How to Tell If a Website Lead Is Spam

Tony CarnevaleNovember 21, 202612 min read
Inbox 112ArchivedSpam 5Deleted
Ben Sorrentino18hSpeed optimization
Alicia Grant1dSpeed optimization
Phoenix, AZ2dTapped your phone number
Priya Nathan2dEmergency repair
Rajat K2dHello, I came across your website and noticed a few issues. We offer affordable SEO packages that will get you on the first page of Google. Interested in a free audit?
Amanda Cole4dHi there, I built an AI agent just for maintenance companies, trained on your website. Would you like a demo this week?
Victor5dDear Sir/Madam, our web development agency provides website redesign services at unbeatable prices.
Legal Notice1wThis is a final notice regarding copyright infringement of images on your website.
Sandra Mills1wWe provide trained virtual assistants. Our remote staff can handle your support inbox.
The demo site's Inbox, then its Spam tab: five pitches the filter kept out of the lead count, from an SEO offer to a fake copyright notice.

The short answer

In a recent two-week count on tracked small business websites, bot visits sent one contact form for every 112 that people sent, so a junk lead is rarely a bot that browsed your site. Look for a person with a pitch or a script that skips the page: check what the message offers, whether the contact details can be real, and how the visit behaved.

What to know

  • Bots made 39% of recorded visits on tracked websites, yet visits judged to be bots sent only one form for every 112 sent by people.
  • Spam leads come in two kinds: a machine filling the form, and a person with a pitch for SEO, a redesign, virtual assistants or links.
  • Links are the clearest sign in a message. One form host found links behind 81% of the submissions its filter quarantined.
  • Ad clicks were almost all people: 1.8% of Google Ads click visits and 1.4% of Meta click visits were bots.
  • A short visit from another country raises doubt, but it should never condemn a message that asks about a real job.
  • Venbit Intelligence files suspected spam under Spam in the Inbox with the reason, out of your lead count, and Not spam puts a lead back.

To tell if a website lead is spam, check three things: what the message asks for, whether the name, email and phone could belong to a real person, and how the visit behaved before the form was sent. A customer asks about a job. Pitches offer you a service, point you to a link, and land seconds after the page opened. The tool below runs those checks on one lead you paste in, and the rest of this page explains each sign.

Check a lead

Check one lead for spam signs

Paste what the form sent, and answer three questions about the visit if you know them. Every sign the rules find gets its own row and a plain reason.

Try an example:

About the visit, if you know

From your analytics or the lead's visitor trail. Leave any of them on Not sure.

Did they read more than one page?
Did they send it within ten seconds of arriving?
Did the visit come from the country you serve?

Paste a lead to read it

The signs it finds in the message, the contact details and the visit appear here, each with what it means.

A rule check in your browser, written for this page from the kinds of signal Venbit Intelligence reads on every form. Nothing you type leaves your browser. No rule set catches every pitch, and a careful one written in a customer's voice will pass. On the tracked websites, visits already judged to be bots sent 4 forms against 447 sent by people (Venbit Intelligence Research, over two weeks in November 2026), so the message and the contact details carry most of the weight here. Throwaway domains: the public disposable-email-domains list.

The check uses the kinds of rule Venbit Intelligence applies to every form it records, rewritten for this page so I can explain each one in a sentence. It runs in your browser, sends nothing anywhere and uses no AI. No rule set catches every pitch, ours included, and a seller who writes in a customer's voice from a patient visit will get past this one.

Bot visits sent 4 forms in two weeks

39%
of recorded visits were bots over 90 days
447
forms sent by people over two weeks in November 2026
4
form sends caught from visits already judged to be bots
Venbit Intelligence Research, websites tracked in the Venbit Approved System

I expected bot traffic to be where junk leads come from. Two in five recorded visits were machines, and that looked like the obvious place to start. Across the two weeks we counted forms, those machines sent 4 of them, while people sent 447. That is one bot send for every 112 from people.

Most bots on a small business website have no use for a contact form. They are crawlers reading pages for search engines, AI assistants and ad networks, and scrapers copying text. The few built to fill forms try to pass as people, and the tracker catches them either at the moment they send or a few seconds later, when a headless browser gives itself away on the thank-you page.

Junk in a contact form has two likely senders: a person with something to sell, or a script that posts to the form without opening the page.

The second sender leaves no visit behind. A script can post straight to the address a form submits to, skipping the website, so no analytics tool records it and the only trace is the email your form sends you. When your mailbox holds junk your analytics never saw, a direct post is one route and a vendor form inside a frame is another. The guide to how much of your website traffic is bots covers the visits the tracker does see.

