Google wants you to go broad. Every recommendation card in your account nudges you toward it. Every Google rep on the phone pitches it. And for the first time in the history of paid search, they might not be entirely wrong-but they're not entirely right, either.
Google Ads is moving from keyword matching to intent modelling.
Broad match has existed since the dawn of Google Ads, but advancements in AI and large language models now allow it to understand search intent. That sounds impressive in a press release. The operational reality is messier. Broad match in 2026 can surface converting queries you never imagined-and it can burn through your monthly budget on irrelevant clicks before Tuesday if the guardrails aren't set. This post isn't a cheerleading piece for broad match, and it isn't a warning siren against it. It's a practitioner's breakdown of how the matching system actually works now, which signals you control, which you don't, and how to structure campaigns so broad match serves your goals instead of Google's revenue targets.
How Broad Match Interprets Intent in 2026
Broad match is no longer the blunt instrument it was five years ago. Machine learning capabilities and large language models now better understand a user's intent. Google trains them with billions of pieces of text so the system can learn all the different variations and meanings of a word or phrase. The result: matching now happens at the semantic layer, not the lexical one.
In the early days of broad match, the technology would connect the dots between keywords in a search query and an ad, but it wouldn't necessarily take the order of the words into account. If someone was looking for a flight from London to Paris, they could have been shown an ad for a flight from Paris to London. Today, broad match knows that going from A to B isn't the same as the other way around.
What changed? Google shifted from synonym tables to neural-network-driven intent models. Broad match is the only match type that uses all of the signals available in Google Ads to understand the intent of both the user's search and your keyword. Those signals include:
- User's recent search activity - if someone searched for "best CRM for agencies" ten minutes ago, their subsequent search for "client management tools" carries different weight than the same query from a casual browser.
- Landing page content -
the content of the landing pages and assets, along with other keywords in an ad group, help Google better understand keyword intent.
- Location and device context -
Google takes the user's location into account to show ads that are geographically relevant, and even if the search query doesn't explicitly mention a location, Google can use the user's geographic information to ensure that the ads are locally relevant.
- Historical account performance -
Google Smart Bidding algorithms are trained by past performance. As certain types of searches consistently lead to conversions, the system becomes increasingly confident to show ads for similar queries.
This is why broad match can sometimes produce counterintuitive results. Modern broad match isn't just matching keywords-it's matching user behavior patterns. Practitioners have seen "Football Statistics" searches convert extremely well for "Football Tickets" broad match keywords because Google shows these ads only to users whose behavior indicates purchase intent, not to everyone searching for statistics.
The Signals You Can Shape vs. the Ones You Can't
You control landing page content, ad group theming, negative keyword lists, audience signals, and conversion tracking accuracy. You do not control Google's user-behavior graph, its LLM interpretations, or its auction dynamics. That division matters because it defines your optimization surface: everything useful you can do with broad match happens at the signal-input layer, not at the matching layer itself.
Why Broad Match Economics Have Shifted
The financial argument for broad match has changed since 2023, and the data is stark. Between June 2023 and June 2025, broad match CPCs rose 29%, while phrase match surged 43%. Advertisers relying on phrase are clearly overpaying for the same queries. That pricing gap creates a structural incentive: Google is making broad match comparatively cheaper while making tighter match types more expensive.
Phrase match has also declined in quality. Close match variants now allow phrase and exact to behave as broadly as broad match, without the AI that helps filter for relevance. This is the part that surprises many advertisers. Phrase match in 2026 isn't the buttoned-up match type you remember from 2019. Both experts agree: phrase match has become a less reliable tool. It may still work in some legacy campaigns or specific cases, but generally speaking, most new accounts can skip it in favor of exact or broad.
62% of advertisers using Smart Bidding now use broad match as their primary match type. That number reflects market reality, not just Google marketing. Still, adoption doesn't prove universal fit. The question is whether broad match is right for your account.
The Non-Negotiable Prerequisites
Broad match without the right foundation is still reckless. Three conditions must be met before you even consider switching.
Conversion Tracking Must Be Accurate and Aligned to Revenue
Broad match is only as intelligent as the data feeding it. If your account sends clear signals about what a valuable conversion looks like, broad match can expand intelligently into relevant territory. If your signals are weak, duplicated, inflated or misaligned with business goals, broad match will expand in the wrong direction.
This is the most common failure point I see in audits. Accounts that track newsletter signups, page views, and form submissions all as primary conversions are giving the algorithm permission to chase volume over value. If your conversion tracking records every form submission as equal, broad match will optimise for volume. You may see cost per lead fall, while sales teams complain about declining quality.
