Google Smart bidding: How To Train Your Algorithm – Kian Njie

08 April 2019

Posted in: BSEO

In an ever-changing landscape of outreach and advertising on Google, Kian Njie, PPC Executive at Inchora, walked us through Google’s Smart Bidding technology and how PPC professionals across the world can use this to further their clients’ interests.

Overview

What was the talk about?

So what is Google Smart Bidding (GSB) and how is it different to the approaches that PPC professionals have been using over the last few years? Simply put, GSB is a platform that uses advanced machine learning techniques to train bidding algorithms to perform better through acting on certain signals. The algorithm will pair the best combination of adverts and customer queries, automatically adjusting the bids and serving the best-performing ads. Essentially, it does a lot of the heavy lifting for you.

Kian touched on how advertising with Google has changed. The first clue is the fact that it is no longer called Google Adwords: it’s now just Google Ads.

The major benefits of GSB versus manual application are that you are able to draw more heavily from Google’s mountain of data (it’s a substantial mountain) plus cut back on the time and energy it takes to build and manage a campaign. Work smart, not hard.

Of course, while it sounds good to having ML applications run your campaigns, you need to justify it with results. This is why it’s so important to have the correct structures in place to measure a campaign’s performance. Kian advises uses Time Decay or Linear Regression models. In an ideal world you want to be utilising a data-driven model, but the only catch with this is that it requires click-through traffic of 15k+ per month which doesn’t make it attainable for many small brands. If you’re a big brand, fill your boots!

It’s important to remember that all things take time, and GSB is no exception. It’s advised that you need to give the technology at least 8 weeks to get to grips with your campaign. This time accounts for planning, the actual learning from the system, and testing. Kian advised steering clear of using Single Keyword Adgroups (SKAGs) as they might confuse the learning process.

In summary, GSB is a powerful and highly-effective tool, but it order to realise its full potential, you need to be able to give it some time and ensure that you are able to test it rigorously.

Potential impact on the industry

GSB is going to allow for the application of much smarter bidding with the end goal (hopefully) of using client advertising dollars better while further increasing the reach to relevant parties.

Beyond this, the hope is that it will allow campaign managers to focus their time and attention on improving other areas of their service offerings.

Key takeaways

  • Pick a realistic goal (+/- 15% of current).
  • Always use portfolio bidding strategies.
  • Try & follow the testing framework .
  • Review your measurement tools.
  • Transition away from SKAGs.
  • Trial setting ad rotation to optimise.
  • Limit KPI changes to +/- 10% jumps.
  • Never set and forget.
Share This