Why Your Google Ads Leads Are Low Quality — and What to Fix First

Step 1: Check what you're counting — conversion definitions set the quality ceiling

Before judging lead quality, check what your account calls a lead. The most common cause of poor quality isn't the traffic — it's the conversion definition. If your conversion action fires on a button click, a pricing-page view or a form start, Google's Smart Bidding is optimising to find more of exactly that: people who click buttons and start forms. Some of those are buyers; many are researchers, and the bidding system cannot tell the difference unless you tell it.

Open Tools → Conversions and audit each action against one question: does this fire on a real business outcome, or on an intent signal? A form submission on a thank-you page is a real outcome. A phone-call extension click is an intent signal — some of those calls become customers, many don't, and counting a tap-to-call as a conversion tells the algorithm to buy you more dials. The count setting matters too: "Every" on a lead action means one determined person refreshing and resubmitting counts five times.

The stricter standard that fixes quality: mark only outcomes sales would accept as a lead — a submitted form with real fields, a connected call above a duration threshold, a booked appointment. Move the softer signals (button clicks, pricing views, page scrolls) out of conversions and into GA4 as key events for observation only. Your bidding will change within weeks, because conversion tracking is the input every bid decision inherits. If the current setup can't tell a real lead from a click, fixing that is step one — the conversion tracking service does exactly this audit.

Step 2: Separate spam from bad quality before optimising anything

Some "low quality leads" are not leads at all. Fake form submissions — bots, scrapers, competitors checking your pricing, and agencies prospecting you — arrive through the same form as real enquiries and sit in the same CRM column. An account can look like it produces garbage when in fact a third of its submissions were never humans, and the humans converting fine are being averaged into a number that says otherwise.

Spam has a pattern. Submissions seconds after the page loads, gibberish or URLs in the name and message fields, disposable email domains, phone numbers with impossible formats, and bursts of submissions in short windows. Pull the last month of form fills, scan the name and message fields, and count the fake ones. The fix is a hidden honeypot field, basic rate limiting, and spam filtering on the form — not a change to the ad account. You cannot optimise what you haven't cleaned.

Also separate the near-spam: people who submit a form but won't answer the phone and don't reply to two emails. That category is genuine low quality (the rest of this article addresses it), but it's a different problem from fake submissions with a different fix. Sort one from the other before you touch a single keyword.

Step 3: Keyword intent — the searches that set the ceiling

After definitions and spam, lead quality is decided at the keyword level. Every keyword you bid on is a claim about intent: "emergency plumber near me" is a person with water on the floor; "how to become a plumber" is a person choosing a career. Both clicks cost money; only one produces a customer. When leads are consistently unqualified, the first place to look is which intents the account is actually buying.

Sort keywords by spend over 90 days and read them as intent claims. The recurring offenders are research and employment intents: "salary", "course", "certification", "how to", "free", "template", "what is". An illustrative pattern that appears in almost every account with a quality complaint: a recruiter advertising "hire developers" also spends steadily on "software developer salary" and "software developer roadmap" — job seeker traffic, billed at the same CPC, feeding the CRM with people looking for work, not staff.

The test for each keyword: would a salesperson want a call from the person who typed this, today? Not eventually, not with nurturing — today. Keywords that fail belong in a separate research campaign with its own (low) budget, where they can't contaminate the bidding data of the campaigns that pay the bills — or nowhere at all. Intent discipline at the keyword level does more for lead quality than any bid adjustment.

Step 4: Broad match — where searches you never chose get in

Broad match lets Google expand your keyword to anything it considers related, and its idea of "related" is broader than yours. A broad match keyword like "business coach" can match "life coach certification", "business coach salary" and "coaching business plan" — three different intents, none of them buyers. Broad match is also the setting Google most often recommends, because it generates volume, and volume without quality is precisely the problem you're diagnosing.

The rule from the waste-reduction framework applies doubly to quality: match type should match data volume. Broad match performs when Smart Bidding has enough conversion data to steer it — and it steers toward whatever your conversion action counts, which is why steps 1 and 4 compound. An account with a loose conversion definition and broad match keywords is optimising to find more form-fillers across the widest net Google can cast.