Fix this before touching match types. Import offline conversion data where possible. When broad match is paired with offline conversion imports, it becomes far more strategic-instead of optimising for any lead, it learns which search contexts produce revenue-generating clients.
You Need Sufficient Conversion Volume
Smart Bidding needs data to learn, and broad match needs Smart Bidding to function properly. Automated bidding is only going to be able to perform if it has enough data to learn on and infer user interests. If you're not hitting a minimum of 15 conversions per month on the desired conversion actions, this strategy could be tough to pull off. Google's own recommendation for Smart Bidding strategies like Target CPA or Target ROAS is more aggressive: aim for 30 conversions in 30 days before using Maximize Conversions. This threshold helps Google's AI optimize bidding and targeting.
For accounts spending under $2,000/month or those with fewer than 15 conversions per month, Manual CPC remains the safest way to control costs. It prevents the AI from over-testing and burning your limited budget during a volatile learning phase.
Smart Bidding Must Be Active
This point deserves emphasis because some advertisers still attempt broad match with manual bids. It's critical to use Smart Bidding with broad match because every search query is different, and bids for each query should reflect the unique contextual signals present at auction-time.
Manual bidding does not make sense with broad match in 2026. Without auction-time bidding adjusting for each individual query, broad match becomes an open spigot with no metering.
Building a Negative Keyword Strategy That Actually Works
Broad match expands reach. Negative keywords ensure that reach remains relevant. This is the most important operational discipline in a broad match campaign, and it's where most accounts fall short.
Account-Level vs. Campaign-Level Negatives
Account-level vs. campaign-level negatives serve different purposes. Account-level negative lists should contain universal exclusions that apply across all campaigns: jobs, careers, salary, free, DIY, how-to, resume, etc. Campaign-level negatives should be specific to that campaign but not relevant elsewhere.
Organize themed negative keyword lists rather than dumping individual terms ad hoc. Create lists like "Job Searches," "DIY Intent," "Competitor Terms," "Informational Queries" and apply them to relevant campaigns. This scales better and prevents the management chaos that emerges when three different team members add negatives independently.
Proactive vs. Reactive Negatives
Most advertisers only add negatives after they see bad queries in the search terms report. That's reactive. The algorithm should theoretically avoid non-converting searches automatically, but in practice it often tests irrelevant searches for weeks before learning they don't convert. Negative keywords prevent this waste upfront.
Before launching a broad match campaign, seed it with negatives you know will be problems. If you sell enterprise SaaS, add "free," "open source," "tutorial," "salary," and "jobs" on day one. Every week for the first 60 days, review search terms and add to those lists. Negative match types haven't changed-they still work exactly as originally designed. That predictability is an asset you should exploit.
Match Type Selection for Negatives
Here's a detail many advertisers miss: When you add a new negative keyword, Google defaults to exact match. That's not an accident. Exact match blocks the fewest searches, which keeps your spend higher. Always manually change negative match types. Phrase match negatives offer a good balance between blocking enough and not accidentally over-blocking and should be your default in most cases.
The Role of AI Max and Where Broad Match Still Fits
AI Max is one of the most important Google Ads developments to understand in 2026 because it changes how traditional Search campaigns can operate. Many advertisers are familiar with broad match, responsive search ads, smart bidding, and automatically created assets as separate concepts. AI Max brings more of those ideas together into a unified Search campaign enhancement framework.
Data shows that advertisers that activate AI Max in Search campaigns typically see 14% more conversions or conversion value at a similar CPA/ROAS. For campaigns that are still mostly using exact and phrase keywords, the typical uplift is even higher at 27%.
But practitioners are divided on whether AI Max is worth the transparency trade-off. The difference is that broad match still respects your keyword list as the starting point. AI Max throws that out and lets Google decide entirely. Store Growers founder Dennis Moons captures the practitioner consensus well: stick to broad match with Smart Bidding. You get the reach without giving up visibility into what's driving your results.
On February 26, 2026, Google announced that beta access to text guidelines is now open to all advertisers worldwide, across both AI Max for Search campaigns and Performance Max campaigns.
Text guidelines let you block up to 25 specific words from AI-generated ad copy. That's progress on the control front, but it's still a narrow guardrail for advertisers used to writing every word of their ads. For most accounts, broad match with Smart Bidding remains the pragmatic middle ground in 2026. It gives Google room to find high-intent queries you haven't anticipated, while keeping keywords as the foundation and the search terms report readable. Reserve AI Max for testing once your core search campaigns are stable and performing.
