Google Ads announced a change to its search term reports, excluding search queries with low search volume. As an instrumental feature used by practitioners, this change to the search term report necessitates a strategic pivot in the management of paid search campaigns.
Google Ads’ paid search platform provides advertisers with reporting tools for the measurement of campaigns. One of the reports most often used is the “search term report” where users can see all search queries that have resulted in an ad impression.
These reports provide insight into the total cost expenditures for every query, frequency which queries have been used within a specified amount of time, number of clicks (and subsequently the cost for each click incurred), and the keyword which Google determined to be the closest and most relevant to the search query.
Additionally, advertisers are better able to visualize the search landscape for any given account. This is a crucial step in optimizing the paid search traffic to maximize the performance of media budgets. Historically, these reports had no threshold for reporting, displaying search terms that only showed up (impressions) once alongside terms that have garnered thousands of impressions in the selected date range.
On Sept. 1, 2020, Google Ads announced1 that it would eliminate terms that it deems “insignificant,” removing low-traffic terms from the reports beginning this month.
This change upends what has been a foundational source of insight into search behavior. It could also greatly impact marketers’ ability to optimize their paid search efforts and realize their current level of performance.
Why Did They Change It?
Google stated that this decision was made in order to “to maintain our standards of privacy and strengthen our protections around user data.” Recent privacy laws such as GDPR and CCPA have been set in place to regulate the collection of tracking data on multiple platforms; they encompass trackable facets of a user’s online journey ranging anywhere from cookies to mobile device IDs. Google Ads has come under recent scrutiny in the world of search marketing, as an error in the platform allowed advertisers to see the actual phone numbers of users who clicked a call extension.2
These inadvertent violations of personally identifiable information (PII) are reasons for a company of Google’s size to take a proactive, rather than reactive, approach to the management of personal information. For similar reasons, Google announced its intention to phase out third-party cookies on its Chrome browser in January this year.3
As Google outlined, the reporting on low-traffic queries will be obscured moving forward. From the standpoint of the company, overtly unique search queries are not unlike a PIN. It can be argued that a user’s search behavior could be unique enough to track them through these query reports. Including high-traffic terms only in reports would act as a stopgap from specifically identifying an individual user.
This is not a novel occurrence within paid search, paid social and programmatic media. Uploading data such as a customer email list allows for effective remarketing efforts or lookalike audience expansion. These first-party data audiences are required to meet a minimum size before they are eligible for direct use. In the case of Google’s customer match feature, a minimum of 1,000 records in a list are needed, as any less could risk identifying an individual user through search behavior.
What Does This Mean For Marketers?
Marketing professionals use the search term report to gain insights that lead to keyword expansions for scaling the campaign’s reach and improving relevance. It can also lead to optimizations that yield greater return on ad spend (ROAS).
A common strategy for identifying new keyword opportunities is to deploy phrase and broad match types, then analyze the resulting search traffic using the search term report. The report might suggest prospective keywords that the marketer can further test to gain scale. Alternatively, the report can reveal poorly performing keywords that can be used as negatives in the campaign to improve relevance and efficiency.
Phrase and broad match keywords are subject to many unique queries. The searches range from short-tail (two words) up to long-tail terms. The long tail, low-traffic search queries are where the majority of optimization insights are found.
The change that Google is implementing to the search term report will result in fewer actionable insights by redacting these low-traffic search queries. This action will blind marketers to a significant degree of optimization potential.
Categories with short conversion funnels may have predictable search traffic requiring little keyword research. Search term reports for these accounts may already be thin, each term likely commanding thousands of impressions each month. In this situation, the impact of Google Ads’ reporting change will be minimal.
On the other hand, high-consideration categories may have a longer conversion funnel that includes a research phase. In-market consumers may use many and varied search terms as they familiarize themselves with the category. These categories will be impacted by the change to a greater degree.
How Severely Could This Impact Performance?
In order to quantify the potential impact of this change, we analyzed search terms YTD for a representative client.
At this time, Google does not specify the criteria used to determine what would be significant or insignificant. For the purpose of this analysis, we assumed that traffic frequency of less than one impression per day (243-day date range from Jan. 1, 2020 to Aug. 31, 2020) would be insignificant. Exact match keywords were excluded.
