Local Services Ads bill per lead, which sounds like the fairest deal in advertising until the leads start arriving: the caller looking for a service you don’t offer, the homeowner three counties outside your service area, the robocall, the competitor fishing for prices, the person who found you in the carousel but wanted the plumber two slots down. Every one of them is a charge on your account — and Google’s own model acknowledges this by providing a dispute mechanism that credits legitimately unqualified leads back. The mechanism works; most advertisers just don’t work it. Accounts routinely leave meaningful money unrecovered every month, either because nobody reviews the lead log, because the dispute categories are misunderstood, or because a few rejected disputes convinced someone the process is theater.
It isn’t theater, but it is a system with rules — and, importantly, with changed rules: Google’s move toward automated lead crediting shifted how disputes are handled, what gets credited without action, and what still needs your intervention and documentation. Treating the current system like the 2021 version (dispute everything, argue every case) wastes time; treating it like a lottery (dispute nothing) wastes margin. The businesses that recover the most run disputes like accounts-receivable hygiene: a weekly review, clear category mapping, evidence habits that make legitimate cases undeniable, and — the part most guides skip — the upstream profile fixes that stop generating disputable leads in the first place.
This guide covers the whole discipline: what genuinely qualifies for credit (and what never will), how the crediting system currently behaves and where manual feedback still matters, the weekly review workflow with documentation habits that hold up, the profile-hygiene fixes that cut junk-lead volume at the source, and how to escalate when a clearly bogus charge slips through.
LSA charges per lead and credits genuinely invalid ones — but only if your operation actually flags them. Credit-worthy categories: spam/bots/solicitation, wrong service (not offered in your profile), out of service area, and certain wrong-number/misdial cases. Never credit-worthy: real prospects you lost — price shoppers, no-shows, missed calls, bad-fit-but-in-scope jobs; the charge is for the lead, not the outcome. Google’s system now auto-credits much of the obvious junk; your job shifted to weekly lead-log review, rating/feedback on lead quality, and flagging what automation missed — with call recordings and notes as evidence. Fix the upstream causes too: stale service selections, over-broad service areas, and vague business info generate disputable leads you pay for first and argue about later. Run it as a 20-minute weekly routine: review, rate, flag, document — and track your credit rate as a standing line item.
What Qualifies for Credit — and What Never Will
The dispute system exists to enforce one principle: you pay for leads that a reasonable person would call a potential customer for the services you advertise, in the area you advertise them. Everything else follows from it:
| Lead type | Creditable? | Notes |
|---|---|---|
| Spam, robocalls, bots, telemarketing/solicitation | Yes | The clearest category — and the one automation now catches most reliably |
| Service you don’t offer and don’t list | Yes | The “don’t list” matters: if the service is checked in your profile, the lead is valid |
| Location outside your configured service area | Yes | Judged against your settings, not your preferences that day |
| Wrong number / clearly not seeking any service | Usually | Misdials and accidental contacts — brief calls with no service intent |
| Duplicate of a very recent charged lead (same customer) | Sometimes | Policy treats repeat contacts within a window as one lead; flag genuine duplicates |
| Price shopper who didn’t book | No | A real prospect you didn’t convert is still a lead |
| Customer who booked elsewhere / ghosted / no-showed | No | Outcome risk is yours — that’s the pay-per-lead bargain |
| Call you missed entirely | No | Painfully: a missed call is a delivered lead. Fix the phones, not the invoice |
| In-scope job you didn’t want (too small, bad timing) | No | Selectivity is a business choice, not a billing error |
Internalizing the right column changes behavior on both sides: teams stop burning credibility disputing every lost prospect — and start reliably flagging the genuinely invalid ones they used to shrug off. If a large share of your “bad leads” land in the non-creditable rows, your problem isn’t billing; it’s intake and follow-up — the lead-quality diagnostics and speed-to-lead fixes recover more money there than any dispute ever will.
How Crediting Works Now: Automation First, You Second
Google shifted LSA toward automated lead crediting: the system analyzes leads (call patterns, durations, content signals) and credits much of the obvious junk — spam, bots, clear solicitations — without you filing anything. Credits land on the account (visible in your lead and billing history) rather than as answered dispute tickets. Three practical consequences:
- Your feedback became training data. The lead inbox’s rating and feedback options (marking leads, noting job outcomes) feed the quality systems that decide both future crediting and, indirectly, what the system learns about your fit. Rating leads consistently — good and bad — is now part of the recovery mechanism, not busywork.
- Manual flagging still matters at the margins — precisely where automation is weakest: wrong-service calls that sound plausible, out-of-area addresses mentioned mid-call, duplicates, and nuanced misdials. These are the cases where your notes and recordings decide outcomes.
