Anti-Fraud Policy

TrafficSigma operates 3-level fraud protection - pre-bid filtering, real-time detection, and post-campaign analysis - across every format on the network. This Policy explains what each level does, what counts as fraud, how we enforce against it, and how advertisers and publishers can raise a traffic-quality dispute. It forms part of the Advertising Agreement and the Publisher Terms.

No system eliminates fraud entirely. The controls described here are designed to reduce invalid traffic materially, and we invest continuously in them, but they are not a guarantee that no invalid traffic will ever be delivered, and nothing in this Policy creates a warranty of a specific fraud rate, conversion rate, or return on investment.

1. What we treat as fraud and invalid traffic

"Invalid traffic" includes, without limitation:

• Bot and automated traffic: crawlers, headless browsers, emulators, scripted agents, and automation frameworks masquerading as human users.
• Data-centre and proxy traffic presented as residential or mobile traffic, including anonymizing VPN and proxy chains used to disguise origin.
• GEO and device spoofing: falsified location, user agent, carrier, OS, language, or device fingerprint.
• Click injection, click spamming, and click flooding in mobile and app environments.
• Impression fraud: stacked, hidden, zero-pixel, or out-of-viewport ad placements and auto-refresh abuse.
• Domain and app spoofing: misrepresenting the inventory on which an ad is served.
• Incentivized or coerced engagement where the campaign was not sold as incentivized.
• Conversion fraud: fake leads, form-filling scripts, recycled or purchased contact data, and self-referred or manufactured conversions.
• Attribution manipulation: cookie stuffing, forced tracker firing, and last-click hijacking.
• Malvertising: creatives that deliver malware, forced redirects, or drive-by downloads.

2. Level 1 - Pre-bid filtering

Before an impression is bought or served, traffic is filtered against source-level and request-level signals. This level includes publisher and traffic-source vetting at onboarding, source reputation scoring, IP intelligence covering known data-centre ranges and anonymizing infrastructure, device and user-agent integrity checks, blocklists of known-bad placements and identifiers, and consistency checks between declared and observed GEO, carrier, OS, and language. Requests that fail these checks are not served.

3. Level 2 - Real-time detection

While a campaign delivers, traffic is scored continuously. Behavioural and statistical signals - engagement timing, interaction patterns, click-to-impression and conversion ratios, frequency and velocity anomalies, duplicate and burst detection, and per-source deviation from expected baselines - are evaluated in real time. Traffic that crosses risk thresholds is blocked, throttled, or routed away from the campaign, and affected placements can be automatically paused pending review.

4. Level 3 - Post-campaign analysis

After delivery, we review campaign and source data retrospectively for patterns that only become visible at scale: cohort quality against advertiser-reported outcomes, conversion-quality signals, chargeback and refund correlation, source-level trend analysis, and investigation of advertiser and publisher reports. Findings feed back into the pre-bid blocklists and detection models, and support credits, payout reversals, and enforcement decisions.

5. Advertiser obligations

• Implement tracking accurately and do not manipulate postbacks, conversion values, or attribution data.
• Report suspected invalid traffic promptly, with the evidence described in Section 8.
• Do not use fraud claims as a mechanism to avoid paying for legitimately delivered advertising.
• Do not run creatives that carry malware, forced redirects, or the deceptive techniques prohibited by our Acceptable Use Policy.

6. Publisher obligations

• Send only genuine human traffic from inventory you own or are authorized to monetize.
• Never buy, generate, incentivize, or otherwise artificially produce impressions or clicks.
• Do not misrepresent the domain, app, format, placement, or geography of your inventory.
• Do not stack, hide, auto-refresh abusively, or place ads so they are clicked unintentionally.
• Disclose any material change to your traffic sources or acquisition methods.

7. Enforcement

Where invalid traffic is identified, we may, proportionately to severity and history:

• Filter, throttle, or block the affected source, placement, or geography.
• Credit an advertiser for verified invalid traffic, or reverse or withhold the corresponding publisher revenue.
• Suspend the campaign, placement, or account pending investigation.
• Terminate the account and forfeit revenue attributable to fraudulent activity, as permitted by the applicable agreement.
• Refuse future business with the individuals or entities involved, and, in serious cases, refer the matter to law enforcement.

8. Raising a traffic-quality dispute

Advertisers and publishers may raise a dispute by emailing [email protected]. To let us investigate effectively, include:

• Your account ID and the campaign, zone, or placement ID concerned.
• The date and time range, in UTC, and the geographies affected.
• Log-level or aggregated evidence: your own tracker or MMP report, click and conversion IDs, IP or device samples, and any third-party verification report.
• A description of the pattern you observed and the remedy you are seeking.

We aim to acknowledge disputes within two business days and to complete an investigation within 14 days, or to explain why more time is needed. Disputes should be raised within 30 days of the delivery in question, so that log data is still available. Where a claim is verified, the remedy is normally a credit to the advertiser account rather than a cash refund; cash refunds remain governed by our Refund Policy.

9. Data used for fraud prevention

Fraud detection involves processing technical data such as IP addresses, device and browser characteristics, timestamps, and interaction patterns. We process this data on the basis of our legitimate interest in keeping the network free of fraud and in protecting advertisers, publishers, and end users, and we retain it only as long as necessary for that purpose. See our Privacy Policy for details and for your rights.

10. Changes to this Policy

Fraud techniques evolve, and we update our controls and this Policy accordingly. Material changes are reflected in the "Last updated" date above.