How do I optimize broad match keywords?

How broad match works (and why “optimize” means guiding, not restricting)

What broad match is really doing now

Broad match is the default keyword match type in Google Ads. Practically, that means your keyword isn’t treated as a strict “contains these exact words” rule. Instead, broad match aims to show your ads on searches that are relevant to the meaning and intent behind your keyword, not just the literal phrasing.

To decide what’s “relevant,” broad match can consider signals beyond the query text itself, including the user’s recent searches, the content of your landing pages and assets, and the other keywords in your ad group (to help infer intent). This is why two advertisers can use the same broad keyword and see very different search term patterns—your account context matters.

Why Smart Bidding is the make-or-break pairing

Broad match creates reach; Smart Bidding creates control. Because every query is different and auction contexts change constantly, broad match is designed to work best when your bids are set auction-by-auction using conversion signals (for example, Maximize conversions or Maximize conversion value, with optional targets). In plain terms: broad match can open more doors, and Smart Bidding decides which doors are worth paying to walk through.

If you try to run broad match with weak conversion tracking or the wrong conversion goals, you’re essentially giving the system a messy scoreboard—and it will optimize toward the wrong outcomes with impressive efficiency.

Build the foundation first: measurement, goals, and campaign structure

Step 1: Make conversion goals “bid-worthy” (Primary) and consistent

Before you scale broad match, confirm that the conversions you truly want are eligible for bidding (Primary) and are the ones your campaign is set to optimize toward. Google Ads allows account-default goals and campaign-specific goals, and Smart Bidding optimizes only toward what’s included in the “Conversions” (and “Conversion value”) reporting for that campaign.

If you change conversion goal setups, expect performance volatility while bidding models adapt. When you must make goal changes, adjust targets gradually rather than swinging CPA/ROAS targets abruptly.

Step 2: Prefer value-based optimization when your leads/sales aren’t equal

If all conversions are not created equal (common for lead gen), broad match improves dramatically when you feed bidding higher-quality signals. Value-based bidding (optimizing for conversion value) is most effective when you can assign meaningful values (revenue, profit proxy, lead scores) and keep uploads/tags current. A practical baseline is choosing a single funnel stage to optimize toward that has a short conversion delay and at least about 15 conversions per month, so the system has enough feedback to learn.

Also, avoid stacking artificial constraints (like tight CPC caps) that prevent Smart Bidding from participating in the auctions it needs to learn and perform.

Step 3: Keep broad match ad groups tightly themed by intent

The fastest way to “break” broad match is to mix intents in one ad group and force one set of ads/landing pages to fit everything. Keep ad groups narrowly themed (one service line, one product family, one core intent). This improves relevance and makes the “other keywords in an ad group” signal work for you, not against you.

When your ad group is tight, your RSA messaging becomes a stronger filter: even if broad match explores, the ad and landing page alignment helps qualify the click and supports conversion rate—feeding Smart Bidding better outcomes.

Step 4: Decide whether to use individual broad keywords or the campaign-level setting

You can run broad match by adding broad keywords normally, or you can enable the “Broad match keywords” campaign setting that applies broad match across the entire campaign. In practice, I recommend starting with individual broad keywords (more controlled rollout), then considering the campaign-level setting only after you’ve proven your conversion measurement and query hygiene are solid.

A repeatable optimization workflow for broad match (what I do in real accounts)

1) Use search term visibility the right way: Search terms report + Search terms insights

Broad match optimization starts with understanding what’s actually triggering ads. The Search terms report shows the search terms that reached a “significant” volume threshold and triggered your ads, along with helpful columns like which keyword triggered and how the match type was classified for that search term.

Because not every query will appear (especially lower-volume queries), use Search terms insights to see aggregated intent categories. This is the best way to spot emerging themes (good and bad) that may not be fully visible at the individual query level.

2) Add negatives carefully (broad match needs guardrails, not walls)

Negative keywords are still your primary safety mechanism with broad match, but you want to use them with restraint and precision. Two key realities matter here: negative keywords don’t match to close variants, and they can be highly restrictive—meaning overuse can block valuable traffic and starve the algorithm of learning.

When adding negatives, focus on excluding clearly irrelevant intent (jobs, definitions, free, DIY, parts you don’t sell, competitor-only research terms if you don’t want them). If you’re managing many campaigns, consider an account-level negative keyword list to apply exclusions broadly across Search and Shopping inventory where appropriate, instead of repeating the same negatives everywhere.

Critical negative keyword checklist (keep it tight):

• Exclude only terms that are truly irrelevant or brand-unsafe (not merely “low converting” in a short window).
• Remember: negatives don’t match close variants—add plural/singular and common variations if needed.
• Prefer intent-based negatives (for example, “jobs,” “salary,” “definition”) over overly-specific negatives that might accidentally block qualified searches.
• Centralize obvious global exclusions using account-level negative keywords so you don’t miss coverage.

3) Promote winners into their own keywords (and keep broad as your exploration engine)

A proven approach is to let broad match discover demand, then “graduate” consistent, high-quality themes into their own dedicated keywords and ad groups (often phrase or exact for tighter steering). This isn’t about reverting to old-school micromanagement; it’s about creating cleaner intent buckets, better ads, and better landing page alignment for your best performers.

In other words: broad match finds what you didn’t think to bid on; your structure captures and scales it profitably.

4) Use audience signals in Observation to help Smart Bidding learn faster

In Search campaigns, audiences are commonly set to Observation by default (meaning your ads can still serve based on keywords, while you collect audience performance data and allow bidding to incorporate signals). This is ideal for broad match because it preserves reach while improving decision-making at auction time.

Layer “your data” segments (site visitors, customer lists where eligible), plus relevant intent segments, then watch how conversion rate and value differ. The goal isn’t necessarily to restrict reach; it’s to feed bidding better context and create audience-based insights you can act on with messaging and landing pages.

5) Tune targets and budgets with patience (broad match often needs room to work)

When advertisers say broad match “doesn’t work,” the most common culprit is targets that are set too aggressively for the learning stage, paired with budgets that don’t allow exploration. If you’re using Maximize conversions or Maximize conversion value with optional targets (behaving like Target CPA/ROAS), set targets based on recent reality, not best-case aspiration, then step them down/up gradually as efficiency stabilizes.

Also, if you apply recommendations to expand into broad match, be aware that incremental volume often requires incremental budget. If budget is capped too tightly, you can end up with a noisy test where broad match can’t properly explore while still meeting efficiency constraints.

6) Know what “good” looks like when you evaluate performance

Don’t judge broad match on a 3-day window. Evaluate after at least a couple of weeks or multiple conversion cycles so the bidding system has time to adapt. In the early phase, you’re looking for whether conversion volume/value increases while efficiency stays within an acceptable range—not whether every new query is perfect.

Then iterate: refine negatives for clearly wrong intent, tighten ad group themes when you see mixed intent, improve landing page alignment for the search categories you want more of, and keep upgrading the conversion signals (values, offline imports where applicable, and consistent goal settings) so Smart Bidding has the right incentives.

If optimizing broad match feels like balancing exploration with control, Blobr is designed to make that process easier: it connects to your Google Ads account, continuously analyzes what’s driving performance (from conversion signals and Smart Bidding behavior to ad group intent and search-term patterns), and turns best practices into clear, actionable recommendations. When you’re tuning broad match, its specialized AI agents can be especially helpful, like the Keyword Ideas Finder to uncover new, relevant opportunities and the Negative Keywords Brainstormer to add the right guardrails without over-blocking—so you can iterate with more confidence while staying in charge of what gets applied.