’t have context about whether data is truly meaningful and findings are often based on correlation – not causation – which means we are still needed.
The advancements in machine learning are a positive change for people who are prepared. The automation already available in tools, such as AdWords, frees up our time to be creative and provide more value to clients.
We can also take advantage of automation with Supermetrics templates that let us easily pull data for review and recommendations. Cross channel performance can be analyzed with the Data Studio connectors so we can view AdWords compared to our other channels. So don’t panic. Be aware of the changes ahead and maybe even a little excited about what’s to come. Who doesn’t love the idea of automating the mundane parts of digital marketing!
About Tina Arnoldi
Tina Arnoldi is Analytics and AdWords Qualified and one of the few people in the United States recognized as a Google Developer Expert(GDE) for marketing. Her agency, 360 Internet Strategy, is also a Google Partner. You can learn more about her on LinkedIn