Check it empirically: filter the search terms report to matches on your broad keywords over 90 days, sort by cost, and estimate the share from searches outside your intent. Where the drift is large, the fix is usually not converting the keyword to exact — losing the keyword's useful discovery — but tightening its negatives (next step) and pairing broad match with a conversion action that only qualified enquiries can trigger.

Step 5: Negative keywords — cut the recurring junk at the source

Every search term you've decided is unqualified will come back through match-type expansion unless a negative keyword blocks it. Negatives are the memory of your quality decisions: without them, you re-buy the same bad searches every week and re-qualify (or reject) them by hand.

Build three layers, as covered in the waste framework: campaign-level negatives for terms irrelevant to that campaign; a shared negative list applied account-wide for the recurring offenders — jobs, hiring, salary, course, free, how to, DIY, template, careers — plus your own brand terms on non-brand campaigns; and a deliberate decision on competitor names, either bid on them knowingly or block them. Accounts with quality complaints almost always have empty or ancient shared lists.

Review search terms weekly while quality is the active problem — fortnightly once it's stable. One caution: an over-aggressive negative can block phrase variations you wanted, so after large additions, watch impression volume on affected keywords for two weeks. The goal is fewer unqualified leads, not fewer leads.

Step 6: Location settings — leads from people you can't serve

Location settings cause a specific and very common quality complaint: leads whose addresses are outside your service area. Google Ads defaults to "Presence or interest" — serving ads to people anywhere in the world who mention your target area. A searcher in another country typing "plumber Manchester" sees the Manchester plumber's ad, clicks, and submits a form from a place the business cannot reach.

The fix takes ten minutes: switch to "Presence: people in or regularly in your targeted locations", then open the locations report and check "matched locations" — where clicks actually came from. In local service accounts this routinely surfaces 10–20% of spend buying clicks from searchers who could never become customers, and each of those clicks displaces a real one. For service-area businesses, also confirm the radius or listed areas in the campaign settings match where sales would actually travel.

This is the highest-speed fix in the whole framework: the settings screen shows the problem and the change is a toggle. If a meaningful share of your unqualified leads are simply people you can't serve, nothing else in this list will matter until this is corrected.

Step 7: Lead form assets — low friction buys low commitment

Lead form assets (the forms that open inside the ad or on Google's own page) trade friction for volume: fewer fields, no landing page, an instant completion. The trade is real in both directions. Lead forms typically produce more leads at a lower cost per lead — and a lower close rate, because the person who filled a two-field form on a phone in ten seconds has made a much smaller commitment than the person who read your page and submitted a real enquiry.

This doesn't make lead forms wrong — it makes them a specific tool. They work best when the follow-up is instant (a call within minutes, or an automated message with a calendar link), when the offer suits a light hand (a quote range, a callback, a checklist), and when the CRM feedback loop exists to prove they close. Where follow-up is slow or the sale needs an engaged conversation, landing page leads usually qualify better despite the higher cost per lead.

If quality complaints started when lead forms were added, the diagnosis is likely structural rather than accidental. Compare close rates by lead source in the CRM before judging; if forms clearly close worse, either pair them with instant follow-up or move that budget to campaigns that send traffic to pages built to qualify. One trap to check: lead form leads can flow into the account via an integration that counts the submission as a conversion automatically — confirm the integration feeds real submissions and that no second tag is double-counting.

Step 8: Qualification — agree in writing what a good lead is

"Low quality" is only measurable once someone defines "high quality". In most small teams the definition lives in salespeople's heads — one rep calls a two-field form a great lead, another calls it tyre-kicking — and the marketing account gets blamed for a disagreement nobody has resolved. Write the definition down: which enquiry types count, which minimum details are required (postcode or city, real phone number, service needed), and what disqualifies instantly (job seekers, students, suppliers, outside the area).

Then make the definition visible where leads are judged. Add a disqualification reason field to the CRM and require it on every rejected lead. Within a month you'll have a table that turns opinion into data: which reasons recur, from which campaigns, keywords and lead sources. That table is the input for every remaining step in this framework — the negatives, the campaign budget shifts and the bidding feedback all read from it.

Response time is part of qualification in practice: leads contacted within minutes convert at multiples of leads contacted the next day, and an account can be blamed for quality when the real leak is speed. Before concluding the leads are bad, check how fast the good ones were actually called.

