What Is Invalid Traffic (IVT)? GIVT vs SIVT Explained

What is invalid traffic (IVT)?

Invalid traffic (IVT) is ad activity, such as impressions, clicks or conversions, that does not come from a real person with genuine interest and therefore should not be counted or paid. It includes bots, known data center traffic, publishers clicking their own ads, and organized fraud. The industry standard from the Media Rating Council (MRC) splits it into two groups: General Invalid Traffic (GIVT) and Sophisticated Invalid Traffic (SIVT).

IVT is a wider term than ad fraud. Some invalid traffic is harmless, like a search engine crawler loading a page, and some is accidental, like a user tapping an ad by mistake. It still gets filtered so that nobody pays for it.

The official definitions

The MRC Invalid Traffic Detection and Filtration Standards Addendum (June 2020 update, with interim updates published in April 2024) defines IVT as traffic that “does not meet certain ad serving quality or completeness criteria, or otherwise does not represent legitimate ad traffic that shall be included in measurement counts.” The MRC accredits measurement companies, so this is the definition audited vendors work to.

Google uses a shorter, publisher-facing version. The AdSense invalid traffic page says invalid traffic “includes any clicks or impressions that may artificially inflate an advertiser’s costs or a publisher’s earnings,” covering “intentionally fraudulent traffic as well as accidental clicks.” Its examples are publishers clicking their own live ads, repeated clicks or impressions from one or more users, publishers encouraging clicks on their ads, and automated clicking tools, robots or other deceptive software. The Google Ad Manager invalid traffic page uses the same wording (both checked on November 5, 2026).

Google also says that even when a third party causes it, “ultimately it is your responsibility as the publisher to ensure that the traffic on your ads is valid.”

GIVT vs SIVT

According to the MRC addendum, GIVT “consists of traffic identified through routine means of filtration executed through application of lists or with other standardized parameter checks.” SIVT “consists of more difficult to detect situations that require advanced analytics, multi-point corroboration/coordination, significant human intervention, etc., to analyze and identify.”

The line moves over time. The MRC expects some SIVT methods to become routine and move into GIVT as detection improves.

Where rewards and incentives sit

This part matters for rewards sites. The SIVT list includes “incentivized human invalid activity” and incentivized manipulation of measurement. But a footnote in the same addendum excludes “disclosed self-directed activity without the express purpose of manipulating measurement such as reward programs,” which should be reported separately as a subset of valid activity. In plain terms: a user who knowingly completes an offer for a reward is not invalid by definition. A user paid to inflate ad clicks or impressions is. Google draws the same line in its own products, where encouraging clicks on ads is invalid. Never reward anyone for clicking or viewing display ads. Incentivized traffic explained covers where incentives are allowed.

How invalid traffic is detected and removed

Detection runs in layers, roughly from cheap to expensive:

1. Lists. The IAB/ABC International Spiders & Bots List, data center IP ranges, and known bad user agents remove most GIVT.
2. Network checks. VPN, proxy and data center detection, explained in What is IP reputation?
3. Device checks. Matching devices across accounts and spotting emulators, covered in What is device fingerprinting? and Emulator and device farm detection.
4. Behavior. Click timing, conversion speed, repeat patterns and outlier rates per source.
5. Human review. Analysts confirm the hard cases, which is what makes it SIVT work.

What happens next depends on the model. On impression or click buys, filtered traffic is removed from billing, and Google notes that invalid activity can make finalized earnings differ from estimated earnings. On performance deals, invalid conversions are rejected or reversed. For a deeper look at fraud types and warning signs, read Ad fraud and bot traffic.

A worked example (illustrative)

A publisher with a game guide site and an earn page gets 50,000 sessions in a month. All numbers here are made up to show the arithmetic.

• 3,000 sessions come from a data center range and a known crawler. Filtered as GIVT: 6% of sessions.
• A paid traffic source sends 4,000 sessions with a 40% bounce in under one second and clicks timed at exact intervals. An analyst flags it as SIVT: another 8%.
• Valid traffic left: 43,000 sessions, or 86%.

On display ads, the publisher is paid on the 86%, and the bad source raises the risk of limited ad serving. On a pay-per-result offerwall, the bot sessions mostly convert nothing, and the few fake sign-ups they produce are reversed when the advertiser rejects them. Cutting the paid source is the fix in both cases.

How publishers prevent invalid traffic

Google’s tips for preventing invalid traffic come down to five habits, and they apply to any monetization:

• Segment your traffic. Break reports down by channel, ad unit, country and source so a bad source stands out.
• Avoid low-quality traffic partners. Be wary of networks or directories that promise cheap visitors.
• Never click your own ads, and ask friends and testers not to either. Use test modes.
• Check your implementation. Ads placed next to buttons or menus cause accidental clicks.
• Report and remove. If you suspect an attack, pull the ads while you investigate.

Rewards sites add account-level checks: one account per person, email verification, VPN blocking, and holds on new earnings. See Stop multi-accounting and VPN abuse and Bot and click spam on an earn page.

Common mistakes

How pay-per-result models differ

On an offerwall you are not paid for impressions or clicks at all. You earn only when a user completes an offer and the advertiser confirms the conversion. That shifts where invalid traffic hurts you. A bot can load your wall a thousand times and earn nothing, because it never finishes an offer. The risk moves to fake or low-quality completions: a sign-up that never verifies, an install that never opens, an account farm. Those get rejected or reversed after the fact, and a reversal takes back the reward you already credited if you do not plan for it.

So on pay-per-result, IVT shows up in your reversal rate rather than in a filtered-clicks column. Publisher earnings and reversals explains how to handle them, and What is an affiliate chargeback? covers the related term.

Where Moonetix fits

Moonetix is a pay-per-result affiliate network, so its publishers are paid per completed offer, not per impression or click. Users on VPNs, proxies, Tor or data center networks are blocked from the wall, offers are reviewed by the team before they go live, and every credit and reversal reaches your server as a signed postback with the same transaction ID. That does not make invalid traffic someone else’s problem: fake accounts on your side still produce reversals, so your own checks matter. Read How Moonetix checks traffic quality and the traffic quality rules.

Frequently asked questions