- The dispute-everything era is over. Blanket disputing of lost prospects doesn’t just fail; a pattern of unfounded flags degrades your credibility with the system and wastes the attention that legitimate cases need. Precision beats volume.
Credits go to accounts whose claims are verifiable in seconds. Three habits make that automatic: keep LSA’s call recording enabled (leads arriving through the platform are recorded when configured — the recording is your entire case for wrong-service and out-of-area claims); train whoever answers to capture the two disqualifying facts early and naturally (“what do you need done, and where are you located?” — the first thirty seconds of a proper intake script doubles as dispute evidence); and log a one-line disposition note on every lead in the inbox while it’s fresh. When the weekly review runs, every flag you file carries a timestamped recording and a contemporaneous note — the difference between ‘claim’ and ‘case.’
The Weekly Workflow: Twenty Minutes That Pay Wages
- Open the lead log for the week (do not batch monthly — feedback windows and memory both decay). Sort by charged leads.
- Reconcile against your CRM/call notes: every charged lead should have a disposition — booked, quoted, lost, invalid-with-reason. Anything without one gets listened to now (recordings) and dispositioned.
- Rate every lead in the LSA inbox — the systematic feedback that trains crediting in your favor. Mark booked jobs too; the system understanding your good leads sharpens its junk detection.
- Flag the invalid ones with category-accurate reasons and the evidence pointer (recording timestamp, note). Category accuracy matters: a wrong-service call flagged as spam reads as an unfounded claim.
- Check credits received against last week’s flags and the auto-credit line; log your monthly credit rate (credits ÷ charged leads) as a standing KPI — it’s both recovered margin and an early-warning gauge of lead-mix drift.
- Feed the patterns upstream: two wrong-service calls for the same unoffered service, or repeated leads from one out-of-area town, are profile bugs (next section), not billing bugs — fix the source and the disputes stop being necessary.
Upstream Hygiene: Stop Manufacturing Disputable Leads
Every recurring dispute category has a profile-side cause — and prevention is worth more than recovery, because prevented junk never taxes your answer rate, your intake time, or your ranking signals:
- Service selections drift. The single biggest generator of wrong-service leads is a profile with services checked that you’ve deprioritized or never really offered — often checked optimistically at setup. Audit the list quarterly against what you actually want to sell; remember that a listed service makes those leads valid and billable.
- Service areas over-reach. Zip codes added “just in case” produce leads you’ll decline — at full price, non-creditable, because the area was yours by configuration. Match the map to the trucks. (The reverse error — too-narrow areas throttling volume — is part of the volume diagnostic.)
- Business details ambiguity. Vague business names, missing hours, and thin profile detail attract mismatched callers; specificity filters at the impression, before the charge.
- Booking flow mismatches. If you accept message leads or bookings, respond inside the expected windows — expired and ignored leads complicate both your ranking and your quality standing, and none of them are creditable.
The costliest dispute-program mistake isn’t missing credits — it’s using disputes as anesthesia for conversion problems. A lead log full of ‘price shopper,’ ‘went with someone else,’ and ‘never answered our callback’ describes an intake and follow-up funnel leaking real revenue: those were reachable customers, the charges are valid, and no credit will ever arrive. The tell is your booked-job rate on connected leads — if competitors book 40–50% of their LSA calls and you book 20%, the recovery opportunity is in scripts, speed, and scheduling, and it’s several times larger than your entire theoretical credit pool. Dispute hygiene and funnel repair are different programs; run both, and never let the first substitute for the second.
When a Clear Case Gets Missed: Escalation That Works
Occasionally an obviously invalid charge survives both automation and your flag. Escalate like a professional, not a plaintiff: use the LSA support channels with a tight, factual package — lead ID, date/time, category claim, one-sentence description, and the evidence (recording timestamp, note). Batch related cases (three wrong-service leads for the same unoffered service tell a systemic story a single call doesn’t). Track escalations and outcomes in the same log; if a category consistently fails despite solid evidence, the economical response is upstream prevention or budget reallocation, not a standing grievance. And keep perspective on scale: escalations are for the exceptional leak — if you’re escalating weekly, either your profile hygiene or your category understanding needs the attention first.
5 Common LSA Dispute Mistakes
- Not reviewing leads at all. The default outcome of no process is paying full price for every junk call automation missed.
- Disputing lost prospects. Price shoppers and no-shows are valid charges; unfounded flags erode credibility that real cases need.
- Wrong-category flags. Accuracy is evidence; a mislabeled claim reads as a false one.
- No recordings, no notes, no case. Evidence habits built into the call cost nothing; reconstructing them later costs everything.
- Treating recurring junk as a billing problem. Two identical invalid leads are a profile bug — fix the services/areas generating them and stop paying the intake tax.
