Get smart with Google smart bidding

by Hailee Sosnowski, 27 Apr 2018

Many years ago, search was much more simplistic than it is today. It involved simply building out your keywords, writing compelling ad copy, pointing to relevant landing pages and pressing play. However, as search engines have gotten smarter, the process has become more intricate — specifically when it comes to bidding. The latest advancement in this arena comes from Google in the form of AdWords Smart Bidding — a portfolio of automated bid strategies that uses machine learning to optimize before every auction. Being the only auction-time bidding currently available to advertisers today, we find ourselves in uncharted but exciting territory!

Before diving into the intricacies of Smart Bidding, it’s important to take a step back and review the differences between bidding options available for search.

  • Manual Bidding is likely the option with which we’re all most familiar. While this tried and true method, if done correctly, has proven successful, it can be an incredibly time consuming process depending on the size of the account. In theory, regardless if an account has 1,000 or 100,000 keywords, someone has to review each and every one of them on a daily basis.
  • Rules-Based Bidding considers bids on a keyword by keyword basis using algorithmic decision making. However, this method still heavily relies on human input as it operates off of defined performance criteria. An example is foldering, in which advertisers must manually group keywords and set goals before a system works its magic. While this provides for more control over bids, brands run the risk of leaving conversions on the table as we can assume that bids are only being adjusted once a day based on historical data.
  • Intra-Day Bidding, often referred to as “real-time bidding,” uses machine learning algorithms to adjust bids throughout the day – usually around four times a day. While Intra-Day Bidding still relies on historical data, it refreshes much more frequently than a traditional Rules-Based system.
  • Auction-Time Bidding, or Smart Bidding, utilizes Google’s machine learning algorithm to optimize bids for each and every auction. More specifically, in addition to the factors available with manual bidding — time of day, remarketing lists, device, location, gender and age — Smart Bidding strategies utilize real-time signals such as a searcher’s operating system, web browser, language settings and actual query to set bids unique to every searcher. By leveraging this data, Google believes it can accurately predict the conversion likelihood of each auction making some searchers more valuable than others via higher or lower bids.

Once it’s clear where Smart Bidding fits into the bid optimization landscape, it’s important to consider the different strategies available to search. These strategies can be broken down into two buckets: awareness based and conversion/revenue based strategies. For now, let’s focus on Smart Bidding’s conversion and revenue-based strategies:

  • Target Cost-Per-Acquisition (CPA): This strategy maximizes conversions while also accounting for your specified Target CPA. For instance, if CPA is under goal, Google’s algorithm will most likely increase bids in an attempt to capture more competitive conversions. It will continue to do so until CPA falls in line with the set target.
  • Maximize Conversions: This is as straightforward as it sounds. By using real-time signals, Google estimates how likely a click is to convert and then raises or lowers the bid. It’s important to note that this strategy will strive to spend your entire daily budget as it seeks out additional conversions.
  • Enhanced Cost-Per-Click (eCPC): This strategy attempts to maximize conversions by raising or lowering bids for each auction based on its predicted conversion rate while doing its best to keep a keyword’s average CPC below its set max CPC. It’s also important to note that eCPC can work in conjunction with third-party automated bidding.
  • Target Return on Ad Spend (ROAS): While this strategy is in many ways similar to Target CPA, it differs in that it optimizes toward revenue instead of conversions. Specifically, AdWords sets bids to maximize revenue within the confines of your predetermined target ROAS.

While it’s easy to go on and on about what Smart Bidding is, you’re probably thinking to yourself, “That’s great, but does it actually work when put to the test?” After doing some testing of our own, one thing has become clear: success is dependent on your goals.

Earlier this year, a leading online retailer came to us with the challenge of increasing revenue by 10% year-over-year — this was no small feat considering their campaigns are already wildly successful. As their goals were already largely ROAS focused, it made sense to employ Smart Bidding’s Target ROAS strategy to maximize revenue. To ensure we gained tangible, head-to-head results, we applied Target ROAS via a Google Experiment to various campaigns with less than average impression share across all match types. The experiment vs.control was split 50/50 to ensure that both were eligible for the same number of auctions. The control utilized a third-party rules-based bid strategy managed via folders and updated once a day, the test utilized Google’s Smart Bidding.

Over the course of two months, Google’s Target ROAS strategy generated 68% more revenue than our control. While this number is undoubtedly impressive, it came with a hefty price tag. Specifically, it spent 96% more than the control. Because our selected campaigns had “unlimited” budgets, Google’s system had room to increase our bids to their maximum potential as it sought to uncover all potential sales. In theory, this additional traffic was more expensive because it was more qualified, and, therefore, more competitive in the auction. This ultimately resulted in more traffic at higher CPCs and a significant increase in spend.

We fully understand that every advertiser does not have a large enough budget to comfortably support a 96% increase in spend. So, what happens if you set more aggressive budget caps? The simple answer is that the system will adjust its approach. Instead of seeking out any and every potential click, Google will limit impressions to only the most valuable searchers — those predetermined likely to convert — in an effort to increase efficiency.

 As Google’s Smart Bidding strategies continue to evolve, it’s important for advertisers to test and determine which are most compatible with their business goals. What works for one business may not work for another. However, as the only real-time bidding options currently available, these strategies have the potential to unlock seemingly hidden conversion and revenue potential.

Want to learn more?

Contact us!