July 19, 2021
iOS 14 Update: Facebook’s Data Challenges Show the Need for Alternative Measurement in the Privacy Era
If you are one of the advertisers who woke up on Wednesday and saw this new notification in your Google Ads account, welcome to the party!
Google Ads has sent a notification to all its advertisers about the change to search terms report
On September 1st, Google started limiting the number of unique search terms it will show in the Search Terms report and will show only the ones that have been searched by “a significant number of users.” It is unclear so far what the word “significant” means in terms of the actual number of users or searches, but official documentation indicates that searches with clicks will be removed from the report. API reporting is impacted too, so it’s not limited only to the Google Ads UI reports.
It means that there will be some completely irrelevant searches you’ll be paying for, but won’t be able to exclude, since there won’t be any direct way to identify them.
Search query reports in Google Ads have never contained all searches your ads were served on, or even every search that led to a click (albeit a very small amount). However, this approach is materially different, because Google is telling us that they intentionally won’t be reporting on specific searches. Google points to data privacy as the motivation behind this change, but as it’s one of the most useful and popular reports among search marketers, that reasoning isn’t going over well.
The impact is pretty shocking. Based on an analysis of our clients, provided below, we’re now seeing 19% of clicks and 31% of cost missing from the Search Terms report starting abruptly on September 1st. Which, according to Google’s explanation, means that number of clicks did not meet their definition of “significant” searches and were somehow more likely to be at risk of exposing sensitive user information.
The below chart shows recent trends between Google Ads’ Search Term report and actual campaign level clicks and cost to identify the percentage of missing cost. We are looking only at Search campaigns and excluding search partners completely.
Looking at last week, our search term report aligned with the total account spend almost identically. In the last two days, on average, we are seeing 21% fewer terms in our search query reports while spend increased compared to our daily average during the last week of August. This will almost assuredly vary by account, but it provides an indicator of just what was lost and the financial impact you might expect.
Even more troublesome, the missing cost has so far well exceeded the missing clicks, which means that the search terms no longer being reported on were more expensive than the avg. CPC our clients are paying at-large. Taking that a step further, on a well-managed SEM campaign, average CPC is typically going to be higher for more valuable searches – those that lead to more and better customers. It goes to show the potential power of that insight that we will all seemingly lose moving forward.
The silver lining, if you can call it that, is that historical reports are not impacted yet, so data prior to September 1st is still complete in Google’s reports. If this data is important to your business, we would advise pulling reports now in the event that this changes.
Search term data is usually used for keyword mining and to find new negative keyword exclusions to prevent serving ads against irrelevant or low-value searches. At Metric Theory, we are often using n-gram models to identify specific words and phrases that are performing well or poorly. Reducing the number of searches in the report by this amount will undoubtedly have a negative effect on these types of analyses.
At the level of change we’re observing, it could have a pretty drastic impact on long-term optimizations and thus the performance of the account. Especially where an advertiser can’t rely on automated bidding (for example, niche B2B advertisers who generate a limited number of very valuable leads), they’ll need search term insights to optimize the keyword set, exclude poorly performing search phrases, and to improve ad copy and landing pages based on the actual search terms data.
Any advertiser using a third-party SEM management tool that automates keyword exclusions will likely suffer because of the lesser amount of actual searches it can use to build proper models.
It’s not the first time (and clearly not the last time!) Google has reduced the level of control advertisers can have over their campaigns. Some of the best examples in recent years are changes to the exact match type, loosening close variants, and “Smart” campaign types with very limited reporting capabilities. In a lot of cases, these types of product changes have proven they can improve performance. However, experienced marketers are leveraging every bit of data and insight they can get to improve their strategy, and getting ready access to real search data has become a pretty important part of that equation, even beyond search engine marketing.
Unless Google completely reverses this change (highly unlikely), there’s no perfect replacement. If you’re looking for ways that you can restore some of this insight to help with search term mining and potential keyword expansion or exclusion:
This is fresh news, so expect more followup from our team, along with emerging solutions and more analysis of the impact for advertisers. If you’re relying heavily on SEM for marketing leads, it may be just one more reason to expand your approach to more channels. If you’re interested in a full-funnel strategy review, you can give us a shout.