For over a decade, our team has been buying programmatic advertising. Over those years, we’ve fine-tuned the process to get our clients the most significant ROI possible with the least amount of wasted advertising dollars while cracking down on ad fraud. We use AI, proprietary tools, and unique subscriptions to provide our clients with the best media-buying service possible. Digital Advertising is riddled with fraud, and being aware of how bad actors skim off marketing funds is one of the most critical aspects of professionally managing advertising investments. We use AI in a variety of ways to improve advertising performance.
For optimizations, AI can assist with hitting certain KPIs, such as low CPMs, CPC, VCR, and CPAs.
An example of using AI in advertising campaigns for bidding would be “if a user is in Los Angeles and visiting a health or wellness website, increase the bid price by X.”
For campaign delivery, it could be used to spend a certain percentage of a campaign budget on specific times of the day, days of the week, etc.
Most DSPs (Demand Side Platforms) have some form of AI or off-the-shelf algorithms that a media buyer can use. However, the built-in SPO (supply path optimization) and feedback loop with our subscription to the best in industry anti-fraud analytics tool are unique to our trusted DSP.
SPO happens organically across all open-market inventory thanks to our DSP’s Global Curation Initiative, where every impression transacted is analyzed to identify the most effective supply path for every publisher, using price, viewability, and “humanness” as key metrics. This strategy allows only directly sold supply without resellers on the DSP.
When running in other DSPs, you’re often exposed to 50+ exchanges by default, and many of these offer zero value to you and are essentially intermediaries along the supply chain that are there to collect a piece of the bid.
For display campaigns, cracking down on ad fraud is paramount. Our DSP only uses 3-4 exchanges, which are a direct path to a publisher’s inventory (domain), and CTV is often only 1-2 exchanges. Typically, advertisers, if they have historical campaign data from their previous DSP campaigns, probably saw conversions happening on websites they’ve never heard of and ad placements that weren’t viewable.
The technical side of cracking down on Ad Fraud
Below, I’ve included some of the more technical stuff (that makes a big difference). Our team has been buying media in DSPs for over a decade, and even when using allowlists, there are often no safeguards to prevent any of the following from happening.
Typically, the ad buyer checks boxes in hopes that they’re avoiding fraud and bots. Still, there’s often no or vague reporting to help determine the effectiveness or lack thereof.
For every impression, the DSP we use provides the most advanced deterministic General Invalid Traffic (GIVT) and Sophisticated Invalid Traffic (SIVT) pre-bid filters on the market, including:
- Anti-Device Spoofing: The DSP we use employs advanced network fingerprinting to validate device type and manufacturer at delivery time (e.g., a declared Apple TV device is effectively a real Apple TV device instead of a Desktop Chrome Web Browser), reporting back results on a per-impression basis. The network analysis occurs at auction time for inventory transacted via our SSP.
- Anti-Publisher Spoofing: ads.txt and sellers.json are strictly enforced and verified on every request.
- 3. Anti-Placement Spoofing: Our trusted DSP ensures video placement quality by detecting and measuring discrepancies between declared placement (e.g., in-stream OTT) vs actual placement (e.g., muted outstream web), proactively blocking the combo SSP / Publisher ID / Site or App ID / Placement ID across the entire DSP.
- 4. Anti-VTR Spoofing: For creatives delivered using our DSP’s ad server, monitoring of assets download, and VAST pixel-firing timing verify no cheating is happening on the VTR measurement.
- 5. Datacenter exclusion: The DSP we use maintains a daily updated list of IP ranges and Autonomous System Numbers (ASN) owned or used by data center and cloud providers. All requests coming from these IP ranges are blocked on the DSP edge.
- 6. Crawler exclusion: The DSP we use maintains a daily updated list of User Agents, IP ranges, and ASNs owned or used by crawlers. All requests coming from these ranges are blocked on the DSP edge.
- 7. VPN exclusion: The DSP we use maintains a daily updated list of IP ranges and ASNs owned or used by VPN providers. All requests coming from these IP ranges are blocked on the DSP edge.
- 8. Proxies exclusion: The DSP we use maintains a daily updated list of IP ranges and ASNs owned or used by proxy providers. All requests coming from these IP ranges are blocked on the DSP edge.
- 9. Residential Proxies exclusion: Our DSP maintains a daily updated list of IP addresses used as exit nodes by residential proxy providers. All requests coming from these IPs are blocked on the DSP edge.
- 10. In-App Web Views: Our trusted DSP proactively blocks all requests generated from In-App Web Views (e.g., the Instagram browser).
- 11. Rewarded Content: The DSP we use proactively blocks all requests generated from rewarded or incentivized content.
- 12. Arbitrage: Our trusted DSP proactively blocks all requests generated by CPC-to-CPM arbitrage (e.g., Taboola or Outbrain) by analyzing the referrer of every request and looking for arbitrage markers.
- 13. Invalid Auction Data: The DSP we use verifies the consistency of all request fields (e.g., domain and page match).
On average, 70%+ of impressions do not pass our filter checks and aren’t bid on. Since we have a proprietary subscription to the best in industry anti-fraud analytics tool, we recommend providing prospective clients of ours use our tags for a while, running display campaigns elsewhere to show the difference our filters make. This tactic is a service we’d happily provide a client doing a lot of digital advertising and wants to know how our offer is superior. Ask us about this today!
Our trusted DSP is working for every impression, providing the most reliable deterministic brand-safety pre-bid filters on the market, cracking down on ad fraud by including:
1. High-Risk URLs: Checking page URLs against a list of universally agreeable “very bad keywords.”
2. Medium-Risk URLs: Checking page URLs against a list of subjectively unsafe keywords (e.g., “murder”).
3. Global Allow List for Mobile App Bundles: Employing a DSP-wide allow list of approved mobile app bundles.
4. Global Allow list for CTV Bundles: Employing a DSP-wide allow list of approved CTV app bundles.
If you’d like to talk to our team about cracking down on ad fraud and improving your digital advertising results, contact us for a complimentary conversation about your current marketing plan.