Spam leads come in two kinds

A machine filling the form

  • Random letters in the name or the message
  • The same text in two boxes, such as the phone number typed into the message
  • Three or more links in a short message
  • Link code or template braces left in the text
  • Sent within seconds of the page opening

A person with a pitch

  • SEO, ranking and backlink offers
  • Website redesign offers, with a free mockup
  • Virtual assistants and offshore staff
  • Link sellers and guest article offers
  • AI agents, chatbots and lead generation
  • Fake notices about copyright, a domain or your Google listing

The machine family is the easier one. Its signs are mechanical, a hidden form field and a time check are built for it, and no customer is lost when it goes. The second family costs more time. These are people working through lists of small business websites, writing full sentences, passing every captcha, and using your form because it is the one inbox on your site they can reach.

Some of them sell the very thing they are doing to you. One of the pitches the Venbit Intelligence filter was taught to catch offered software that sends one message through thousands of contact forms, and it arrived through a small business's contact form.

Read the message for an offer

A customer asks about your work, and a pitch tells you about theirs.

  • Links

    A customer describes a job in words. One link is worth a second look, and three in a short message is a pitch or a script. Links to Google Drive, Dropbox or iCloud are how customers share photos, so leave those out.

  • An offer in place of a question

    The message names something it wants to sell you: rankings, a redesign, staff, articles, funding or an AI agent.

  • Who the sentences are about

    Customer messages describe their own house, office or deadline. A pitch is written about your website, your rankings and your reviews.

  • Letters no person types

    A name or message made of mixed capitals with almost no vowels is a script filling a required box.

  • The same text twice

    The phone number pasted into the message box, or the name repeated as the message, is a tool putting one value into every field it must fill.

  • A move to a chat app

    A request to carry on over WhatsApp or Telegram is a seller moving you off your own records.

Links are the one sign with a published number behind it. Splitforms, a form host, released statistics on 993 forms from 3 July to 30 September 2026. Its rule-based filter quarantined 23.1% of the 21,050 submissions that reached storage, and links or suspicious URLs were the most common flag, on 81% of the quarantined ones.

Read the second half of that study before the 23.1% alarms you. On the 183 forms with 20 or more submissions, the median spam share was 1.4%, and about three quarters of all the forms received no quarantined spam at all. Splitforms describes its own figure as a floor, because spam stopped by a hidden field or a captcha never reached storage. The pattern is lopsided: spam piles onto some forms and leaves most of them alone.

Check the name, email and phone

Contact details catch the leads the wording misses. Each check below is cheap, and none of them proves anything alone, because real people mistype, hide behind masking addresses and keep old area codes. Count the signs and decide on the pile.

DetailA sign of spamNot a sign by itself
EmailAn address from a throwaway mail serviceA Gmail, Yahoo or Outlook address
EmailA free address named for a marketing trade, selling a company's serviceAn address on the sender's own company domain
EmailA link in the message to the domain of the sender's own addressA link to photos on Google Drive, Dropbox or iCloud
PhoneAll one digit, a run such as 1234567, or a 555 numberAn area code from another state
PhoneToo few digits to dial, or more than any number hasNo phone number at all
NameA web address or a phone number in the name boxA first name only, or a nickname
NameRandom letters no person is calledAn unusual surname
What the contact details say. Each row pairs a sign of spam with something that looks similar and is not a sign by itself.

Free email is where owners go wrong. Most customers write from Gmail or Yahoo, so the provider tells you nothing. What counts is a free address beside a message that sells a company's service, and a free address named after the trade it sells, such as one built around the letters SEO.

For phone numbers, count the digits. A US or Canadian number has ten, or eleven with the leading 1. Fewer than seven cannot be dialled, and a string of one repeated digit is a placeholder typed to get past a required box. A different area code means nothing, since people keep their numbers when they move.

Check the visit behind the lead

The visit is the evidence the form itself cannot give you. Before someone writes to you, they arrive from somewhere, read some pages, spend some time on the site and sit in some country. A person shopping for a job reads before writing. Form-filling scripts and sellers working a list land on the contact page, paste, and leave.