When Broad Match Is Wrong for Your Account
Not every account should run broad match. Recognizing where it fails is as important as understanding where it excels.
High-CPC, Low-Volume Verticals
For high-CPC accounts or tight budgets, starting with phrase match and graduating winners to exact match is still the smarter play. In legal, B2B SaaS, or financial services, a single mismatched click can cost $50–$150. The exception is high-CPC accounts where intent changes with a few words and each mismatch costs real money. There, it's worth trading volume for accuracy.
B2B Accounts with Niche Terminology
If you're in B2B or a niche space, you'll likely still rely on exact match and manual control. But if you're in B2C with conversion data to work with, broad match is likely your best tool. An industrial manufacturer targeting "hydraulic cylinder repair" cannot afford broad match interpreting that as "hydraulic press YouTube" queries.
New Accounts Without Historical Data
In new accounts without historical conversion data, broad match can be difficult to control. Starting with phrase match allows you to build performance history, understand user behaviour and gather data. Once that foundation is in place, you can consider layering in broad match.
A Practitioner's Framework for Testing Broad Match
If you've met the prerequisites, here's a systematic approach to introducing broad match without blowing up what's already working. Step 1: Isolate the test. Run broad match keywords in their own dedicated campaign so you can easily track performance and control their budget separate from exact match campaigns. Never convert existing phrase or exact campaigns to broad wholesale. Step 2: Set aggressive Smart Bidding targets. Set a tROAS goal higher than the average non-brand actual ROAS, or 1.5x to 2x your goal. This constrains the algorithm's exploration and prevents it from chasing marginal traffic early. Step 3: Layer audience signals. Add audience signals-layer in data segments, remarketing lists, and demographics. Google then aims broad match at people closer to your buyers. Use observation mode, not targeting mode, so you inform the algorithm without restricting its reach. Step 4: Seed negatives proactively. Add your standard negative keyword lists before the campaign goes live. Include competitor terms, job-related queries, informational intent terms, and anything your existing campaigns have already identified as waste. Step 5: Review search terms aggressively for 60 days. Review search terms every few days. Kill bad themes with negatives. Copy strong new queries into phrase or exact in your main campaigns. This turns broad match into a discovery engine feeding your controlled campaigns. Step 6: Evaluate after 4-6 weeks with sufficient data. Don't judge broad match performance in the first seven days. Give new bidding strategies at least two weeks to gather sufficient data, learn the patterns, and fully stabilize before making any major changes or judgments. Realistically, you need four to six weeks of data before the comparison is meaningful.
Consolidated Account Structures Beat Granular Segmentation
The old playbook of single-keyword ad groups (SKAGs) and match-type-segmented campaigns is dead. Over-segmentation limits machine learning. Many accounts are structured around legacy control models-separate campaigns for minor variations, overuse of match types, fragmented budgets. AI systems learn better from consolidated, high-volume data.
Gone are the days of building 20 ad groups with hyper-segmented SKAG structures. Thanks to broad match and Google's machine learning, practitioners recommend starting small-even with just a few keywords and two or three ad groups.
The modern structure looks like this: group campaigns by business objective (not match type), use 3–8 thematically similar keywords per ad group, and let Smart Bidding optimize across queries. Group your keywords into ad groups and campaigns with similar themes. This makes it easier for Google to understand your keywords, select the best one, and determine which ad should serve for each search query.
Landing page strategy matters here too. Pages need to answer broader search intent, provide more topic depth, offer stronger alignment between ads, search context, and destination content, and include clear explanations, not just conversion widgets. When broad match sends a wider variety of queries to your pages, those pages need to address more than one narrow intent. --- The shift to broad match isn't optional. Google's pricing structure, AI improvements, and product roadmap all point the same direction. That control is being replaced by calibration. The advertisers who thrive won't be the ones clinging to exact match structures from 2018, and they won't be the ones blindly accepting every Google recommendation either. They'll be the ones who understand that broad match amplifies the signals you feed into the system. If those signals are weak or misaligned with commercial outcomes, broad match can scale the wrong type of traffic very quickly. Clean conversion tracking, disciplined negative keyword management, audience layering, and consolidated account structures-these are the inputs that separate profitable broad match campaigns from expensive lessons.
Your ability to adapt as match types continue shifting matters more than picking the "right" match type today. Meet Google halfway. Feed it better data. Build the guardrails. Then let the machine do what machines are actually good at-processing millions of auction-time signals faster than any human ever could.
Ready to optimize for the AI era?
Get a free AEO audit and discover how your brand shows up in AI-powered search.
Get Your Free Audit