Broad Match

How Broad Match Works in the Age of AI

Broad match has existed since the early days of Google Ads, but its behavior has changed fundamentally with the rise of machine learning. In the original system, broad match simply expanded keyword targeting to include synonyms and related phrases, often resulting in irrelevant traffic that wasted budget. Advertisers learned to avoid it in favor of phrase match and exact match, which gave them tighter control. Modern broad match operates differently. Google's AI now interprets the semantic intent behind a search query and matches it to your keyword based on meaning rather than literal text overlap. A broad match keyword like "women's hiking boots" might trigger your ad for a search like "best waterproof trail shoes for women," even though none of the original keyword words appear in the query. This shift reflects a broader change in how Google's algorithms work. As Patrick Gilbert explains in Never Always, Never Never, ad platform AI uses both structured and unstructured learning to identify patterns in user behavior. Broad match gives the algorithm a wider field of data to learn from. Instead of restricting the system to a narrow set of predefined queries, it allows the AI to discover which search behaviors actually correlate with conversions.

Broad Match and Campaign Liquidity

The case for broad match is rooted in a concept Patrick Gilbert calls liquidity, drawn from a 2019 Interactive Advertising Bureau paper on machine learning in advertising. Liquidity describes the condition where every dollar is allowed to flow to the most valuable impression, made possible when advertisers reduce constraints and let the algorithm read the terrain. There are four dimensions of liquidity: placement, audience, budget, and creative. Broad match directly increases audience liquidity by expanding the pool of searches your ads can appear for. The narrower your keyword targeting, the fewer people the algorithm can learn from. Broad match removes that constraint. The IAB paper put it plainly: "In every case, allowing machine learning algorithms to process as much data as possible improves liquidity. By removing restrictions on a campaign, media teams can test more opportunities simultaneously and ensure budgets are applied to find the right people and produce the most cost-efficient results." This is why Google pushes advertisers toward broad match. It is not arbitrary. The platform's bidding algorithms need sufficient data volume to exit the learning phase and make confident predictions. Broad match feeds the algorithm more signals, which accelerates learning and improves the quality of bid decisions across a wider range of auctions.

Enjoying this? Never Always, Never Never goes much deeper into the mental models and decision frameworks that shape how we think.

Get the Book on AmazonLearn more about the book

When Broad Match Works and When It Does Not

Broad match is most effective when three conditions are met. First, the campaign uses Smart Bidding (Target CPA, Target ROAS, or Maximize Conversions). Smart Bidding evaluates each auction independently using hundreds of real-time signals like device, location, time of day, and user behavior. It can decide that a broad match query is worth bidding on aggressively or not worth bidding on at all. Without Smart Bidding, broad match is just expansion without intelligence. Second, the campaign has sufficient conversion volume. Google's algorithms need data to learn from. Campaigns with fewer than 15 to 30 conversions per month may not generate enough signal for the algorithm to make informed decisions about which broad match queries are valuable. In low-volume accounts, tighter match types may still outperform. Third, negative keywords are actively managed. Broad match will surface irrelevant queries. That is expected. The advertiser's job is to monitor search term reports and add negatives for queries that do not align with the business. This is not a "set it and forget it" strategy. It is a partnership between human judgment and machine learning. Broad match fails when advertisers pair it with manual bidding, ignore search term reports, or use it in campaigns without enough conversion data to guide the algorithm. In those cases, the expanded reach generates traffic without the intelligence to evaluate whether that traffic is valuable.

The Evolution from Control to Collaboration

The history of keyword match types reflects a broader shift in the relationship between advertisers and ad platforms. In the arbitrage era, control was the competitive advantage. Advertisers who could identify the right long-tail keywords, set precise bids, and segment audiences manually outperformed those who could not. Exact match and phrase match were the tools of precision. As Google's AI capabilities have grown, the competitive advantage has shifted from manual control to strategic collaboration with the algorithm. The advertiser's role is no longer to micromanage every keyword and bid. It is to set the right objectives, feed the system clean data, and provide the liquidity the algorithm needs to perform. Patrick Gilbert frames this as the central tension of modern advertising. Many advertisers remain obsessed with resetting learning phases and tweaking campaign structures, convinced that one more adjustment will reveal a profitable configuration. But as he writes, "If the underlying strategy is flawed, no amount of platform optimization will save it." Broad match is a tool that amplifies whatever strategy you feed it. A strong value proposition with clean conversion data and compelling creative will benefit from broad match's expanded reach. A weak strategy will simply spend budget faster on the wrong people.

Related Terms