A £10,000 month produces 200 form submissions. Sales accepts 24. Google reports a £50 cost per lead; the commercial number is £417 per sales-qualified lead.
That gap is the problem. Improving B2B Google Ads lead quality does not start with another bid adjustment. It starts by separating four failures that look identical in a dashboard: irrelevant searches, poor-fit prospects, automated spam and conversions that were never real enquiries.
If those categories stay mixed, optimisation makes the problem worse. A cheap form completion is rewarded whether it came from a procurement director, a job hunter or a bot. The account then finds more of whatever produces the easiest completion.
The remedy is a controlled chain: define acceptance, trace rejections back to traffic, use ads as filters, harden the form and measure the actions the business values.
Define B2B lead quality before touching the account
“Poor quality” is not a usable diagnosis. Marketing and sales need a short, shared definition for each stage:
- Raw submission: anything recorded by the form, phone or chat tracking.
- Valid enquiry: a unique, contactable human asking about something the company sells.
- Marketing-qualified lead: a valid enquiry that meets agreed fit rules, such as geography, company type and minimum project value.
- Sales-qualified lead: sales has confirmed a credible need, buying route and timeframe.
Write the fit rules down before changing the campaigns. For a UK managed service provider, that might mean a UK organisation with 50–500 employees, a live infrastructure requirement and at least £2,000 monthly budget. It should not mean “sounds promising”.
Give sales no more than six rejection codes: spam, duplicate or unreachable, wrong geography, wrong customer type, below commercial threshold and no active requirement. Require a code within two working days. Free-text notes can add context, but they cannot support reliable analysis on their own.
Consider a realistic before-and-after:
| Metric | Before | After |
|---|---|---|
| Media spend | £10,000 | £10,000 |
| Raw submissions | 200 | 125 |
| Valid enquiries | 120 | 106 |
| Sales-qualified leads | 24 | 35 |
| Raw cost per lead | £50 | £80 |
| Cost per sales-qualified lead | £417 | £286 |
After query exclusions, clearer ad qualifiers and form controls, raw cost per lead is 60% worse. Yet sales-qualified volume is 46% higher and its unit cost is 31% lower. An optimisation process focused on the cheapest form fill would reject the better result. The guide to B2B cost-per-lead economics explains why the denominator matters as much as the spend.
Diagnose junk leads by source, not anecdote
Take the last 100 submissions with mature sales outcomes. At lower volume, use every lead from two complete sales cycles. Join each record to campaign, ad group, available search term, network, location, device, form and rejection reason. Some low-volume queries will remain hidden because Google’s search terms reporting aggregates them for privacy, so keep an “unattributed search spend” line rather than pretending the report is complete.
Then distinguish the symptoms:
| Symptom | Likely failure | First evidence to inspect | First lever |
|---|---|---|---|
| Jobseekers, students or DIY researchers | Query intent | Search terms and rejection notes | Negative keywords and intent separation |
| Right service, wrong company size or budget | Fit filtering | Company data by campaign and ad | Qualifiers and form questions |
| Enquiries outside the served market | Location settings | User-location report and form country | Presence targeting and exclusions |
| Repeated, gibberish or impossible details | Automated abuse | Submission patterns and server logs | Form protection and validation |
| Many “leads”, few actual records | Tracking inflation | Tag firing and CRM count | Conversion-action repair |
| Plausible contacts that sales will not pursue | Weak buying intent or unclear criteria | Call notes, query theme and rejection code | Shared definition, ads and offer |
One awful screenshot is not a trend. Calculate invalid rate and sales-qualified rate by campaign and query theme. High invalidity points to traffic or form abuse. High validity but low sales acceptance points to fit, offer or intent. Those are different problems and need different fixes.
Improving B2B Google Ads lead quality at the query level
A keyword is an instruction to Google; a search term is what the person typed. Match types control how far Google can interpret that instruction, not whether the query will repeat the keyword literally. That is why the search terms report, not the keyword list, is the starting point.
Use a weekly four-step review:
- 1. Label visible terms as qualified, ambiguous or irrelevant using sales outcomes, not the marketer’s opinion.
- 2. Build negative themes for genuine exclusions such as jobs, salaries, courses, definitions, templates, free tools, support and consumer variants. Apply them only where the term is disqualifying; a software company with a free trial should not blindly exclude “free”.
- 3. Move recurring qualified queries into tightly themed ad groups so their spend and acceptance rate can be judged separately.
- 4. Separate vendor-ready searches from early research, and branded demand from non-branded acquisition. A “how does payroll work” query is not inherently bad, but it should not consume the same budget as “payroll software for 100 employees”.
Exact and phrase match are useful for establishing a clean baseline. Broad match can find incremental demand, but its extra reach is an expansion decision, not a default cure for low volume. Test it in a contained campaign only when rejection coding and query reviews are already working. Campaign architecture, budgeting and bidding sit within the wider B2B Google Ads strategy.
Check location options as well as the map of targeted areas. Google’s default advanced location option can include people who show interest in a location. A firm serving only the UK usually needs presence-based targeting, explicit exclusions and a review of where users were actually located.
Segment Google Search from Search Partners. If partner traffic produces a materially lower valid or sales-qualified rate over a useful sample, switch it off for a controlled test. Apply the same discipline to devices and hours: five poor leads are not enough to erase mobile or weekends. As a working threshold, wait for roughly 30 mature outcomes in a segment or two sales cycles, whichever gives the more credible view.
Use the ad to repel poor-fit clicks
Most B2B ads try to maximise eligibility. Better ads state who the offer is for and what commitment it assumes.