Austin, TXOrganic searchBecame a lead3s ago
ComputerSafari3 pagesFirst visit · Started 2:14 PM, 4m ago
Home Page0:38Scrolled 75% downClicked “See care plans”
Website Maintenance Pricing & Care Plans1:52Scrolled 100% downClicked “Get a free quote”
Contact Usjust openedReading
Sent the quote request
Left the site2:18 PM
A real lead's visit on the demo site: a Google search from Austin, Texas, then the home page, the pricing page and Contact Us, where the form was started and sent about four minutes in.
  • More than one page

    Moving from the home page to services and pricing before the form is what shopping looks like. One page with the form on it adds weight only beside other signs.

  • Under ten seconds

    Nobody reads a page and types a message in ten seconds. A send that fast was pasted or scripted. The guide on how long visitors take to contact a business shows what normal looks like.

  • The country you serve

    A local business hears from few people abroad. People travel and use VPNs, so a foreign visit raises doubt without settling anything. See traffic from China and Singapore for the countries bots favour.

One rule matters more than the three facts. The visit can raise a doubt, but it cannot convict a message that reads like a customer. People paste from their notes, browse on a work VPN and fill forms in a hurry on a phone, and a real lead lost to a rule costs more than a pitch that gets through.

Venbit Intelligence follows the same rule. What the visit looked like is weighed together with the wording, and a message in a customer's voice is never filed as spam on the visit alone. Every lead in the Inbox carries its visit as a visitor trail, the same view Feed and Analytics gives every visit to the site.

Ad clicks were almost all people

Share of visits that were bots

  • All recorded visits39%

    Every visit that ran the page script

  • Visits from a Google Ads click1.8%

    Fewer than 1 in 50

  • Visits from a Meta ad click1.4%

    About 1 in 70

Venbit Intelligence Research, tracked websites, over 90 days. An ad click is a visit that arrived carrying the ad platform's click id.

When junk leads arrive from an ad campaign, the first suspect is bots clicking the ads. Our data points elsewhere. Of the visits that arrived with a Google Ads click id, 1.8% were bots, and on Meta the share was 1.4%, against 39% across all recorded visits.

So a junk lead from an ad is no proof that machines are clicking your ads. Read the message and the contact details first. Sellers click ads too, and a pitch typed by a person who arrived on a paid click is still a pitch. The guide on why ad clicks do not match website visits covers the gap between the clicks you pay for and the visits you record.

What to do with a spam lead

  1. 1

    Run the three checks before you pick up the phone

    Message, contact details, visit. Paste the lead into the tool above, or read it against the signs on this page. Two minutes here can save a conversation with a seller.

  2. 2

    Leave a pitch unanswered

    A reply starts a sales thread and confirms that a person reads this inbox. File it as spam and move on.

  3. 3

    Keep spam out of your lead count

    A pitch counted as a lead makes your website look better than it is and every lead look cheaper. Take spam out before you judge a month, and before you compare your conversion rate with anyone's.

  4. 4

    Never block a whole country or email provider

    Real customers write from Gmail, travel abroad and browse through VPNs. A rule that blocks a provider or a country loses leads you never hear about.

  5. 5

    Add a hidden field and a time check to the form

    Both are built for the machine family: a hidden field only a script fills, and a send that arrives faster than anyone can type. Neither does anything against a person with a pitch, which is the job of a filter that reads the message.

See spam leads in Venbit Intelligence

Venbit Intelligence reads every form send as it lands. A send that reads as a pitch or as junk goes to the Spam tab of the Inbox, stops counting as a lead, and skips the push to your iPhone and the win email. Nothing is deleted, and one tap on Not spam brings a lead back to the Inbox.

Inbox 112ArchivedSpam 5Deleted
Rajat K2dHello, I came across your website and noticed a few issues. We offer affordable SEO packages that will get you on the first page of Google. Interested in a free audit?
Amanda Cole4dHi there, I built an AI agent just for maintenance companies, trained on your website. Would you like a demo this week?
Victor5dDear Sir/Madam, our web development agency provides website redesign services at unbeatable prices.
Legal Notice1wThis is a final notice regarding copyright infringement of images on your website.
Sandra Mills1wWe provide trained virtual assistants. Our remote staff can handle your support inbox.
An SEO pitch in the demo site's Spam tab: it came straight to Contact Us, read 1 page in 3 seconds from Ashburn, Virginia, and Venbit's note says why it was filed.

Each filtered lead carries a note from Venbit naming the reason, such as Filtered: mentions SEO or Filtered: fake copyright threat, so you can see why it moved. The filter's strength is a setting in Settings, under Leads, in Spam protection. Standard is the full filter and the default, Moderate takes only the most obvious spam, Lite takes junk only, and Off filters nothing.

