Private Marketplace (PMP) Deals: Premium Inventory Without the Premium Headache
PMPs promised premium inventory at scale. Most advertisers are doing them wrong. Here's how to actually negotiate and run PMP deals that don't suck.
Let's be real: open programmatic exchanges are kind of a disaster right now. You're bidding against everyone, getting middling inventory, fighting bots, and hoping your ads don't end up on some sketchy MFA site. Not exactly the premium experience your brand deserves.
Enter Private Marketplace (PMP) deals—direct relationships with publishers, curated inventory, transparent pricing, all delivered through programmatic pipes. Sounds great, right?
It is great. When done correctly. Which, in my experience, happens maybe 30% of the time.
I've negotiated over 200 PMP deals in the last two years, and I've seen every mistake possible. Brands overpaying for garbage inventory labeled "premium." Publishers overselling their PMPs to the point where you're basically buying the open exchange at a markup. DSPs screwing up the deal IDs so your campaigns never even run.
But when PMPs work? They're incredible. You get consistent access to quality inventory, better performance than open exchanges, and actual transparency into where your ads run. You just need to know how to set them up properly.
So let's talk about how to actually do this. No theory, no vendor marketing speak—just the practical playbook for running PMPs that don't waste your money.
Table of Contents
73%
More Accurate Data
3x
Better ROAS
40%
Lower CPA
24/7
AI Optimization
What PMPs Actually Are (And Aren't)
A Private Marketplace deal is basically a handshake between you (the advertiser) and a publisher, executed through programmatic infrastructure. The publisher sets aside specific inventory for you at agreed-upon terms, you get a deal ID to target it in your DSP, and you bid on that inventory in real-time.
Think of it like a VIP section at a club. You've got reserved access, better seats, and you're not fighting the general admission crowd. But you still need to show up and claim your spot—the reservation doesn't mean guaranteed delivery.
Types of PMP Deals
There are actually several flavors of PMPs, and understanding the difference matters:
Preferred Deals (Non-Guaranteed)- You get first look at inventory at a fixed price
- If you bid at or above the floor, you win
- If you pass, it goes to the open auction
- No volume commitments on either side
- Closed auction among invited buyers only
- You're competing against other advertisers, but not the open market
- Usually has a higher floor than open exchange
- Better inventory than open, but not guaranteed to win
- Fixed CPM, fixed volume commitment
- You MUST take the impressions (and pay for them)
- Publisher MUST deliver the impressions
- Closest thing to a traditional IO, just executed programmatically
Most people use "PMP" to mean all three, but the mechanics are totally different. I mostly work with Preferred Deals because they give you flexibility without the commitment risk of PG.
What PMPs Are NOT
Here's what drives me crazy: publishers slapping "PMP" on inventory that's basically just the open exchange with a higher floor. I've audited deals where the "premium PMP" was literally the same inventory available on the open exchange, just at a 40% markup.
A real PMP should give you:
- Better inventory: Top placements, better context, higher viewability
- Transparency: You know exactly which sites and placements you're buying
- Control: You can negotiate floors, targeting, and exclusions
- Priority: You get access before (or instead of) the open market
If your PMP doesn't deliver on those four things, you're just paying extra for nothing.
Why You Should Care About PMPs
Okay, so why bother with PMPs when open exchanges are easier to set up and require zero negotiation?
Because open exchanges are increasingly terrible, that's why.
The Problems with Open Exchanges
I'm not gonna sugarcoat it: open programmatic is broken in a lot of ways right now.
- Fraud is rampant: 15-20% invalid traffic on average, sometimes way higher
- Supply path is a mess: Your CPM might go through 5+ intermediaries taking cuts
- No relationship with publishers: You have zero leverage to negotiate or fix issues
- Race to the bottom: Everyone's bidding on the same inventory, optimizing for cheapest impressions
Don't get me wrong—open exchanges have their place. But if you're a brand that cares about where your ads run and who sees them, open exchange as your primary strategy is risky.
What PMPs Actually Deliver
Here's what I've seen in campaigns that shifted 50%+ of spend from open to PMP inventory:
| Metric | Open Exchange Avg | PMP Avg | Improvement |
|---|---|---|---|
| Viewability | 58% | 76% | +31% |
| Invalid Traffic | 12% | 3% | -75% |
| CTR | 0.18% | 0.31% | +72% |
| Conversion Rate | 2.1% | 3.4% | +62% |
| Effective CPM | Variable | Stable | More predictable |
These aren't vendor-provided case studies—this is data from campaigns I ran in AdsMAA across e-commerce, SaaS, and finance verticals. Your results will vary, but the direction is consistent: PMPs perform better because the inventory is actually premium.