If one defines “insignificant” traffic as having less than one impression per day, marketers are potentially blind to 99% of unique searches.
It is unsurprising that the data displays such a large number of insignificant traffic. This is characteristic of these match types. The high number of search terms with insignificant volume are the terms pay-per-click (PPC) teams scour for relevance, new keyword ideas, measurement of customer sentiment, and negative keyword optimization.
Another area of concern is the potential cost of these terms:
When comparing the share of spend to the entirety of all paid search campaigns (exact match included), 15.6% of media spend will not have any reportable search terms.
Google has been championing its machine learning algorithms for years, touting performance that parallels or exceeds a manually optimized campaign. This may be another move by Google to increase users’ reliance on what is becoming a black-box system.
Promising as it may be, machine learning in paid search does not guarantee performance. A number of factors from bid price, ad relevance, estimated impact of the ad for the user, and performance history are used by the auction to determine how high an ad can place on a results page.
It is not uncommon for automated bidding to increase ad spend, sometimes requiring up to 30 days for a “learning period” in which the advertiser is told not to make any changes to the campaign as the AI is gathering data to optimize.
Such learning periods may be acceptable with larger budget accounts, but is generally unacceptable for accounts where immediate sustained impact is called for. This is where the manual analysis and optimization of paid search accounts fall on PPC practitioners.
It can be argued that machine learning’s ramp-up period can be outperformed by an experienced PPC subject matter expert (SME) using the reporting features within Google Ads. Limiting the insights the PPC SME can derive from search queries hampers what has been a go-to method for ensuring traffic relevance, and makes the machine-learning option more appealing to less savvy paid search users.
Moving Forward: What Should Marketers Do?
As requirements for each paid search program are unique, there are no clear-cut guidelines for marketers. The degree to which an account may be affected depends on the nature of paid search queries for that particular product category or vertical market.
Fortunately, other platforms that report on similar aspects of search behavior are not affected by this change. Microsoft Ads’ search query reports (which, as of this writing, have not adopted this change) can be a good interim source of data to maintain search traffic relevance.
Practitioners may also rely more heavily on syndicated insights tools such as SEMRush or SpyFu to gain perspective on the search landscape. High-level sources such as Google Trends can help fill in this newly created gap as well.
Finally, expect to see increasing collaboration between PPC and SEO channels as organic search data becomes a more important resource for paid search professionals.
A Refresher On Match Types
PPC’s use of keywords and the degree of relevance that a keyword can be applied to a search query depends on the use of “match types.” These specify to Google how identical a search query should be to a keyword within the account in order to potentially serve an ad.
Match types can be classified into the following:
- Exact: Has to match the search query precisely as it is specified in the keyword
- Phrase: Can either match as exact does, or include additional terms before/after the keyword
- Broad (and broad match modifier): The most liberal match type. May or may not include the keywords specified in the search query. Synonyms and related terms may be seen as relevant enough to match these keywords. Can match any of the two previous keyword match types as well
Search term reports are used to list queries which have resulted in the client’s ad being shown. A report solely focusing on exact match types yields little insights and surprises, as the traffic is predictable. It is when the other two match types are used that marketers can begin to see phrases or word combinations that may or may not be relevant.
These terms can be either adopted as new keywords, or negated much in the same way a keyword can be implemented:
- Exact Negative: Excludes search terms that precisely match the negative keyword
- Phrase Negative: Excludes any searches with the negative keyword included at any point
- Broad Negative: Excludes any searches with the negative keywords included in any order at any point within the query
While some search terms within an account are perennially high-traffic, there is often a chance of an irrelevant term being low-traffic, yet commanding a high cost-per-click. This irrelevant term can be implemented as a phrase negative.
Associating a set of negative keywords from the onset of an account or campaign can help to limit the likelihood, but ongoing analyses of the search term reports usually follows to find account-specific negatives to implement. No two PPC accounts are alike; negative keyword lists will never be the same, especially after an account has been subject to competent optimization.
Digital Media Director
Digital Media Supervisor