Beside it sits Ignore forms nobody typed in. A send with no typing behind it is the mark of a script, and with the switch on it stops counting and waits under Spam. The visit behind a lead counts as well: when the tracker judges that visit to be a bot, the lead leaves the Inbox, even after it arrived, because a headless browser can look clean as it sends and give itself away on the thank-you page twenty seconds later.

WebsiteMaintenance.com
Last 30 days
344visits
Portland, OROn Pricing
Bothell, WAOn Website Repair
Everett, WAOn Care Plans
Seattle, WAOn Home
Bot trafficBots are recorded but hidden from every number by default.
Datacenter browserScript · Ashburn, VA
No user agentScript
GooglebotGoogle · Search engine
GPTBotOpenAI · AI crawler
AhrefsBotAhrefs · SEO tool
Visits reaching the demo site: people join the visit count, and bots are named and set aside under Bot traffic, out of every number.

Sends from visits already known to be bots wait in a folded Verified spam section at the foot of Spam, and none of them count as a lead. When a rule is added for a new kind of pitch, the filter reads the leads already sitting in your Inbox again, and one that slipped through last month moves to Spam with nobody touching it. A lead you marked yourself stays where you put it.

There is one limit to state. The filter works on what Venbit Intelligence records, so if your form also emails you, that email still arrives with the pitches in it, and a script that posts to the form without opening the page leaves no visit to judge. Work from the Inbox, where the pitches are already set aside, and treat the email copy as a backup.

How we measured this

The research figures come from Venbit Intelligence Research: websites tracked in the Venbit Approved System. Visit and bot counts cover 21 August to 18 November 2026, and the form counts cover the last two weeks of that window, where a form counts as sent when the tracker saw the send. A bot is a visit the tracker's rules judged to be automated, and an ad click is a visit that arrived carrying the ad platform's click id.

The tool's rules were written for this page. They follow the kinds of signal the product reads and leave out its word lists and weights on purpose, since a published list is a map for getting around it. The throwaway-domain check uses the public disposable-email-domains list, and nothing you paste into the tool leaves your browser.

See which leads were spam on your own site

Venbit Intelligence reads every form as it lands, files pitches under Spam with the reason, and keeps them out of your lead count. One line to install, free for 14 days.

In Venbit Intelligence

More website analytics guides

Every question in this series, from Website Analytics for Small Business Owners.

Leads inbox4
Ad performance4
SEO tracking3
Visitor behaviour4
Feed and notifications3
Goals and measurement4
Privacy, AI and teams3
Venbit Intelligence Research19

Tony Carnevale

Founder, Venbit

Tony started Venbit in Mill Creek, WA in 2011 and still does the work: web design, SEO, and the website analytics product the company builds and runs for its own clients. The numbers and screens in his articles come from that product, not from a content mill.

Questions & Answers

  • An offer of SEO, a redesign, staff or links in place of a question about a job points to a pitch. In the contact details, look for a throwaway address, a placeholder phone number or random letters in the name. A visit of seconds on one page from abroad adds weight. On tracked websites, bot visits sent 4 forms in two weeks, so start with the wording.

  • Any public contact form can be found by sellers and scripts that work through lists of business websites. On tracked websites, visits judged to be bots sent 4 forms in two weeks against 447 sent by people, so the junk that gets through comes from people with a pitch, from tools that act like people, or from scripts posting to the form without loading the page.

  • Some do, and far fewer than bot traffic suggests. Bots made 39% of recorded visits on tracked websites, yet visits already judged to be bots sent 4 forms in two weeks while people sent 447. Most bots read pages and leave. The ones that fill forms try to look human, and a hidden form field plus a time check are built to stop them.

  • Not in our data. Of the visits that arrived with a Google Ads click id, 1.8% were bots, and 1.4% for Meta ads, against 39% across all recorded visits on tracked websites. A junk lead from an ad campaign is no proof of click fraud. Read the message and the contact details before you blame the clicks, because sellers click ads too.

  • A captcha stops part of the machine family, the scripts that fill a form with no person behind them. It does nothing against a person typing a pitch, because that person passes the test. In our data, bot visits sent 4 forms in two weeks against 447 from people, so the message itself needs reading. A hidden field and a time check cost visitors nothing.

  • It depends on the form more than the business. Splitforms counted 993 forms: 23.1% of stored submissions were spam, but on forms with 20 or more submissions the median share was 1.4%, and about three quarters of forms got none. On tracked websites, visits judged to be bots sent 4 forms in two weeks against 447 from people.

More guides with a tool

See all 23 guides