How to fix: Skewed keyword match-type distribution

TL;DR

Your campaign's keyword mix is dominated by a single match type — usually 100% broad or 100% exact. All-broad without the 30-conversions-per-30-days Smart Bidding gate burns budget on queries the algorithm cannot score, typically 15-30% of spend on irrelevant terms. All-exact starves the account of discovery volume that feeds long-tail conversions, capping reach below available demand. Fix: build a balanced ladder — broad on conversion-rich themes (only when gated), phrase on mid-funnel, exact on bottom-funnel and brand.

Why it matters

Google Ads has three match types and they are designed to work together, not in isolation. Broad match expands a keyword to related searches (synonyms, intent variants, contextual co-occurrences). Phrase match expands to queries that contain the keyword's meaning. Exact match restricts to queries with the same intent as the keyword itself [1]. Each type plays a different role in the auction funnel, and Google's 2024-2025 documentation explicitly frames broad match as the discovery layer that Smart Bidding learns from, not a standalone strategy [2].

When the mix collapses to a single match type, the funnel breaks in two predictable ways.

Case 1: 100% broad match, gate uncleared. Smart Bidding needs roughly 30 conversions in the last 30 days to predict per-query bids reliably (see Smart Bidding under 30/30). Below that floor the algorithm runs in permanent exploration, and broad match — which expands aggressively into related but lower-intent queries — feeds it a high-noise signal. The account burns 15-30% of clicks on terms that have no realistic path to conversion. The visible symptom is a Search terms report dominated by tangential queries: a SaaS account targeting "project management software" gets clicks on "free Excel templates", a B2B legal account on "lawyer salary jobs". Smart Bidding cannot solve this — it has not yet learned what converts.

Case 2: 100% exact match (or near-100%). The opposite failure mode. Exact match restricts the campaign to queries the account already knows convert, but it also blocks the discovery surface that finds new converting queries. Long-tail demand passes the account by. Impression share for eligible queries is high, but eligible-query volume is artificially small. This commonly appears on legacy accounts that survived the 2018-2021 era of exact-match-only safety: the account is profitable but stalled, and additional budget produces no incremental volume because there are no broader-match keywords to spend it on.

The healthy distribution uses each match type for its design intent. Broad match expands top-converting themes — once the Smart Bidding gate is cleared, broad terms feed the algorithm the variety it needs to find auctions you would not have keyword-matched manually [2]. Phrase match holds the mid-funnel: queries that contain your intent but vary in modifier order. Exact match locks down bottom-funnel intent (high-CTR commercial queries, branded variants) where you want maximum bid control and minimum query drift.

The recoverable performance is meaningful. On audited Search accounts, rebalancing from over-broad to gated-broad-plus-phrase typically lifts conversion rate 5-12% within 30 days. Adding phrase-match coverage to a near-100%-exact campaign typically lifts eligible impressions 10-20% within 30 days, with conversion volume tracking the impression lift after the initial Smart Bidding stabilization.

How to fix

1. Pull the keyword report grouped by match type. Open Campaign → Keywords → Search keywords and segment by Match type. Note the percentage of spend (not just keyword count) in each bucket. A campaign with 200 broad keywords and 5 exact keywords can still spend 80% on the 5 exact keywords if those are the converters — the spend split is what matters.
2. Verify the Smart Bidding gate before deciding broad-match weight. Check the campaign's 30-day conversion count on the Overview tab. If the campaign is below 30 conversions in 30 days, broad match cannot be safely scaled. Either bring the campaign over the threshold first (see Smart Bidding under 30/30) or restrict broad match to a small test budget while the rest of the campaign runs on phrase and exact.
3. If above 60% broad and the gate is not cleared: convert the worst-performing broad keywords to phrase match. Sort the keyword report by cost descending, then add a Conversions column. For broad keywords with zero or near-zero conversions and meaningful spend, change the match type to phrase (Keyword settings → Match type). Keep broad match only on the top three to five converting themes.
4. If above 80% exact: add phrase-match versions of the top converting exact keywords. For each exact keyword that drives meaningful conversions, add the same root as phrase match in the same ad group. Then open the Search terms report and look for queries that triggered your exact keywords through close-variant matching — those are candidates for new phrase additions.
5. Add the Search terms report into the cadence. Every 14 days for the first 60 days post-rebalance, run the Search terms report. Add irrelevant queries as negatives at the ad-group or campaign level (Tools → Shared library → Negative keyword lists). Promote high-converting search terms to their own exact-match keywords. This is how the ladder stays balanced as Smart Bidding learns.
6. Re-check the distribution at 14 days. Confirm CTR did not drop more than 10% (a sign of over-tightening) and that wasted-spend percentage stayed flat or improved. Expect a 7-14 day Smart Bidding learning-phase dip after large match-type rewrites — do not roll back inside that window.

Common mistakes

• Switching to 100% broad because Google's UI recommends it. Google's auto-recommendations promote broad match without gating on the 30/30 conversion floor. On a sub-threshold campaign, accepting the recommendation guarantees the algorithm runs broad without the data to score it. Always check the conversion gate first.
• Treating match types as keyword settings instead of campaign strategy. Match type is a portfolio decision: broad for discovery, phrase for mid-funnel, exact for control. Setting match type per-keyword without thinking about the overall ladder produces incoherent campaigns.
• Adding broad-match modifiers (BMM) syntax. BMM was retired in 2021 — the +keyword syntax now resolves to phrase match. If your account still has BMM strings, audit them and explicitly choose phrase or broad.
• Mixing match types within a single ad group without negatives. If "running shoes" sits as broad, phrase, and exact in the same ad group, the broad version will canniblize impressions from the exact version. Either separate match types into different ad groups, or accept that the broad version dominates and remove the duplicates.
• Letting close-variants do all the discovery work. Exact match's close-variant expansion has widened significantly since 2018, but it is not a substitute for actual phrase or broad coverage. Close variants stay within near-intent of the exact keyword; phrase and broad surface intent variants you would not have keyword-matched manually.

FAQ

Should broad match be the default match type in 2026?
Only on campaigns that have cleared the 30-conversions-per-30-days Smart Bidding gate and have clean negative-keyword hygiene. Google's documentation promotes broad-plus-Smart-Bidding as the modern Search default [2], but the conversion floor is a hard prerequisite — without it the algorithm cannot score broad-match queries.

What about broad-match modifier?
BMM was retired in 2021 and merged into phrase match. The +keyword syntax still parses but treats the term as phrase. Clean up legacy BMM strings and explicitly choose phrase or broad.

How do I know if my exact-only campaign is leaving demand on the table?
Check the Search terms report. If close-variant matches surface queries your exact keywords did not cover (different modifier order, synonyms), those queries are signal that phrase match would expand reach. Also check Impression share lost (rank) — if it is high while Impression share lost (budget) is near zero, the campaign has unused budget that broader match types could absorb.

Does PMax replace the need for a match-type strategy?
No. PMax handles its own query routing internally, but Search campaigns still need a balanced match-type ladder. The two campaign types are complementary, not substitutes.

Sources

[1] Google Ads Help — About keyword matching options: https://support.google.com/google-ads/answer/2497836
[2] Google Ads Help — About broad match keywords: https://support.google.com/google-ads/answer/7478529