What is invalid traffic? Invalid traffic (IVT) is any ad impression or click that comes from a non-human or fraudulent source rather than a real person who might buy something. Think bots, crawlers, click farms, hijacked devices, and data-center traffic dressed up to look like a customer. It splits into two grades: general (GIVT), the easy-to-spot stuff, and sophisticated (SIVT), the kind built to fool you. Either way, it burns budget and reports a "result" that never happened.
What is invalid traffic, in plain English?
Every time your ad loads or gets clicked, somebody pays. The whole model assumes a human is on the other end. Invalid traffic breaks that assumption. It is traffic generated by software or fraud schemes that mimic a real audience well enough to trigger an impression or a click, so the platform logs activity, the dashboard shows numbers, and your spend goes out the door for an audience of zero.
Here is the part agencies tend to skip past: IVT is not a rare edge case you only worry about at enterprise scale. It is baked into the open programmatic ecosystem, and the bill is enormous. Industry estimates for 2025 global ad fraud losses range from roughly $32.6 billion on the conservative end to far higher depending on how "fraud" versus "wasted spend on invalid traffic" is defined, and the trajectory is up, not down, as AI-driven buying expands the attack surface. The exact number is genuinely contested. The direction is not.
The reason this matters for your account is simple. IVT does not just waste money. It poisons your data. Bots inflate your click-through rate, crater your conversion rate, and feed garbage into the optimization engines that decide where your next dollar goes. Modern campaigns lean hard on automated bidding, and those algorithms are only as smart as the signals you feed them. So you are not just paying for fake traffic once. You are training your campaigns to go find more of it, then paying again to reach the same bots tomorrow.
The two grades of IVT: GIVT vs SIVT
The advertising industry, through the Media Rating Council (MRC) and IAB, splits invalid traffic into two official tiers. The distinction matters because they require completely different defenses, and because the cheap kind and the expensive kind hide in completely different places.
General Invalid Traffic (GIVT)
GIVT is the obvious stuff. It is identified through routine, list-based filtration: known bots, spiders, and crawlers; data-center traffic; non-browser user agents; and pre-fetched or pre-rendered impressions. Per the MRC's definitions, GIVT is what you can catch with standard published lists and basic checks. Most reputable platforms filter the bulk of GIVT before it ever hits your invoice. It is table stakes, not a selling point. If an agency brags about blocking GIVT, they are bragging about remembering to lock the front door.
Sophisticated Invalid Traffic (SIVT)
SIVT is the expensive problem. This is invalid traffic engineered to look human, and it only gets caught through advanced analytics, multi-point corroboration, and sometimes manual human review. It includes hijacked devices and ad tags, malware and adware, click farms, device emulators, cookie stuffing, domain spoofing, and bots that mimic real browsing behavior using residential proxies. Because SIVT actively tries to pass as a person, it is harder to detect and represents the most financially damaging slice of the fraud problem. When someone says "ad fraud," they almost always mean SIVT. It is also the part that survives a casual platform filter and quietly shows up as "engaged traffic" in your dashboard.
| GIVT (General) | SIVT (Sophisticated) | |
|---|---|---|
| What it is | Known bots, crawlers, data-center traffic, pre-fetch | Spoofing, malware, click farms, device emulation, hijacked sessions |
| How it is caught | Routine, list-based filtration | Advanced analytics, corroboration, human review |
| Detection difficulty | Low (mostly filtered automatically) | High (built to look human) |
| Financial damage | Lower | Higher, the costly part of fraud |
Where IVT comes from (and why it keeps showing up)
Invalid traffic is not random. It exists because somebody profits from it. Follow the money and the same sources turn up again and again:
- Bots and automated scripts. Software that loads pages and clicks ads at scale, often to drain a competitor's budget or pad a fraudulent publisher's revenue.
