Choosing A Mobile DSP: 12 Questions To Ask Before You Sign

• 1. Which supply do you access, and how? Ask which exchanges and SSPs the DSP buys from, which formats (in-app video, playable, native, display, CTV) and in which geos. A good answer is specific and admits gaps.
• 2. How do you avoid duplicate supply paths? The same impression can be offered through several resellers. A good DSP prefers direct, authorised paths, checks app-ads.txt and sellers.json, and can explain its supply-path optimization.

• 3. What will I see in reporting? You should be able to see performance by app or placement, creative, OS, geo and device — not only campaign totals. Ask for a sample report.
• 4. Can I control allow and block lists? You should be able to block apps, categories and placements, and run allow-list campaigns on proven supply.

• 5. What does your bidder optimize toward? Installs, events, revenue? A strong answer describes how post-install MMP events feed the model, how it handles sparse deep events, and how long learning typically takes.
• 6. How do you handle iOS? Ask about SKAN 4 support, conversion-value schema design, crowd-anonymity planning and how iOS performance is reported and optimized.

• 7. Which MMPs do you integrate with, and how are discrepancies resolved? The DSP should work with your existing MMP and agree that the MMP is the source of truth for attribution and billing on outcome-based deals.
• 8. What is your fraud approach? Look for pre-bid filtering, post-install monitoring, consumption of MMP rejection postbacks, and a clear policy for how rejected conversions are treated. Be wary of vague claims or unverifiable certifications.

• 9. How are you paid? Platform fee, margin on media, CPI/CPA pricing or a mix. Understand how much of your budget reaches media. Transparency on fees is as important as the level.
• 10. What are the minimums and contract terms? Minimum spend, notice periods, payment terms and whether tests can run without long commitments.

• 11. What creative support do you provide? Spec reviews, format guidance, testing frameworks and creative-level reporting all shorten the path to scale.
• 12. Who will actually run my account? Ask who your day-to-day contacts are, their experience, availability across time zones and how often you will review performance together.

• Guaranteed results before any data has been reviewed.
• Refusal to share placement-level reporting.
• Performance figures without context, sources or definitions.
• No clear policy on fraudulent or rejected conversions.
• Pressure to sign long contracts before a test.

Once a DSP passes due diligence, test it properly. Agree the KPI and target in writing, give the model enough budget and time to learn, keep creative comparable to other channels, and read results on mature cohorts. Ask for weekly reviews during the test so you see how the team works, not just what the numbers say.

Evaluating many platforms at once spreads budget and attention too thin for any test to reach statistical confidence. Most teams get better decisions by shortlisting two or three DSPs on the due-diligence questions, then testing one or two properly in parallel with comparable budgets, creatives and KPIs. A focused evaluation produces clearer results, faster, and leaves budget for scaling the winner.

Before evaluating any platform, be clear about your own requirements. Which KPI defines success, at which cohort day? Which geos and platforms matter most in the next two quarters? How much budget can you commit to a fair test, and for how long? Do you need a self-serve workspace, a managed team, or both? Which data must flow into your own BI? Teams that answer these questions internally first get far more useful answers from vendors, because they can ask about the situations that actually matter to them instead of generic capabilities.

Programmatic buying spreads budget across thousands of apps and placements, through several intermediaries, in milliseconds. That complexity is what makes a DSP powerful — it reaches users other channels cannot — and also what makes it easy to hide inefficiency. Two platforms can show similar headline CPIs while one is buying clean, retained users through direct supply paths and the other is buying recycled, low-quality inventory through long resale chains. The difference only appears when you ask the detailed questions and look at the data.

Due diligence also sets up the working relationship. A DSP that answers precisely, shares sample reports and explains its limitations is likely to be the same partner when results are mixed and decisions are hard. One that answers vaguely before you sign rarely becomes more transparent afterward.

On supply, a good answer names the categories of supply accessed, explains how the platform chooses between paths to the same impression and admits where coverage is thin. On optimization, it explains which events the model uses, how it handles low-volume deep events and what a realistic learning period looks like for your budget. On measurement, it accepts your MMP as the source of truth and describes how discrepancies are investigated. On fraud, it distinguishes pre-bid filtering from post-install review and gives a written policy for rejected conversions.

Notice what is missing as much as what is said. Unqualified performance promises, averages without definitions, and certifications that cannot be verified are all signals to dig deeper.

Involve the people who will live with the decision. UA managers should test the workspace and reporting. Data or BI teams should check that exports and APIs fit your stack. Finance should review commercial terms and invoicing. Product or creative teams should understand the format and spec requirements. A short scorecard — each question rated by the relevant owner — makes comparisons between platforms objective and keeps the decision grounded in evidence rather than the most persuasive presentation.

• Score each of the twelve questions from one to five, with notes on the evidence provided.
• Weight the questions by what matters most to your app — iOS measurement for an iOS-heavy title, fraud controls for fintech.
• Request references to documentation or sample reports rather than relying on verbal answers.
• Re-score after the test with real data, and compare against the pre-test expectations.

Selecting a DSP is the start, not the end. Schedule an onboarding plan with clear owners, agree the review cadence, and document KPIs, targets and attribution settings in one shared place. Revisit the original twelve questions after the first quarter: are the answers you received in the pitch still true in practice? That review keeps both sides accountable and gives you a clear basis for scaling the partnership.

We are happy to walk through every question above in a call — supply access via RTB across in-app, web and CTV; placement-level BI; optimization on your MMP events; SKAN 4 support; layered fraud controls; transparent commercials; and a hands-on team available around the clock.