Fake Leads, Bot Clicks and Spam Calls: How Local Service Businesses Lose Ad Budget and Rankings Data Without Seeing It
Published · Last updated:
This guide reflects Google Ads, Local Services Ads, Cloudflare and OpenAI crawler rules as of October 2026. Credit policies and bot settings change often. It is general marketing guidance, not legal advice.
Fake Leads, Bot Clicks and Spam Calls: How Local Service Businesses Lose Ad Budget and Rankings Data Without Seeing It
Fake leads, bot clicks and spam calls start as invalid traffic: bots, accidental clicks and competitors clicking to raise your spend. Google does not charge for invalid clicks it catches before billing and credits some it finds later, but that filter works on clicks. The fake leads Google Ads never flags, plus the bots that inflate your Search Console data, are the ones you have to find yourself.
- Google defines invalid traffic as traffic that does not represent genuine interest, including bots, accidental clicks and competitors clicking to raise your spend; it does not charge for invalid traffic caught before billing and credits invalid traffic found later "where appropriate and possible". [1]
- The problem is large on both sides of the auction. Google blocked or removed over 8.3 billion ads in 2025 [2], and Thales, which sells bot protection, reported that bots made up 53% of the traffic it observed in 2025. [3]
- Local Services Ads credits are automated now. Google reassesses charged leads with models, most credits post within 30 days, and leads marked "job type not serviced" or "geo not serviced" are no longer credited. [4]
- What the top-ranking fake-lead guides we reviewed do not show: an agency's own data. One Ambrose page at position 4.7 drew 32,417 Search Console impressions and zero clicks in three months, the signature of rank-tracker scraping [5], and the blocked "AI crawlers" in a 24-hour trace of our site were impostors. [6]
- Most Ambrose clients invest roughly $1,200 to $4,000 a month in marketing, often across more than one channel; the platforms, the number of markets, ad spend and scope move the number. Book a free consultation and we will walk through your Invalid clicks column and lead log with you.
Picture a plumber who sees forty form fills in a week and books six jobs. A med spa gets a run of calls that hang up after two seconds. Neither owner can tell how much of that activity was ever a person, and neither knows what the ad platform already caught. Google's own definition is a good place to start: invalid traffic covers bots, accidental clicks and competitors who click to raise your spend. [1]
This guide separates what the platforms filter and credit from what lands on your desk. It also covers what most fraud guides skip: bots that distort your Search Console data, and bot settings that block Google or ChatGPT along with the bad traffic.
Invalid Traffic Is Any Click That Was Never a Customer
Invalid traffic is any ad click or impression that does not come from genuine interest in your business. Google names three sources: non-human traffic such as bots and automated software, accidental clicks, and people who click on purpose to drain your budget, including competitors. Invalid traffic is the root of most fake leads and bot clicks a local service business pays for.
Google's definition is short enough to quote in full: "Invalid traffic refers to traffic that doesn't represent genuine interest in your business." [1] The page then lists the categories. Non-human traffic covers bots and automated software. Accidental clicks cover the thumb that lands on an ad by mistake. A third category covers deliberate clicks, and Google states plainly that this "includes competitors who manually click to increase your spend". [1]
That definition sets the boundary of what Google will act on. A click is in scope. A form fill typed by a person with no intention of hiring you is a different problem.
The scale on the platform side is large. Google reports that in 2025 it blocked or removed over 8.3 billion ads and suspended 24.9 million accounts, and that 602 million of those ads were associated with scams. [2] Those figures describe bad advertisers, not bad clicks on your ads, but they show the volume of abuse an ad platform screens every year.
The traffic side is just as busy. Thales, a vendor that sells bot protection, reported that bots made up 53% of the traffic it observed in 2025 and bad bots 40%. [3] That is one vendor measuring its own network, not the bot share of your account.
