Prompt

How do I evaluate whether a premium video inventory seller is credible and unbiased?

Advertising · DSP & SSP / Dsp ssp1 observationLast seen Jul 27, 2026

Latest observation

Jul 27, 2026 · OpenAI APIWeb search: off

To evaluate whether a premium video inventory seller is credible and unbiased, look at both proof of quality and proof of independence.

1) Check their supply quality

Ask for evidence on:

  • Publisher relationships: Are they directly authorized by the publishers, or are they reselling through layers?
  • Placement transparency: Can they show where ads run, on what content, and in what formats?
  • Inventory type: Is it premium in-stream, out-stream, CTV, or a mix? “Premium” should be defined clearly.
  • Viewability and completion rates: Request historical performance by placement.
  • Brand safety controls: Ask what categories, keywords, and exclusions they use.
  • Fraud prevention: Look for measures against bot traffic, domain spoofing, and IVT.

2) Assess whether they are unbiased

A seller can be credible but still biased if they only push inventory that benefits them most. Check:

  • Disclosure of conflicts of interest: Do they own media, represent only certain publishers, or receive preferential economics?
  • Selection criteria: How do they decide which publishers or placements are “premium”?
  • Comparable alternatives: Do they present multiple options, or only one preferred package?
  • Measurement transparency: Do they use independent measurement tools or only their own reporting?
  • Optimization incentives: Are they paid on spend, margin, or performance? Incentives can shape recommendations.

3) Verify with independent sources

Use outside validation:

  • Third-party verification: IAS, DoubleVerify, MOAT, Pixalate, etc., if applicable.
  • Ad specs and policies: Compare claims against publisher or platform documentation.
  • Reference checks: Ask for advertiser references or case studies with named verticals.
  • Market reputation: Look for consistency across multiple buyers and agencies.

4) Ask the right questions

A credible, unbiased seller should answer clearly:

  • What is your direct vs. indirect supply?
  • Which publishers own this inventory?
  • Can I get log-level reporting or placement-level reporting?
  • What is your policy on hidden fees or markups?
  • How do you define “premium”?
  • What independent verification do you support?
  • What inventory would you not recommend for my campaign, and why?

5) Watch for red flags

Be cautious if they:

  • Use vague terms like “premium” without definitions
  • Refuse to disclose publisher names or placement details
  • Overstate exclusivity or guarantee results without data
  • Only provide their own dashboards and no third-party verification
  • Push a one-size-fits-all package
  • Avoid discussing make-goods, refunds, or underdelivery policies

6) Run a test

Before committing:

  • Start with a small pilot
  • Set clear KPIs: viewability, completion rate, CTR, brand safety, attention, conversions
  • Compare against another vendor or benchmark
  • Review results at the placement level, not just aggregate

Simple credibility scorecard

Rate each 1–5:

  • Direct publisher authorization
  • Transparency of placements
  • Independent verification support
  • Brand safety/fraud controls
  • Disclosure of conflicts
  • Consistency of performance data
  • Willingness to answer tough questions

A seller scoring high across all categories is more likely to be both credible and unbiased.

If you want, I can turn this into a vendor due diligence checklist or a questionnaire you can send to sellers.

Observation history

Unlock observation history

Create a free Obsurfable account to unlock historical responses, model comparisons, and deeper brand intelligence.

How did Obsurfable measure this prompt?

Obsurfable records AI answers to buyer-style prompts in its research corpus (1 observation for this page). Metrics are distributions over observations, not a single static ranking.

Which AI systems does Obsurfable collect answers from?

OpenAI, ChatGPT, Google, Gemini, Google AI Mode, Anthropic, Claude, Perplexity, Grok, DeepSeek, Mistral, Copilot, and Meta AI.