Google Ads Bidding: Is enhanced CPC really working?
In the last year, Google Ads has announced several changes to bidding models, specifically “smart bidding”, in an effort to automate bidding for advertisers. While typically one of the more time consuming aspects of account management, bidding is also one of the most important aspects. New bidding models and updates to existing models have been rolled out to accommodate a variety of advertiser goals, from a target CPA, a target ROAS, or driving quality traffic to the site. They are designed to put more control in to Google’s hands to allow advertisers to focus on other aspects of account management, but are these bid strategies always worth it?
Our client, an online retailer, was starting to see an overall increase in branded CPCs. These campaigns were all set to enhanced CPC bidding.
Enhanced CPC (eCPC) bidding is a form of smart bidding from Google that allows Google to raise or lower your max CPC bid depending on if they believe a conversion will happen with a certain click. Prior to mid-2017, there was a 30% bid cap on eCPC, so Google couldn’t bid any more or less than 30% of the bid that you set. Now that 30% bid cap is no longer in place, the value of the actual max CPC bid set is less meaningful. Google attempts to average out actual CPCs close to what the max CPC bid is set at, but it is not guaranteed.
For the purpose of this test, we tested turning off eCPC in our branded campaigns so we were 100% on manual bidding with the max CPC bid we set. Our hypothesis was that with eCPC enabled, Google was bidding up when not necessary, driving up our actual CPCs. By nature of branded campaigns, regardless of the competition in the marketplace, CPCs should always be relatively low since a very high quality score of 9 or 10 out of 10 is typically achieved. For this client, branded CPCs had increased over 60% from January 2018 to September 2018, going from $0.13 to $0.21.
For this test, we switched over all branded campaigns from eCPC to manual CPC bidding on 9/25/2018. This test was monitored using FT Optimize’s internal A/B testing system, so we were using internal revenue numbers along with Google metrics such as cost and click data to easily measure pre/post performance in an easy-to-read dashboard. The timing for the test was also optimal, because we were able to gather results prior to the peak season of November to mid-December holiday shopping.
Prior to launching the test, we determined that we’d run the test for 14 days to evaluate initial performance. If we saw positive results (decrease in CPCs and even or increased revenue), we would continue running the test. If we saw poorer results (even or increasing CPCs and decreasing revenue), we would look to see if there were any other factors impacting performance and revert back to eCPC bidding if necessary. Of course, we were still monitoring this test daily to ensure no drastic changes were occuring. Additionally, to keep the test pure, we refrained from making any other major changes in the account during this time to isolate the data to this specific test.
In the first 14 days we were seeing positive results. CPCs were down 5.5% and revenue was up 2.5%. With branded revenue already high, the 2.5% increase in revenue drove an extra $1.6k in revenue in the first 14 days. We continued on with the test.
The 30 day results were very similar to the 14 day results, so the test continued on with manual CPC bidding. CPCs were down 6% and revenue was up 2.3%, with cost remaining the same.
As we approached November, seasonality started to take effect and overall volume began to increase in branded searches. In the coming months, CPCs continued to drop, from an average of $0.21 in September down to $0.17 in January 2019. In the same time period the prior year, CPCs had only fluctuated +/- $0.01.
In accounts with $1k in monthly spend or $1MM in monthly spend, every advertising dollar counts. Even in branded campaigns that can be considered “easy wins”, keeping costs low is integral to being cost-effective in driving revenue. In a large account like this one, when daily branded click volume hovers around 2k/day, even a cent or two difference in CPCs can impact efficiencies and daily spend. Being able to decrease CPCs a few cents leading into peak holiday season allowed us to become more cost efficient and save over $5.5k from 9/25/18 - 12/24/18 compared to if CPCs had remained at $0.21 throughout peak.
The success of CPC bidding test in branded was a springboard to testing in other areas of nonbranded search. Using FT Optimize’s unique bidding algorithms to make impactful bidding changes at scale, we’ve seen manual CPC bidding more successful in taking over control of CPCs, performance and account management vs. putting it in the hands of Google’s algorithm. As always, controlled testing with a plan of action and success metrics is key for any marketing plan. Has Google’s smart bidding helped or hurt your account? What bidding strategies have you seen successful?