The never-a-customer test is the simplest way to sort the problem. For every click, lead or call, ask one question: could this have become a booked job? A bot cannot. A competitor cannot. A misdial cannot. A real homeowner outside your service area cannot either, even though a person made the call. The first three are invalid traffic in Google's terms [1]; the fourth is a targeting problem, and Google treats it differently when it decides credits.
Where Fake Leads Come From in Google Ads, Local Services Ads and Meta
Fake leads reach a local service business through four routes: invalid clicks on search ads, low-quality or spam leads on Local Services Ads, unverified form fills on social lead ads, and impostor businesses that take calls meant for real ones. Each route shows up in a different report, and a different party, sometimes nobody, credits you for it.
The table below sorts each fraud type by where it shows up, who credits it, and what you have to do yourself.
| Fraud type | Where it shows up | Who credits it | What you must do |
|---|---|---|---|
| Bot and competitor clicks on search ads | Google Ads clicks; the Invalid clicks column | Google: not charged if caught before billing; credited as "Invalid activity" where appropriate and possible [1] | Add the Invalid clicks column and read your invoice adjustments [1] |
| Spam or low-quality Local Services Ads leads | Local Services Ads lead log (calls, messages, booking requests) [7] | Google: automated model reassessment; most credits within 30 days [4] | Keep job types and service area accurate, because "job type not serviced" and "geo not serviced" leads are no longer credited [4] |
| Unverified social lead form fills | Meta lead ads (leads objective) [8] | No Meta credit policy is cited in this guide | Confirm every form fill by call-back before you count it as a lead |
| Impostor business listings | Google Maps and Business Profiles | Google removes listings and has sued operators [9] | Search your own trade and city on Maps for listings using your name |
| Misrepresented purchased leads | Lead marketplaces | FTC enforcement, as in the HomeAdvisor consent order [10] | Track booked jobs by lead source, not lead count |
Local Services Ads work on a different meter. You pay per lead, not per click, and Google counts a text or email, a voicemail, an answered call, a returned missed call or a booking request as a charged lead. [7]
That is where spam calls cost money. Google assesses each lead at first contact and does not charge for leads it judges invalid or low quality, but it does charge for an answered call in which you speak with the caller [7], so a wrong-trade caller or a spam call that passes that first check can become a charged lead unless Google's models later credit it. [4] The meter stays the same as Local Services Ads move into Performance Max, where Google says advertisers still pay only for valid leads. [11] Our guide to Google Local Services Ads vs Google Ads in 2026 covers that move in detail.
Meta campaigns that use the leads objective are optimized to drive leads, and LocaliQ's 2026 benchmark put the beauty and personal care cost per lead on them at $50.91. [8] No Meta credit policy for fake form fills is cited in this guide, so your own call-back is the filter.
Impostor listings are the strangest route. Google reports it removed over 10,000 fake Business Profile listings run by a group impersonating locksmiths and sued the people behind them; in December 2025 the main defendant settled and the court issued a stipulated permanent injunction. [9] A customer who calls one of those listings never reaches you at all.
In April 2023 the FTC finalized a consent order requiring HomeAdvisor to pay up to $7.2 million in redress, settling a complaint that charged it with false, misleading or unsubstantiated claims about the quality and source of its leads. [10] Measure jobs, not promises.
Not sure how many of last month's leads were real? Book a 15-minute consult and we will review your Invalid clicks column, your Local Services Ads lead log and your form fills with you before anyone talks budget.
Book Your Free Consultation →What Google Credits and What It Does Not
Google does not charge for invalid clicks it detects before billing, and it credits invalid traffic found after billing where appropriate and possible. Local Services Ads credits are automated: models reassess charged leads, and most credits post within 30 days. Leads outside your listed job types or service area are no longer credited, and health care verticals receive no lead credits at all.
On search, display and other click-based campaigns, the rule has two halves. "You aren't charged for invalid traffic detected before billing." [1] If Google finds invalid traffic after your invoice, "you'll receive a credit where appropriate and possible", and that credit appears as an "Invalid activity" adjustment. [1] The qualifier matters. A credit is not automatic for every click you believe was fake, which is why the Invalid clicks column, not your own suspicion, is the number to start from.