Frequently Asked Questions
How quickly do I need to flag an invalid lead?
Treat it as a same-week obligation and you’ll never be caught by a window. Lead feedback and dispute mechanisms operate on limited timeframes (Google has used windows measured in days, and the automated-crediting era ties much of the evaluation to signals gathered around the lead itself), so a monthly batch review structurally forfeits some recoveries and degrades your evidence — recordings get harder to interpret and dispositions harder to reconstruct. The weekly twenty-minute routine exists precisely to make timing a non-issue: every charged lead gets dispositioned, rated, and (if invalid) flagged while the call is fresh and the recording is one click away. If you’re starting from a backlog, still do the historical pass — rating past leads trains the quality systems even where crediting windows have closed — but fix the cadence going forward; recency is half the case.
Google auto-credited some leads I never flagged. Do I still need a dispute process?
Yes — automation moved the floor up, not the ceiling. What it reliably catches is the signal-obvious junk: robocalls, bots, telemarketers, calls with patterns the system has seen across thousands of accounts. What it structurally can’t judge without your input: whether the plausible-sounding caller asked for a service you don’t offer (it doesn’t know your real scope beyond the checkboxes), whether the address mentioned mid-call sits outside your area, whether this ‘new’ lead is last Tuesday’s customer calling back, and every nuanced misdial. Those margins are where your recordings, notes, and category-accurate flags earn their credits — and your systematic lead rating (including marking the good ones) is what keeps the automated layer calibrated to your account rather than to averages. The right mental model: automation handles the bulk; your weekly review handles the judgment calls; skipping the second because the first exists leaves the judgment-call money on the table permanently.
What credit rate is normal — how do I know if I'm recovering enough?
There’s no published benchmark, and the honest range varies by category, market, and how clean your profile is — but the useful practice is tracking your own two numbers monthly: invalid-lead rate (leads you dispositioned as junk ÷ charged leads) and credit rate (credits received ÷ charged leads). In a well-run account the two converge: most of what you legitimately flag, plus the auto-credited layer, comes back. Warning signs in the gap: a high invalid rate with a low credit rate means either your flags are miscategorized/under-evidenced or your ‘invalid’ definition includes non-creditable lost prospects — audit a sample against the qualification table. A persistently high invalid rate itself (say, well above the 15–25% junk-mix range typical of unmanaged accounts) points upstream: service selections, service-area over-reach, or a category with a spam problem worth raising through support with batched evidence. And if both numbers are near zero while your team swears the leads are bad — the leads are probably fine, and the funnel is the patient.
Can I dispute a lead where the caller hung up after ten seconds?
It depends on what those ten seconds contained, which is why recordings decide these. Genuinely creditable versions: dead-air robocall patterns, instant hang-ups with no human engagement, obvious misdials (‘sorry, wrong number’), and solicitations that identify themselves fast — flag these with the spam/wrong-number categories and let the recording speak. Non-creditable versions that feel similar: a real prospect who hung up because the call went to voicemail, sat in a hold queue, or got a confusing greeting — that’s a delivered lead your phone experience lost, and it’s also quietly damaging your responsiveness ranking signal. The systemic view matters more than any single ten-second call: a pattern of ultra-short calls concentrated at specific hours usually means a coverage gap (calls ringing out at lunch, after hours mis-set), and the fix — call routing, missed-call text-back, honest hours in the profile — recovers far more than the disputes would. Flag the true junk; instrument the rest.
Does disputing lots of leads hurt my LSA ranking?
Legitimate, category-accurate, evidence-backed flagging doesn’t penalize you — it’s the intended use of the system, and the feedback loop it feeds exists to improve your lead mix. What plausibly does hurt is the behavior cluster around bad disputing: blanket-flagging lost prospects (unfounded claims erode how much weight your future feedback carries), letting leads sit unrated and unanswered (responsiveness is a real ranking input, and ignored message leads are a real quality signal), and the upstream sloppiness — stale services, over-broad areas — that both generates the junk and muddies what the system learns about your fit. Note the asymmetry worth internalizing: answering fast, rating everything, and flagging precisely improves your standing on every axis simultaneously — ranking, crediting, and lead mix — while dispute-heavy accounts with poor intake operations get the worst of all three. If ranking anxiety is holding your team back from flagging obvious junk, redirect the anxiety to answer rate and review velocity; that’s where LSA ranking actually lives.
How much LSA spend are you leaving unrecovered?
We’ll audit your lead log against your CRM, quantify your true invalid-lead and credit rates, install the weekly review workflow with evidence habits — and fix the profile settings quietly manufacturing the junk in the first place.
Get an LSA Account Review Start with a PPC Audit