“Payroll software for UK employers with 50+ staff” filters harder than “Easy payroll software”. “Managed IT from £2,000 per month” may lower click-through rate, but it prevents microbusinesses from paying you a visit neither side needs. Other useful qualifiers include sector, service area, implementation window, contract model and the problem being solved.
Use only thresholds the sales team actually enforces. Repeat critical qualifiers across enough responsive search ad assets that they do not disappear in weak combinations. Judge the change on qualified-lead cost, not click-through rate alone: a fall from 7% to 5% CTR can be a win if the sales-qualified share doubles.
The choice between ongoing Google Ads management for B2B demand and a bounded Google Ads restructuring project is secondary to the brief: it should specify accepted-lead criteria, exclusions and the cost per qualified outcome.
Improve lead quality at the form without building an obstacle course
Form length is the wrong debate. Every field and control should do one of three jobs: block abuse, establish fit or prepare sales for the conversation.
Block machines
Use layered protection: server-side validation, a honeypot, rate limiting, duplicate suppression and a challenge only when behaviour looks risky. Validate that email, telephone and website fields are structurally possible without rejecting legitimate UK and international formats. Log blocked submissions and inspect false positives; a CAPTCHA that stops real directors on mobile has not improved quality.
Fire the lead conversion only after the server accepts the submission. A button click, form start or thank-you-page reload is not a lead. Count phone calls only after a duration supported by call reviews—for example, 60 seconds if that is where genuine enquiries separate from misdials. Keep downloads, chat opens and other micro-actions observable but secondary. Google normally uses primary conversion actions for bidding, while secondary actions remain observational unless included in a custom goal.
Qualify people
Ask three to five questions that sales genuinely uses: company, role, organisation-size band, requirement, timeframe or budget band. A “reason for contact” field can route job applications, supplier pitches and support requests away from the sales form. Conditional fields let a complex buyer provide detail without making every visitor complete the longest route.
Requiring a work email can reduce consumer enquiries, but it can also exclude founders who use Gmail and buyers in sectors with weak domain-email adoption. Test it against valid and sales-qualified rates. It is not a substitute for query control or anti-spam protection.
Measure B2B Google Ads lead quality in pounds
The operating scorecard should show raw cost per lead, valid-enquiry rate, raw-to-sales-qualified rate, cost per sales-qualified lead, meeting-held rate and opportunity rate. Report them by campaign and intent group; an account average can hide a strong category campaign behind cheap branded leads.
Use 2026 B2B Google Ads benchmarks to spot an unusual CPC, click-through rate or raw conversion rate, not to define quality for your company. Likewise, ROAS ranges by industry are a sense-check. Your margin, close rate and sales capacity determine what a good lead is worth.
Suppose a won customer contributes £12,000 in first-year gross profit and the business permits £3,000 of that for media acquisition. If 25% of sales-qualified leads close, the maximum affordable cost per sales-qualified lead is £750:
£3,000 × 25% = £750
If only 20% of raw submissions become sales-qualified, the implied ceiling for a raw lead is £150:
£750 × 20% = £150
That £150 is not a target to chase. It is the ceiling created by the current funnel. Improve qualification or close rate and the allowable acquisition cost changes.
Fix the system in sequence: confirm conversion counts; clean queries, locations and traffic sources; add honest ad qualifiers; strengthen form controls; then reconsider bidding. Change one major layer at a time and wait for outcomes to mature. Once qualified stages are available as reliable campaign signals, choose the stage that reflects commercial value. The technical process for returning CRM outcomes to Google Ads is a separate implementation task.
FAQ
Why am I suddenly getting more junk leads from Google Ads?
Start with the change history and the rejection mix. A match-type expansion, new network, broader location option, altered conversion goal or weakened form control can change quality quickly. If the increase is mainly repeated or nonsensical data, inspect form abuse. If the enquiries are human but irrelevant, inspect queries, geography and ads.
Should B2B advertisers avoid broad match?
No. Broad match is a reach tool, not a quality guarantee. It is risky when the account has vague conversion goals, no rejection data and irregular search-term reviews. Establish performance with tighter matching first, then test broad match in a contained budget against cost per sales-qualified lead.
Will adding more form fields improve lead quality?
Only when the fields change routing or qualification. Company size, requirement and timeframe can help. Asking for a postal address when sales never uses it merely adds friction. Combine a small number of commercial questions with anti-spam controls, and judge the result on valid and qualified rates rather than conversion rate alone.
What is a good B2B sales-qualified lead rate?
There is no universal percentage because a £500 monthly product and a £100,000 consultancy engagement impose different thresholds. Build an eight-to-twelve-week baseline by campaign, then set a target from close rate and allowable acquisition cost. A rising raw CPL can still be healthy when cost per qualified opportunity falls.
How long does a lead-quality fix take to show?
Bot controls and tracking repairs can change counts immediately. Query, ad and form changes need enough new leads to compare. Commercial quality takes at least the time required for sales to accept or reject them, and sometimes a full sales cycle. Avoid declaring victory from the first five submissions.
Summary
- Define valid, marketing-qualified and sales-qualified leads before optimising campaigns.
- Trace fixed rejection codes back to queries, locations, networks, ads and forms.
- Use negative themes and honest ad qualifiers to prevent avoidable clicks.
- Layer form protection with only the qualification questions sales uses.
- Optimise for cost per qualified outcome, even when raw cost per lead rises.
Actualyse runs precision B2B Google Ads programmes optimised for lead quality, not click volume. Book a call to talk through where yours stands.