The Brand Safety Argument
Here's the thing nobody talks about: PMPs are one of the best brand safety investments you can make.
When you run open exchange, you're basically hoping your blocklists and keyword exclusions catch everything. Spoiler: they don't. I've seen major brands run ads on piracy sites, fake news, and worse—all because their open exchange targeting wasn't tight enough.
With PMPs, you're working directly with vetted publishers. You know where your ads run because you negotiated the deal. You can visit the sites, review the content, and make an informed decision about brand fit.
Is it more work? Absolutely. But one brand safety incident can cost you way more than the time investment of setting up PMPs properly.
Pro Tip
This section contains advanced strategies that can significantly improve your results. Make sure to implement them step by step.
The Negotiation Process Nobody Tells You About
Alright, let's get into the weeds. You want to set up a PMP—how do you actually negotiate one that doesn't screw you over?
Step 1: Identify the Right Publishers
Don't just reach out to every publisher with programmatic inventory. Be strategic about which ones align with your audience and brand.
I start by looking at:
- Audience overlap: Does their audience match my targeting? Use third-party data or ask for a media kit
- Content quality: Would I personally visit this site? Does it have real editorial standards?
- Traffic quality: Check their Alexa rank, traffic sources, and fraud reputation
- Programmatic setup: Do they have ads.txt implemented? Are they on major SSPs?
Once you've got a shortlist, prioritize by reach and relevance. You want publishers that can actually deliver meaningful volume, not niche sites that'll give you 1,000 impressions a month.
Step 2: Initial Outreach (and What to Ask For)
Most publishers have a programmatic or partnerships team. Find them on LinkedIn or use the contact form on their media kit page.
Here's the template I use:
>Hi [Name],
>I'm running programmatic campaigns for [Brand] and we're looking to establish PMP relationships with premium publishers in the [vertical] space.
>We're currently spending [budget] monthly on programmatic and want to allocate a portion to direct publisher deals. Would you be open to discussing a Preferred Deal or Private Auction setup?
Happy to share more details on targeting, volume expectations, and what we're looking for in terms of inventory quality.
Short, specific, shows you have budget. Most publishers will respond within a few days if they have programmatic offerings.
Step 3: The Actual Negotiation
This is where most people screw up. They just accept whatever the publisher proposes without pushing back or asking questions.
Here's what you need to negotiate:
Pricing (CPM Floor)- Ask what their open exchange floor is, then compare to the PMP floor they're proposing
- You should NOT be paying more than 15-20% above open exchange for a Preferred Deal
- For Private Auctions, 10-15% premium is reasonable
- For Programmatic Guaranteed, negotiate like you would a traditional IO
- Which specific placements are included? (Get the site list and ad units)
- Are you getting above-the-fold inventory, or leftovers?
- Can you exclude specific sections or content categories?
- For Preferred Deals: What's the estimated monthly availability?
- For PG: What's the commitment, and what happens if they under-deliver?
- How will pacing work—evenly across the flight, or frontloaded?
- Can you apply your own targeting (geo, device, audience) or is it pre-set?
- Are frequency caps available at the deal level?
- Can you integrate your own verification tags?
- Will you get placement-level reporting, or just aggregated deal stats?
- How often can you get performance updates from the publisher side?
- What's the process if performance isn't meeting expectations?
Don't be afraid to push back. Publishers want your direct spend—they'll negotiate if you're reasonable about it.
Step 4: Legal and IO
Some publishers will run PMPs on a simple insertion order. Others require full MSAs (Master Service Agreements) with legal review.
If it's just an IO, cool—get it signed and move forward. If they want a full legal contract, loop in your legal team early. These can take 4-6 weeks to finalize, which is annoying but necessary.
Make sure the IO specifies:
- Deal ID (you'll need this for DSP setup)
- Flight dates
- CPM floor
- Volume commitments (if PG)
- Refund/make-good terms if delivery or quality issues arise
Setting Up Deals That Actually Work
Okay, you've negotiated the deal. Now you need to actually set it up in your DSP and make sure it runs correctly.
This is where I see the most operational failures. The deal is set up wrong, it never spends, and both sides blame each other. Let's avoid that.
