Early in 2021 Google announced changes to match types that took full effect this July. Several changes were introduced, but the removal of broad-match modifier (BMM) and expansion of phrase match got the most attention and discussion in the industry. While there was uncertainty after the announcement and speculation on how phasing out BMM would affect match type strategy and campaign structures, in practice, there was very limited change for our clients here at Merkle. Moving forward, we don’t foresee any concerns or performance implications from leveraging phrase instead of BMM, which was widely adopted historically. However, there were less-discussed changes to broad matching that might warrant a change in your approach. 

Understanding the changes to broad match

The other change that was introduced and overshadowed by the BMM phase-out was related to broad match. Besides the keyword itself, broad match now also takes into account: 

- User’s recent search activity 

- Content of the landing page  

- Other keywords in the ad group 

As a result of this change, matches were expected to become more relevant, presenting expansion opportunities for advertisers. This new matching behavior seems to make broad match similar to Dynamic Search ads (DSAs), though a little more restrictive by taking into account the actual keyword and others in the ad group. Additionally, unlike DSAs, broad match still gives advertisers full control of ad copy and easier management of potential landing pages to consider for matching. 
 

Opportunities for advertisers 

Historically, pure broad match has driven some questionable matches and Google’s introduction of BMM seemed like a way to handle advertiser concerns. After having a poor experience with pure broad match, it might be hard to open that book again, but combined with advancements in bidding solutions, testing is warranted. Specifically, we have set up several broad match test campaigns that leverage smart bidding or SA360’s auction-time bidding. While the matching is less restrictive, having the bids curated to each individual auction and user query gives an advantage that did not exist in some historic experimentation. 

Based on our testing to date the results are promising, and we are looking to expand it across our advertisers, especially retailers headed into the peak season and looking to maximize revenue. Additional traffic varied from 2 to 10%, and was mainly incremental, meaning these were net new queries that haven’t – and couldn’t - match to other keywords or DSAs in the account. On top of that, when combined with smart bidding, the return was close to goal, even if intuitively the queries seemed to be less relevant. 

Conclusion 

Recent match type changes have been positive overall: without disruption to performance, they introduced a way to capture new emerging queries while maintaining account structures. Revamping phrase match gave a purpose for this under-utilized match type and provided a clearer definition for broad. Additionally, the improvements in relevancy for broad match itself now present an opportunity for advertisers to expand more safely and drive additional volume. Nicely done with this one, Google!