Step 9: CRM feedback — close the loop between sales and the ad account

Once leads are tagged with their campaign, keyword and search term (the click-to-closed-won tracking guide covers the mechanics), the quality data starts answering the questions this article started with. Compare close rate and revenue by keyword and campaign: the account that looked mediocre at cost per lead may contain one keyword closing at 30% and another closing at zero — information no ad-platform report can show you, because the platform only knows about form fills.

Feed the findings back into the account at two levels. Simple level: bid and budget weight toward the keywords and campaigns that produce customers, negative or restructure the ones that produce forms, and add the disqualification reasons from step 8 to your search-term review as recurring checks. Structured level: where volume justifies it, offline conversion imports (next step) let Smart Bidding learn the difference directly.

The loop is the actual job of management, and it's why accounts drift when nobody runs it: without CRM feedback, every optimisation decision is made on form-fill counts — the exact metric that created the quality problem. An hour a month comparing closed deals to their sources is the highest-leverage hour in this whole framework.

Step 10: Offline conversion imports — teach the bidding which leads become customers

The final step automates what step 9 does by hand. Offline conversion imports (or enhanced conversions for leads) pass real sales outcomes — qualified, quoted, won — back into Google Ads, so Smart Bidding optimises toward customers rather than submissions. A campaign that produced 40 cheap form fills and 2 sales stops looking successful; a campaign with fewer, better leads starts getting the budget it deserves, and the bidding system finds more of the closers.

Requirements are practical: you need a qualifying event worth importing (usually a qualified-lead or closed-won stage) and enough volume for the algorithm to learn — Google's guidance puts the useful threshold at around 30 of the imported outcome per month, so smaller accounts should start by importing a mid-funnel stage such as qualified lead rather than closed-won. Click identifiers (GCLID or enhanced conversions from email) must be captured at the form, which is why step 9's tracking is a prerequisite.

Set expectations honestly: OCI changes what the bidding optimises toward; it doesn't rewrite history or guarantee outcomes. Give it a full learning cycle after import — and expect the interim weeks to look turbulent as the account re-learns what "good" means. This is also the point where the quality problem stops being a weekly manual review and becomes a measured system. The B2B offline conversion guide covers the implementation in detail.

The lead quality checklist

All ten checks in one list, in the order that keeps each step's inputs trustworthy:

1. Audit conversion definitionsEvery conversion action fires on a real business outcome, not a click or page view; lead actions count "One".
2. Check conversion tracking integrityTags fire on the actual success event, verified in Tag Assistant or GTM preview — not assumed.
3. Separate spam from bad qualityScan the last month of submissions for bots and fakes; fix with honeypots and filtering, not ad settings.
4. Score keyword intentRead 90-day keywords as intent claims; move research and employment intent out of buying campaigns.
5. Audit broad match driftFilter search terms on broad keywords; measure the share from outside your intent; tighten with negatives rather than deleting reach.
6. Review negative keyword listsThree layers maintained weekly while quality is the active problem; watch impressions after large additions.
7. Fix location targetingSwitch to "Presence", then check the matched-locations report for clicks from places you can't serve.
8. Judge lead form assets by close rateCompare forms vs landing pages in the CRM; pair forms with instant follow-up or shift budget to pages that qualify.
9. Write the qualification definitionAgree with sales what counts and what disqualifies; require a reason on every rejected lead; check response speed too.
10. Close the loop into biddingTag every lead to its source; weight budget toward closers; import qualified outcomes once volume justifies it.

When to bring in outside help

Most of this framework is owner-executable: the conversion audit, locations, negatives and qualification definition are days of work, not months. Below roughly $1,500/month in media spend, doing it yourself with the checklist above is usually the right call.

Outside help earns its cost when the loop, not the list, is the problem — when quality is fine after a cleanup and then drifts, or when CRM feedback and offline imports need plumbing between systems you don't want to wire yourself. That's the ongoing case for Google Ads management, which starts at $399/month; the weekly search-term review and the CRM feedback loop are most of what keeps quality from regressing.

If you'd rather know which of these ten causes is yours before deciding anything, the $299 paid media audit works through this exact framework on your account and hands you the findings in writing — the causes, the affected spend, and the order to fix them. It's a standalone review, yours to act on yourself. Like everything else here, it isn't the cheapest option on the market; the point is that the diagnosis gets done properly.