The broad match keyword type in Google Ads is often frustrating for advertisers. Unlike the more specific keyword match types of phrase and exact, broad match allows more unrelated queries to trigger an ad.
Consider a search for “winter jackets.” In exact match, searchers must type “winter jackets” or a close variant, such as the singular version, for the ad to show. In phrase match, users can have terms before or after the keyword. Queries such as “men’s winter jackets” and “blue winter jackets that are comfortable” would trigger ads.
Broad match is different — any of these queries could trigger an ad:
- “Winter coats”
- “Jackets for winter”
- “Cold weather jackets”
- “Heavy jackets”
The intent of those queries ranges from specific to general. And general, undefined intent produces irrelevant traffic, prompting advertisers to avoid broad-match keywords. But that approach is changing with Google Ads’ automation and machine learning.
Increasingly, Google relies on user intent through signals such as search history and demographics to align ads with searchers. Keywords remain a factor in ad impressions but not the sole determinant. Google recommends broad match because it generates the most data for its algorithm.
Still, for top performance advertisers should combine broad match keyword types with smart bidding and responsive search ads. Smart bidding examples include setting a target return on ad spend or cost per acquisition. A target ROAS of 300% instructs Google to optimize toward that percentage. Smart bidding adjusts bids based on the likelihood of a conversion and the associated goal.
Responsive search ads allow Google to show different combinations of ads based on the searcher. An identical query from two searchers could trigger two different ads, each targeted to that person. RSAs are the default ad type in Search Network campaigns.
Testing Broad Match
In early April 2023, I began using broad match across a client’s campaigns. The goals of those campaigns were generating revenue and driving leads. Each campaign had a target ROAS bid strategy. I added broad-match keywords to the phrase- and exact-match types already present.
Then I added more broad match keywords in early May after seeing conversions and cost per conversion steadily improve throughout April. Those additional keywords increased May conversions substantially while cost per conversion declined.
I also reviewed the Search Network volume before and after implementing the broad match keywords. To no surprise, clicks, cost, conversions, and revenue all saw at least 70% increases from April 3 through May 28 versus February 6 through April 2. Adding broad match keywords was the primary reason for the influx of traffic, although I also included new phrase- and exact-match keywords in the second eight weeks.
Here’s a look at Search Network conversion and revenue changes period over period.
Search Network saw 7,000 more conversions in April and May despite the cost per lead increasing by $67. Revenue increased by $1.5 million while the ROAS dropped. Thus adding broad match keywords contributed significantly more conversion and revenue volume while efficiency declined.
Google Ads will never ignore keywords in my view, although it will likely de-emphasize them. Google considers many factors to show the right ad to the right user. The query is only a start. We already see that phrase- and exact-match keywords can appear for query variations. Moreover, Dynamic Search Ads and Performance Max campaigns — both significant traffic generators — show ads on the Search Network and don’t use keywords.
Hence broad match keywords will be a primary way for advertisers to scale growth on the Search Network. The strategy allows Google increased control over which queries show ads, but advertisers can gain conversion and revenue volume. And deploying traditional optimization tactics such as negative keywords and replacing low-performing assets can target growth.
I will continue using broad match alongside phrase and exact match keywords in campaigns meeting goals. Performance may be stagnant at first as the algorithm learns. But I will know after a few weeks how the broad match keywords are performing. I’ll consider the broad-match test successful if the conversion and revenue volume are higher, even if the efficiencies are the same or slightly lower.