How much does ad fraud cost advertisers each year?

Ad fraud cost advertisers an estimated $84 billion in 2023 — roughly 22% of all digital ad spend — according to Juniper Research, and the same firm projects losses climbing to $170 billion a year by 2028. The figure that should worry a media buyer more: the average invalid traffic (IVT) rate measured across the industry sat at 20.64% in 2026 Fraudlogix data, meaning roughly one in five ad interactions your tracker logs may never have come from a real person.

TL;DR
  • Ad fraud cost advertisers $84 billion in 2023, about 22% of global digital ad spend (Juniper Research).
  • Juniper projects yearly fraud losses reach $170 billion by 2028.
  • Average invalid traffic rate in 2026: 20.64% of ad interactions (Fraudlogix).
  • Programmatic waste runs at 24.5% of spend (ANA Transparency Study).
  • Mobile ads carry the highest fraud rate, around 30% (Juniper Research).

How much does ad fraud cost advertisers each year?

The headline numbers, with their sources and scope:

Measure Figure Source
Global ad fraud losses, 2023 $84 billion (22% of ad spend) Juniper Research
Projected yearly losses by 2028 $170 billion Juniper Research
Average invalid traffic (IVT) rate, 2026 20.64% of interactions Fraudlogix dataset
Programmatic ad waste 24.5% of spend ANA Transparency Study
Fraud rate on mobile ads Around 30% Juniper Research

These are not three versions of the same number. The $84 billion is money stolen outright — clicks and impressions no human ever saw. The 20.64% IVT rate is the share of traffic a typical advertiser's tracker flags as invalid in 2026. The ANA's 24.5% is estimated waste across programmatic buying, which includes fraud plus other leakage. Read together, they put "what fraud costs you" somewhere between a fifth and a quarter of spend before you fix anything.

What that means for a single campaign

Benchmarks only bite when you translate them to your own budget. At the 2023 loss rate of 22%, a campaign spending $10,000 a month is funding roughly $2,200 a month of bot clicks and unseen impressions. At the 20.64% IVT rate, that is about 2,000 of every 10,000 logged interactions. Neither number is a verdict on your specific account — fraud concentrates by offer, GEO and channel — but it sets the default assumption you defend against rather than ignore.

Mobile deserves its own line. Juniper Research measured a fraud rate around 30% on mobile ads, driven by tactics like SDK spoofing that slip past standard filters. If your tracker shows an unusually clean mobile funnel that never converts, that mismatch is a classic spoofing signature, not good news.

Where the money actually goes

Four mechanisms eat the $84 billion (2023) on the way to the projected $170 billion (2028):

  • Bot traffic and AI agents. Bots now mimic scrolling, pausing and form-filling to pass basic filters and pollute the conversion data your bidding algorithms learn from.
  • Click farms and incentive fraud. Human clickers drain CPC campaigns and poison CPA funnels with conversions that never pay out.
  • SDK and API spoofing. Mobile attribution gets faked at the measurement layer, so stolen installs look like legitimate ones in your reports.
  • Stacked and hidden ads. Several ads layered into one slot; only the top one is ever visible, yet every layer bills.

Why ad fraud numbers vary so much between studies

You will see "ad fraud costs" quoted anywhere from tens to hundreds of billions. Three reasons:

  • Definition. Some studies count only detected invalid traffic; others estimate undetected fraud on top of it.
  • Scope. Global estimates, programmatic-only studies and single-channel snapshots measure different worlds.
  • Year and method. Juniper's $84 billion (2023) came from analyzing more than 78,000 datasets of ad activity; a 2026 snapshot like Fraudlogix's 20.64% IVT rate measures observed traffic, not realized losses.

Treat any single figure as a floor or a midpoint, never a ceiling, and judge protection by your own tracker data rather than a report average.

How to keep your share of that cost down

You cannot shrink the global number, only your slice of it. The controls that move a campaign's effective fraud rate:

  • Filter at the source. Run traffic on networks that filter bots before they bill. EZmob campaigns can run ClearTrust bot monitoring that blocks fraudulent traffic before it reaches your stats — see how to automatically block ad fraud on EZmob campaigns with ClearTrust.
  • Verify conversions server-side. Postback-based tracking catches bots that never load your thank-you page. How to track EZmob conversions in Google Analytics 4 covers the measurement side.
  • Watch funnel mismatches. High clicks with zero scroll depth, or healthy click-through rates feeding a dead conversion rate, are fraud signatures — an optimization problem never fixes them.
  • Cap frequency and budget new GEOs separately. Fraud clusters. A small test budget per new GEO caps how much a bad zone can steal while you measure it.

EZmob runs self-serve pop, push, display and native campaigns through the same kind of discipline — set your own bids, cap frequency, and let a bot filter stand between the auction and your invoice.

Related questions

Is ad fraud getting worse in 2026?

By projection, yes: Juniper Research puts 2023 losses at $84 billion and expects roughly $170 billion a year by 2028, with AI-powered botnets and autonomous agents driving the steepest part of that curve.

What percentage of ad spend is wasted on fraud?

Between 20% and 25% in recent measurements: 22% of ad spend was lost to fraud in 2023 (Juniper Research), the average IVT rate measured 20.64% (Fraudlogix, 2026), and programmatic waste ran at 24.5% (ANA Transparency Study).

How do you know if your campaign has bot traffic?

Compare click quality to outcomes. Real traffic shows scroll depth, time on page and conversions that scale with clicks; bot traffic shows clicks without engagement and conversion data that stays flat or jumps unnaturally. A bot filter on the campaign plus server-side postbacks makes the split visible within days.

FAQ

How much does ad fraud cost advertisers each year?

Juniper Research estimated global losses at $84 billion in 2023, about 22% of digital ad spend, and projects losses reaching $170 billion a year by 2028.

What share of my traffic is fake?

The industry-average invalid traffic rate measured 20.64% in 2026 Fraudlogix data. Your real rate varies by offer, GEO and channel, so measure it in your own tracker.

Which ad format has the most fraud?

Mobile ads measured the highest fraud rate, around 30%, per Juniper Research, driven by SDK spoofing that bypasses standard attribution filters.

Can ad fraud be stopped completely?

No. Even fully protected campaigns retain some exposure, so the goal is cutting your effective fraud rate and verifying conversions server-side.

Does EZmob protect campaigns from bots?

EZmob campaigns can run ClearTrust bot monitoring, which blocks fraudulent traffic before it reaches your campaign stats and your bill.

Is the $84 billion figure still current?

It is the last fully published Juniper estimate, from 2023. Projections put losses in 2026 above that level and rising to $170 billion a year by 2028.

One last thing

The cheapest fraud control costs nothing: a kill rule set before launch. Decide in advance how many clicks you will pay for with zero verified conversions before pausing a placement, and enforce that rule in your tracker — not by feel on day three. Bots count on advertisers who keep paying while the data goes quiet.

See what filtered traffic actually delivers in EZmob's published case studies, or start advertising on EZmob with your own bids and budgets.

Scroll to Top