Google’s bidding algorithm has developed and increased in complexity over the years. In the early 2000s, Google AdWords was powered by a ‘simple’ rule-based algorithm. Nowadays, it is powered by machine learning and can take into account hundreds (if not thousands) of data points.
People from all over the world use search engines to find the information they are looking for, learn about topics of interest, and make important decisions. So, Google is trying to show them the most relevant and reliable information available.
And they take the same approach with the ads they show for each search query.
Google Ads auction
A lot happens in the background whenever a user searches on Google. Besides pulling the most relevant result from its index to show the organic results, Google also decides which ads to show.
And for this, it has an auction-based system in place. But it isn’t a traditional auction where the highest bid wins. This would make Google a lot of money, but will not be helpful for the users which might turn them away. And remember: Google’s product is the user base. Without a user base, they would not be able to sell ad placements to advertisers.
So, Google has come up with an auction model that includes markers for relevancy. These are compiled into one metric called the Ad Rank. At a high level, the Ad Rank score is based on 6 factors:
- Your bid
- The quality of your ads and landing page
- The expected impact from your ad assets and other ad formats
- The competitiveness of an auction
- The ad rank thresholds
- The context of the user search
How Google’s algorithm looks at your ads
If you want your ad to enter the auction and show your ad in the search results, you must focus on improving your ad rank. However, from those factors, there are only a few you can influence. So let’s break them down and see if and how you can improve them.
The keyword bid
Nowadays, if you are running Google Ads campaigns, you are most likely using an automated bidding strategy. So you shouldn’t worry about the bid amount.
The quality of your ads and landing page
This is something you can control. You must make sure that the search intent of the keywords you are targeting aligns with the message of your ad copy and the landing page they lead to. If you want to assess the quality of your ads, you can use the metrics Google provides in the Google Ads account.
The expected impact from your ad assets and other ad formats
With Search ads, besides the headlines and descriptions, you can add extra information in the form of ad assets. Think of these as add-ons for your ads. You can add sitelink extensions, call extensions, price extensions, image extensions, etc.
While you can’t estimate the expected impact of adding them, the best thing to do is to use as many of them as possible. Besides the benefits they provide to your Ad Rank, your ads will be more prominent and take more space in the search results.
The competitiveness of an auction
This is clearly something you can’t control. So the best way of tackling this is to do your best to craft irresistible ads and continuously look for new keyword opportunities.
The Ad Rank thresholds
In the auction, Google uses a minimum Ad Rank requirement which determines if your ad will participate in the auction or not. So if your ads don’t meet the threshold, you need to improve your ads to have a chance in the auction.
The context of the user’s search
In deciding which ads are best for a user’s search, Google will look at the context of the search. This means that it will check the search term, the user’s location, the type of device they are using, the time of search, etc.
Google Ads bidding algorithm details
With the auction fundamentals explained, it is time to see how the Google Ads bidding algorithm uses this information in the automated bidding process.
Manual CPC
With a manual CPC strategy advertisers were setting a maximum bid they were willing to spend on a click. And then Google would take that information and use it in the bidding auction.
Note that a bid of $3 per click doesn’t mean you would pay $3 for every click. It is the maximum amount you are willing to pay. The actual click cost can be lower depending on the Ad Rank and the competitor’s bids.
But this was like looking at a movie poster. You only get some context of what the movie is about, the title, and the main actors, but you won’t understand the whole story.
Smart bidding
So Google introduced Smart bidding. This is leveraging all the data Google has from users and advertisers and uses a feature called ‘auction-time bidding’.
Smart bidding uses goals to set the bid for each auction. If it is more likely to achieve the goal with that particular click, it will bid more for it.
But how does Google measure that probability? Well, with a nightmarish level of user data.
Think about it. When a user is looking up something, Google knows:
- the search term
- demographics (age, gender, parental status, and income)
- the location
- the device they’re using
- the operating system
- browser language settings
- the previously visited websites
- the things they buy
- they’re favorite shops (especially if they have loyalty cards in Google Pay)
- their habits, hobbies, and interests (based on YouTube and Google Searches)
- the comments they left on YouTube
- the apps they use (if they’re using Android)
And all of this data is combined with the data you provide about your users:
- conversion tracking data
- customer lists
- remarketing lists
- the target audiences and interests
- the way they previously interacted with your website
So when you give the algorithm a goal, let’s say a ROAS target to hit, it will start looking for users searching for the keywords you are targeting.
Leveraging all this data, Google’s bidding algorithm can understand each user’s search intent. It can estimate quite accurately where they are in the buyer’s journey and act accordingly.
Let’s take two examples:
- A user is interested in cooking, frequently visits recipe blogs, is subscribed to various cooking channels on YouTube, and searches on Google ‘what makes a good knife’
- A user is interested in cooking, frequently visits recipe blogs, is subscribed to various cooking channels on YouTube, visits your knives store, reads a few blog articles about different types of knives and different types of steel knives are made of, and searches on Google ‘best Chinese Chef’s knife for home cooks’.
In the first example, the bidding algorithm will have lower confidence in achieving a conversion, so it will place a lower bid. But in the second example, it knows the probability of that user buying a chef’s knife from your store is higher. And the bid it places will reflect that.
Identify optimization opportunities
With all this automation built into the campaigns, it can be hard to spot ‘traditional’ ways of optimizing your campaigns. So advertisers will have to think more strategically about the campaigns and think of more high-level ways to steer the algorithm toward their business goals.
Have a simple campaign structure
The days of overly complicated structures, when you had a single keyword for each ad group (SKAGs), are over. Google’s bidding algorithm needs as much data as possible to make the correct decisions.
Opting for fragmented campaigns and ad groups today will only hurt your performance.
Imagine that Google is a book publisher who knows the readers really well. You want to publish a book, but instead of submitting your whole manuscript, you only send one page per day. Can the publisher really know if your book matches their readers? And how long will it take to reach a conclusion?
Get more cold traffic
Google will always try to achieve the goal you set for the campaign. This can be the target ROAS, target CPA, Maximize conversions, or Maximize converted value. The easiest way of doing this will be to go for warm traffic because it has a higher probability of converting.
Knowing that Google will have a bias toward users further down in the funnel, you can counter this with negative keywords and audience exclusions.
For example, you can exclude all brand-related keywords from your cold traffic/generic keywords campaigns. Or you can upload a customer list and use it to exclude those users from your campaign targeting.
Think about the user’s search intent
Try to group your keywords based on a theme, but also on the search intent. Don’t mix upper funnel related keywords with lower funnel related keywords in the same ad group.
Use negative keywords
Don’t forget to use negative keywords to let Google’s bidding algorithm what’s not important or useful for your business.
Exclude any keywords that might indicate user intent that you don’t want to target. Here are some examples:
| bargain | deal of the day | free | low price |
| career | deal of the week | inexpensive | review |
| cheap | discount | internship | salary |
| clearance | discount code | job | sample |
| deal | discounted | low cost | vacancy |
Use Marketing Efficiency Ratio
If possible, look at the Marketing Efficiency Ratio (MER) instead of ROAS. This will give you a broader understanding of your performance.
This is not a Google specific suggestion, but it can help you steer the Google Ads algorithm towards better results.
MER will put your ad spend from all platforms into context, allowing you to make decisions based on a broader picture.