Many owners still expect to dispute Local Services Ads leads one by one. Google's current page describes an automated process instead. Google says charged leads "get reassessed by our models over time, and may be issued credits automatically if later determined to be low quality." [4] Timing is stated too: "In most cases, credits will be applied to your account balance within 30 days." [4]
The exclusions are where owners lose money without noticing. Google no longer supports credits for leads marked "job type not serviced" or "geo not serviced". [4] Lead credits are also unavailable for health care verticals, tax specialists and advertisers in EMEA. [4]
| Situation | Credited by Google? | Source |
|---|---|---|
| Invalid click detected before billing | Not charged in the first place | Google Ads Help [1] |
| Invalid traffic found after the invoice | Credited where appropriate and possible, shown as "Invalid activity" | Google Ads Help [1] |
| Low-quality Local Services Ads lead | May be credited automatically if the models later judge it low quality, usually within 30 days | Google LSA Help [4] |
| Local Services Ads lead outside your job types | No longer credited | Google LSA Help [4] |
| Local Services Ads lead outside your service area | No longer credited | Google LSA Help [4] |
| Local Services Ads lead in a health care vertical | No lead credits available | Google LSA Help [4] |
Fake leads Google can detect are largely Google's problem. Leads caused by a loose service area or a job type you do not really offer are now yours. An HVAC contractor that lists duct cleaning without doing it pays for those calls, because "job type not serviced" is no longer a credit reason. [4]
How Bots Distort the Data You Use to Judge Marketing
Bots do not only waste ad budget. Rank trackers and scrapers that load search results create impressions in Search Console with no human behind them, which lowers your click-through rate and makes a well-ranked page look like it is failing. On Ambrose's own site, one page at position 4.7 drew 32,417 impressions and zero clicks in three months.
We checked our own data before writing this. From July 3 to October 2, 2026, the ambrosemarketing.com Search Console property recorded 259,538 impressions and 545 clicks, and 86% of those impressions were on desktop. [5] One page at an average position of 4.7 drew 32,417 impressions and zero clicks. [5] That pattern is the signature of SERP rank-tracker scraping, which was proven on this property in July 2026. [5]
Zero clicks at that volume and position points to something reading the results page that never intended to click.
Here is the arithmetic, run on those Search Console figures. [5]
| Measure | All impressions | Without the one scraped page |
|---|---|---|
| Impressions | 259,538 | 227,121 |
| Clicks | 545 | 545 |
| Click-through rate | 0.21% | 0.24% |
One page took 12.5% of the site's impressions and none of its clicks. Remove it and the site-wide click-through rate rises by about 14%, from 0.21% to 0.24%, without a single new visitor. That calculation is ours, built from our own export: the page figure comes from the Pages report and the totals from the property chart, which Search Console aggregates differently, so treat the split as approximate. It shows how one scraped page can move a headline number. [5]
This matters to an owner who only cares about leads, because Search Console is where most businesses judge whether SEO is working. A falling click-through rate reads as a content problem. If scrapers are behind it, the decision you make from it is wrong. [5]
Three signals in Search Console are worth checking each month:
- Device skew. A share of desktop impressions far above your real customer mix, like the 86% on our property, deserves a closer look. [5]
- Impressions without clicks at a strong position. A page in the top five with thousands of impressions and no clicks is the clearest single sign. [5]
- Sudden impression jumps with flat clicks. Growth that never reaches your phone is not growth.
The fix is not to ignore Search Console. It is to read it by page and device, with the bot share in mind.
Blocking Bots Without Blocking Google, ChatGPT or Your Ad Checkers
Blocking bad bots is safe only when the real search and ad crawlers still get through. Cloudflare's Block settings for AI training now also stop Googlebot, Bingbot and Applebot from crawling for search, Bot Fight Mode cannot be skipped with custom rules, and OpenAI requires its ad crawler on ChatGPT ad landing pages. Check access by source, not by the name a bot claims.
