Estimated Reading Time 11 minutes
How Google Smart Bidding uses Artificial Intelligence to drive successful PPC strategies?
The rise of machines and machine based learning is nothing new. What does the increase mean for us? Can Artificial Intelligence (AI) drive successful PPC strategies? More specifically, using Google’s smart bidding machine learning algorithms.
First, let us understand a few definitions:
What is Artificial Intelligence? (AI)
Artificial Intelligence is a huge concept in computer science, where machines are able to carry out tasks in a creative and intelligent way which mimics humans abilities.
What is Machine Learning? (ML)
Machine learning is a subsection of AI which trains a machine how to learn and uses patterns and data to automatically learn and draw conclusions, without specific programming.
What is Google’s Smart Bidding?
Some of you may already know what smart bidding is but for those of you who are new to this topic or would like to know a little bit more – smart bidding is an automated bid strategy which uses machine learning capabilities to optimise for conversions within your PPC campaigns.
Smart bidding gives digital marketers more time to manage other tasks and develop strategies whilst Google’s algorithms learn and develop across campaigns, whilst manually adjusting keyword bids.
How Smart Bidding Works:
Smart bidding incorporates three systems to work:
- True Auction Time Bidding: Bidding automatically changes for every single individual auction, which happens thousands of times a second and is based on relevant contextual signals at the present time
- Adaptive Learning at the Query Level: (Unique to smart bidding). Algorithms have moved on from learning at the keyword level to learning at the search query level, which provides richer data for the algorithm to work
- Richer User Signals: Smart bidding uses contextual signals and users operating systems like time of day, re-marketing lists, devices, location (etc) to determine optimal bids
Now we know a little bit more about AI, ML, and how smart bidding works – here are three reasons how these systems can drive successful PPC strategies.
1- Speed up the Process
One way smart bidding algorithms can drive a successful PPC campaign is its advantage of speed. Jeff Baymn wrote an interesting article in the White Papers discussing the future of PPC and how this industry lies in machine learning – he titled the paper ‘Adapt or Be Left Behind’. Controversial, I know?
One of the points I want to pick up from his paper is where he mentions ‘machines can crunch data, spot trends, and execute routine tactical tasks faster than we ever could’. These algorithms can analyse thousands of signals to help understand a user journey and intent – using operating systems such as web browser, time of day, locations settings; algorithms can understand the data of the audience and their potential to convert (remember – smart bidding is all about conversions!) faster than we could. Bearing in mind, these signals are analysed in real time, therefore the system will optimise bids based on all these contextual signals and place ads in auctions which are likely to convert.
2- Contextual Signals
This ties in nicely with the advantage of using contextual signals to support your PPC campaigns. I mentioned these signals in the previous point, but as an account manager, if you had access to more insightful details about your potential user, you would have a better chance at targeting and capturing that user attention. These signals are built up of;
- User Behaviour – time of day, day of week, interests, cross-device usage, previous sites visits
- Demographics – age, gender, device type, physical location, language, browser or the operating system
- Ad Characteristics – ad format and ad performance
- Onsite Behaviour – how many pages did a user view, when a user left your site, value of the products a user has viewed on your site
3- Automated Bidding to Reach Your Goal
Keeping in mind that smart bidding is focused on achieving conversions, not clicks or impressions – automated bidding allows you to choose a bid strategy in order to achieve specific goals. In other words, you can select your bidding strategy, choose your goal and let machine learning tailor the right bid to each and every auction, removing the ‘heavy-lifting’ from your busy routine and giving you time to work on other tasks.
These are the types of bid strategies:
- Target CPA (cost per acquisition) – sets bids to help gain as many conversions as possible at the target CPA you set
- Target ROAS (return on ad spend) – sets bids to help to get as much conversion value as possible at the target return on ad spend that you set
- Maximise conversions – sets bid to help gain the most conversions whilst spending your budget efficiently, within your daily budget
- eCPC (enhanced CPC) – adjusts your manual bids to help you get more conversions, while trying to achieve the same cost per conversion
From my experience and general conversations with peers and colleagues, I have found that people still have mixed feelings about using Google’s smart bidding within their PPC campaigns. Mainly because digital marketers feel as if they have less control over their campaigns and it is a huge step to go from your comfort of manual bidding to trusting an automated machine bidding algorithm.
From our own experience, we’ve tested smart bidding on several campaigns and found that Google will push bids over hundreds of pounds, which lets face it is never sustainable. I think Google sometimes forgets that it’s never an easy conversation explaining that we are testing a new system (which may or may not give results) and yet, we are spending hundreds of pounds with possibly no results. We also found that if you revert the strategy, you have to keep a close eye on max CPC’s as we have seen huge increases in prices. Let’s not forget Google makes a massive amount of money from this and although smart bidding could work, we feel you may have to tackle a long journey before seeing the results you want.
To recap I have discussed three advantages to using machine learning through faster processes, contextual signals and automated bidding to reach your goals; all of which in their own categories are beneficial to you as a marketer. I fully understand how AI can be a step in the right direction for PPC, and of course, some account managers might embrace the idea of switching from manual bidding to automated and allow a machine to help and support them with their campaigns. For others, myself included, I feel there is still a lot more to learn and maybe it’s the ‘old school’ part of me that doesn’t feel quite ready to adopt automated bidding and essentially trust a machine to take over campaigns. Perhaps my obsessive compulsive order to control won’t let me let go of manual bidding.
If you would like to learn more about smart bidding, AI or ML or would like to discuss anything above – please contact the team on the details below.