- Made-for-advertising (MFA) sites. Low-quality, ad-stuffed pages that exist only to harvest programmatic spend. MFA placements have exploded, and they are a primary distribution channel for invalid impressions.
- Click farms and device farms. Rooms full of real (or emulated) devices generating engagement for pay.
- Domain and app spoofing. Fraudsters misrepresent garbage inventory as premium inventory so it sells at premium prices.
- Hijacked devices and malware. Real consumer devices infected and quietly puppeteered to load ads in the background, with the owner never the wiser.
The open programmatic supply chain is where most of this hides. Between you and the actual ad slot sit layers of ad exchanges, supply-side platforms, and resellers, and each hop is a place where a fraudster can relabel junk as premium. The less visibility you have into where your impressions run, the more room IVT has to operate. That is not a coincidence; opacity is the business model.
Who detects and filters invalid traffic
IVT filtering happens at several points, and no single one is bulletproof on its own:
- Ad platforms (Google, Meta, the major DSPs) filter a large share of GIVT and some SIVT automatically, and credit back invalid clicks they catch after the fact.
- MRC-accredited verification vendors (the IAS / DoubleVerify / Pixalate tier) specialize in SIVT detection and independent measurement. MRC accreditation specifically requires demonstrated SIVT detection, not just GIVT, which is why "verified by a third party" carries more weight than "the platform says it is fine."
- Your DSP settings. This is where a competent buyer earns their keep: pre-bid fraud filters, inventory allowlists, and excluding suspect domains before the bid is even placed. Filtering at the demand-side platform is cheaper and cleaner than clawing money back afterward, because a credit returns the cash but never returns the polluted conversion data it left behind.
IVT is also closely tied to two neighboring concepts worth knowing: viewability (was the ad even in a position to be seen by a human) and brand safety (did it run next to content that protects or wrecks your reputation). Clean traffic, viewable placements, and safe environments are three legs of the same stool. A buyer ignoring one is usually ignoring all three.
How to prevent invalid traffic
You will not get IVT to zero. Anyone promising that is selling. You can drive it down hard, and the gap between a sloppy buy and a tight one is real money. The practical levers, roughly in order of impact:
- Buy through transparent supply paths. Fewer hops between you and the impression means fewer places for spoofed inventory to hide. Supply-path optimization is not jargon, it is fraud reduction with a fancy name.
- Filter pre-bid, not just post-bid. Block known-bad domains, MFA sites, and data-center IPs before you pay, so the money never leaves the account.
- Use allowlists for high-stakes spend. Run premium budget against vetted, named inventory rather than the open auction's long tail, where the worst of the fraud lives.
- Layer independent verification. Platform filtering plus a third-party MRC-accredited verification partner catches more SIVT than either does alone, and gives you a number the platform cannot grade itself on.
- Watch your own data for tells. Impossible click-through rates, conversion rates that collapse, traffic spikes from odd geographies or at 3 a.m., engagement with zero downstream action. These are IVT fingerprints, and they show up in your own reports before any vendor flags them.
- Demand reporting that shows the filtration. If your agency cannot tell you what percentage of your traffic was flagged invalid and what they did about it, that silence is the answer.
This is exactly the kind of unglamorous work that separates a real programmatic advertising operator from a spray-and-pray media buyer. It does not show up in a flashy case study. It shows up in your cost per acquisition, and it is why we are blunt about what programmatic costs before a dollar goes live.
Stop paying for an audience that isn't there
Invalid traffic is the quiet tax on lazy media buying. The agencies that don't talk about it usually aren't doing much about it. We filter it pre-bid, verify it independently, and show you exactly what we caught, because you should never have to take our word for where your money went.
Want a straight read on how much of your current spend is real? Email us at admin@moonsauceagency.com or book 30 minutes. No hard sell, just straight answers, and we'll tell you the truth even if it means fewer bots and a smaller invoice.
Keep reading: What is viewability · What is brand safety · What is a demand-side platform (DSP) · Browse the full glossary