Cloudflare's September 2026 change makes the risk concrete: its Block and Block on pages with ads settings now apply to mixed-use crawlers including Applebot, Bingbot and Googlebot, so either setting affects search as well as training. [12] Cloudflare's own advice is direct: "To stop training and keep search, use Disallow AI Training." [12]
Bot Fight Mode is the second trap. Cloudflare says it identifies traffic matching known-bot patterns and issues challenges, and "You cannot bypass or skip Bot Fight Mode using WAF custom rules or Page Rules." [13] No custom rule can carve out an exception for a crawler you need.
Ad checkers are the third. OpenAI's ad crawler, OAI-AdsBot, only visits pages submitted as ChatGPT ads, and OpenAI tells advertisers "You must allow OAI-AdsBot." [14] It recommends allowing its search crawler, OAI-SearchBot, too. [14] That search crawler is separate from GPTBot, OpenAI's training crawler, so you can block training and still allow the crawler that surfaces sites in ChatGPT search. [15] Our guide to ChatGPT ads for local service businesses covers the ad side.
Names can lie. We traced our own site to prove it. Over the 7 days to October 5, 2026, Cloudflare's AI Crawl Control logged about 8.42 thousand AI crawler requests to ambrosemarketing.com, and about 4.16 thousand received an HTTP 403 block. [6] That looks like half the AI crawler requests were shut out.
The source tells a different story, at least for the day we traced. When we traced the last 24 hours by IP address, the blocked requests using crawler names came from two cloud servers impersonating Applebot, Claude and Perplexity. [6] The real Applebot, from Apple's own 17.x address range, received clean responses, and so did Googlebot, Bingbot, OAI-SearchBot and ChatGPT-User. [6]
The same trace found one real problem. Perplexity's genuine crawler was being rate limited by the site's hosting platform, with 102 requests answered with HTTP 429 against 114 served with HTTP 200 in that day. [6] One engine simply read part of the site.
Use this checklist before you tighten any bot setting:
| Check | Safe setting | Source |
|---|---|---|
| AI training control | Disallow AI Training, not Block | Cloudflare blog [12] |
| Bot Fight Mode | Know that WAF rules cannot exempt crawlers from it | Cloudflare Docs [13] |
| ChatGPT ad landing pages | OAI-AdsBot allowed; OAI-SearchBot recommended | OpenAI [14] |
| ChatGPT search visibility | OAI-SearchBot allowed, independent of GPTBot | OpenAI [15] |
| Blocked "crawlers" | Traced by IP before you assume a real crawler was blocked | Ambrose trace [6] |
A user agent is a label anyone can type. Judge a block by where the request came from.
A Monthly Lead-Quality Audit You Can Run in an Hour
A monthly lead-quality audit takes about an hour and covers five places: the Google Ads Invalid clicks column and invoice credits, the Local Services Ads lead log and its automated credits, form fills checked by call-back, Search Console impressions by page and device, and your bot settings. The goal is one number: booked jobs per real lead.
Run these steps on the same day each month, in this order:
Read the Invalid clicks column (10 minutes)
Add it to your campaign table in Google Ads and note the filtered count for the month. [1] Then check the invoice for "Invalid activity" adjustments. [1] That is the fraud Google already caught.
Check Local Services Ads credits (10 minutes)
Compare charged leads against credits applied. Most credits post within 30 days, so review the month before last, not last week. [4]
Fix the settings Google will not credit (5 minutes)
Remove job types you do not perform and tighten the service area, because "job type not serviced" and "geo not serviced" leads are no longer credited. [4]
Call back every form fill (15 minutes)
Mark each one real, fake or unreachable. For Meta lead ads, the call-back is the only filter this guide can point to, because no Meta credit policy is cited here.
