Publisher Guide: Compliance

Bot and fraud traffic: how networks detect it

This guide explains, in general terms, how lead networks and buyers detect bot and fraud traffic, and what that means for an affiliate who wants every lead to count. By the end you will be able to recognize the patterns that make traffic look invalid, check your own campaigns for them, and set up a routine that catches problems before they reach your reports.

Fraud often enters a campaign without the affiliate knowing: a bad placement, a bot clicking ads, or a script on a partner site. Knowing how detection works helps you find it first.

Published 5 October 2026. How we write and check guides: editorial standards.

Apply as a publisher

Already generating leads on your own forms? Sell your leads instead.

What counts as invalid traffic

A lead is only worth something if a real person, who is interested in the offer, filled in the form and agreed to be contacted. Invalid traffic is anything that breaks that: automated clicks and form fills, real people submitting someone else's details, the same person submitting many times, or leads built from data that did not come from the form. Ad platforms use the same idea for clicks. Google, for example, describes invalid clicks as clicks on ads that do not represent genuine interest in the business. Lead networks apply the same test one step later, to the form fill.

Invalid traffic matters for consent too. TCPA consent captured on the lander only means something when a real person gave it. See TCPA basics for lead gen affiliates. The FTC has also looked at practices in the lead generation industry, including at its Follow the Lead workshop.

How detection works, stage by stage

Networks and buyers check traffic at several stages. No single signal proves fraud. Detection looks for combinations of signals and for patterns across many leads from the same source.

Where networks look for invalid traffic, in general terms
Stage	What is checked	What it can look like in your traffic
Click	Network origin of the visit, browser and device details, and whether they are consistent with each other	Many clicks from server networks rather than home or mobile connections, or devices that report impossible combinations
Session	How the person moved through the page: scrolling, typing, time on page, focus changes	Forms filled faster than a person could read them, or every session taking exactly the same time
Form data	Whether the phone number and email are valid and reachable, whether names and addresses make sense, and whether they match the visit location	Disconnected numbers, nonsense names, or a stated address far from where the visit came from
Duplicates	Whether the same person, device or details appear more than once	The same phone number on several leads within a short time, or one device submitting many different people
Consent record	The consent certificate that records what happened on the lander	Sessions where the form was completed without the page behaving as a real visit would
After the sale	Whether buyers can reach the person and whether the person remembers asking	Low contact rates, "I never filled out a form" responses and higher returns from one sub ID

Because leads are reported by tracking link, patterns show up at the level of your tracking link and sub ID. A problem in one placement can affect how all of that sub ID's leads are judged.

How fraud reaches honest campaigns

Bad placements. Native ad networks and display placements can include sites with automated traffic. One bad site can produce a large share of your clicks.
Bots that click ads. Some bots click ads to make a site's ad inventory look busy. They may also fill in forms.
Bought traffic you cannot see. If someone sells you clicks, you inherit whatever is inside them.
Form-filling scripts on partner pages. A pre-lander or partner site with the wrong script can interact with forms in ways a person would not.
Your own testing. Test leads with made-up details look exactly like fraud to a buyer.

None of these require bad intent from you. That is why networks judge traffic by its patterns and why the steps below focus on visibility: if you can see where each click came from, you can remove a bad source before it shows up in returns.

Steps to keep your traffic clean

01
Step 1: Know where every click comes from
For each campaign, write down the platform, the placements or sites, and who controls them. If you buy traffic from another party, ask for placement-level detail before you send it to your tracking link.
02
Step 2: Use the platform's own controls
Ad platforms filter some invalid clicks themselves. Read your platform's help pages on invalid traffic, turn on the placement and site reports, and exclude placements that send clicks but no leads.
03
Step 3: Split traffic with sub IDs
Give each placement group, site group or audience its own sub ID. When a pattern appears, you can find the source in minutes rather than pausing a whole campaign. See tracking links and sub IDs explained.
04
Step 4: Watch the ratios
For each sub ID, compare clicks to leads, leads to sold leads, and sold leads to returns. A sudden change in any ratio, or one sub ID far from the others, is the first sign of a problem.
05
Step 5: Look at timing
Plot leads by hour. Real people follow daily patterns. Bursts at odd hours, or perfectly even spacing, suggest automation. The day-parting guide shows what normal patterns look like.
06
Step 6: Never send test leads
Do not submit made-up details on a live lander to check a link. Click your tracking link to check the redirect, and email us if you need to test a full submission.
07
Step 7: Pause first, then ask
If you see a pattern you cannot explain, pause the sub ID and email us with the dates and what you saw. Finding it yourself is better than finding it in returns.

Have clean traffic and the reporting habits to keep it that way? Apply as a publisher.

Apply as a publisher

Traffic quality checklist

You can name the source and placement behind every click.
Each placement or site group has its own sub ID.
You exclude placements that send clicks with no leads.
You review clicks, leads, sold leads and returns by sub ID on a fixed schedule.
You check lead timing by hour for bursts or even spacing.
No one on your team submits test leads on a live lander.
You pause and report patterns you cannot explain.

What happens to invalid leads

Leads that look invalid are less likely to sell, and unsold leads are not paid. Leads that sell and are later found to be invalid can be returned, and returns are deducted under the return terms agreed during onboarding. Every publisher is reviewed before going live, and traffic quality is part of every publisher relationship after that. Read why leads get returned and our traffic guidelines.

The cost goes beyond the leads themselves. Buyers bid on leads in real time through ping/post, and they bid with the quality of past leads in mind. Clean traffic keeps your leads attractive to buyers. See how ping/post affects affiliate payouts.

Common mistakes

One sub ID for everything. Without a split, one bad placement makes all your traffic look suspect.
Judging by click volume. A placement with many cheap clicks and no sold leads is a warning, not a win.
Assuming the platform filters everything. Platform filters catch some invalid clicks. They do not judge lead quality.
Testing with fake details. It looks like fraud and wastes buyer time.
Waiting for returns to find problems. By then the leads have already been sold and contacted.

Frequently asked questions

Will one bad lead get my account flagged?

Detection looks at patterns, not single leads. A few bad leads happen to everyone. A pattern from one sub ID, such as many unreachable numbers or duplicates, is what draws attention.

How do I know if a native placement is sending bots?

Compare each placement's clicks to leads and sold leads. A placement that sends many clicks but few leads, or leads that rarely sell, deserves a closer look. Exclude it and watch whether your other numbers improve.

Can I test my tracking link without submitting a lead?

Yes. Click the link and confirm it opens the right lander. To test a full submission, email us first rather than entering made-up details.

Is SEO traffic safe from bots?

Bots crawl and sometimes submit forms on every kind of site, including SEO and rank-and-rent sites. Watch the same ratios and timing patterns by sub ID. A sudden rise in leads with no rise in real search visits is worth checking.