Tag spam calls (5 minutes)
On Local Services Ads, an answered call in which you speak with the caller is a charged lead [7], so note robocalls and wrong-trade calls and watch whether the automated reassessment credits them. [4]
Read Search Console by page and device (10 minutes)
Look for a page with strong position, high impressions and no clicks, or a desktop share far above your customer mix. [5]
Write the month's result as one line: real leads, booked jobs, and booked jobs per real lead.
If you sell online as well as locally, the same discipline applies to checkout fraud; see our guide to card testing and ad fraud on Shopify.
Where Is Your Lead Leak
This tool does not estimate cost and is not a guarantee of placement or citation. Source: the fraud-type table, the credit table, the Search Console and crawler sections, and the monthly audit in this guide.
Choose the symptom you see most often. The tool shows where that problem shows up, who credits it, and the first check to run, using the rules and data in this guide. It does not estimate cost and is not a guarantee of placement or citation.
Frequently Asked Questions About Fake Leads and Bot Clicks
Why am I getting so many fake leads from Google Ads?
Does Google refund click fraud?
Can bots distort my Search Console data?
Will blocking bots on Cloudflare hurt my SEO or my ads?
Are Local Services Ads spam calls and leads credited?
References
- Google Ads Help. About invalid traffic (11182074). no page date. support.google.com
- Google. 2025 Ads Safety Report. 2026-04-16. blog.google
- Thales (Imperva). 2026 Bad Bot Report. 2026-04-29. cpl.thalesgroup.com
- Google LSA Help. Lead credits (15100654). no page date. support.google.com
- Ambrose Marketing. First-party data for ambrosemarketing.com (Ambrose GSC export + domain property). 2026-10-04 export; read 2026-10-05.
- Ambrose Marketing. First-party data for ambrosemarketing.com (Cloudflare AI Crawl Control + GraphQL request analytics). read 2026-10-05.
- Google LSA Help. How Local Services Ads pricing works (7195435). no page date. support.google.com
- LocaliQ. Facebook Advertising Benchmarks 2026. 2026-09-23. localiq.com
- Google. Google affirmative litigation (Google LLC v. Asayag). 2025-04-07 (blog); injunction 2025-12-05. blog.google affirmativelitigation.withgoogle.com
- Federal Trade Commission. FTC approves final order against HomeAdvisor. 2023-04-21. ftc.gov
- Google Ads Help. Local Services Ads transition to Performance Max (17213585). no page date shown; read 2026-10-05. support.google.com
- Cloudflare blog. Have it both ways: stay discoverable in search while disallowing AI training. 2026-09-15. blog.cloudflare.com
- Cloudflare Docs. Bot Fight Mode. last updated 2026-08-03. developers.cloudflare.com
- OpenAI. Advertiser guidance for allowing OpenAI web crawlers (Help 20001243). updated 2026-09-21. developers.openai.com help.openai.com
- OpenAI. Overview of OpenAI crawlers. no page date; read 2026-10-04. developers.openai.com
Real Leads Are the Only Number That Pays
Fake leads, bot clicks and spam calls are not a single problem with a single fix. Part of it Google already filters and credits: invalid clicks caught before billing are not charged, and Local Services Ads leads later judged low quality may be credited automatically. [1] [4] Part of it is now yours, because leads outside your job types and service area are no longer credited. [4] And part of it never touches an ad at all: bots that inflate your Search Console impressions and impostors that use crawler names to look legitimate. [5] [6]
The fix is a habit, not a tool. Read the Invalid clicks column, check credits a month later, call back every form fill, read Search Console by page and device, and confirm that your bot settings stop impostors without stopping Google, Bing or ChatGPT. Then judge the month by booked jobs per real lead. This guide is one part of paid and organic search as one system, where invalid traffic is one of four leaks that raise cost per lead.
Want a second pair of eyes on your leads?
We will walk through your Invalid clicks column, your Local Services Ads credits and your crawler access, and show you which leads were real. You can also explore our paid advertising services.
Book Your Free Consultation →This article is general marketing information as of its last-updated date, not legal advice. Platform credit policies and bot settings change; confirm current rules on each platform's own help page before you act.