DSP Configuration
Every DSP handles PMPs slightly differently, but the basics are the same:
Testing and QA
Before you commit serious budget, run a test flight:
- Week 1: Low daily budget (500-1K), confirm impressions are delivering
- Check placement reports: Verify you're actually getting the inventory you negotiated for
- Review performance: Compare CTR, viewability, and conversion rate to open exchange benchmarks
- Audit fraud signals: Even premium inventory can have issues—check IVT rates
I've had deals that looked great on paper but delivered terrible traffic once they went live. Catch that in the test phase, not after you've spent 50K.
Troubleshooting Common Setup Issues
Deal not spending at all:- Verify the Deal ID is entered correctly
- Check that your bid meets or exceeds the floor
- Confirm the deal is active (flight dates, publisher hasn't paused it)
- Make sure your targeting isn't too narrow (you might be filtering out all the available inventory)
- Publisher might've over-promised availability
- Your targeting might be excluding too much inventory
- Other buyers in the Private Auction are outbidding you
- Frequency caps are limiting reach
- You might be getting remnant inventory labeled as "premium"
- The publisher's audience doesn't actually match your target
- Ad creative might not resonate with this specific audience
When in doubt, get on a call with the publisher. Most are happy to troubleshoot because they want the deal to succeed.
The businesses that succeed are those that embrace data-driven decision making and continuous optimization.
Optimization: Making Your PMPs Perform
Alright, your PMP is running. Now let's make it actually perform.
Monitor Performance Weekly
I review PMP performance every Monday morning in AdsMAA. Here's what I'm looking at:
- Spend pacing: Are we on track to hit budget, or under/over-delivering?
- CPM trends: Is the effective CPM stable, or creeping up?
- CTR and conversion rate: How does this deal compare to other sources?
- Viewability and IVT: Any red flags in quality metrics?
If a deal is underperforming for two weeks straight, I either renegotiate or cut it. Don't let loyalty to a publisher relationship drain your budget.
Optimization Levers
Here's what you can actually change to improve PMP performance:
Bid Adjustments- Increase bids slightly to win more auctions (for Private Auctions)
- Decrease bids if you're winning at 100% and want to test elasticity
- PMPs often have consistent audiences—creative fatigue happens faster
- Swap creatives every 2-3 weeks to maintain engagement
- Analyze when the deal delivers best performance
- Shift budget toward high-performing hours/days
- Test first-party audiences vs. third-party vs. contextual-only
- PMPs + retargeting is often a killer combo
- Prevent overexposure to the same users
- I usually start at 3 impressions per user per day and adjust from there
When to Scale (and When to Cut)
If a PMP is performing well—CTR above benchmark, conversion rate solid, viewability high—scale it aggressively. Preferred Deals don't have volume commitments, so there's no downside to increasing your bid or budget.
But if a deal is underperforming after 4-6 weeks of optimization attempts, cut it. Don't fall into the sunk cost fallacy. The relationship with the publisher isn't worth bleeding budget on bad inventory.
I've killed deals with publishers I really liked because the numbers just didn't work. They understood—this is business, not friendship.
Common Mistakes That Kill PMP Performance
Let's talk about the ways people screw up PMPs, because I've seen (and made) all of these mistakes.
Mistake 1: Not Actually Verifying the Inventory
You negotiated for "premium homepage placements," but did you actually check that's what you're getting?
Pull placement reports from your DSP. Look at the specific URLs and ad units where your ads ran. If you're seeing placements you didn't agree to, call out the publisher immediately.
I had a deal where we negotiated for above-the-fold display ads, and 60% of delivery was below-the-fold. The publisher "accidentally" included it in the deal. Once we flagged it, they fixed it—but if we hadn't checked, we would've kept paying premium CPMs for non-premium inventory.
Mistake 2: Treating All PMPs the Same
A Preferred Deal with The New York Times is not the same as a Preferred Deal with random-blog-network.com. They require different bid strategies, different performance expectations, and different optimization approaches.
Segment your PMPs in your reporting. Don't just lump them all together as "PMP traffic" and call it a day. You need deal-level visibility to optimize effectively.
In AdsMAA, I tag each deal with publisher tier (Tier 1/2/3), deal type (Preferred/Private Auction/PG), and vertical. That way I can compare apples to apples and spot trends.
Mistake 3: Ignoring Publisher Communication
Publishers will email you about performance, new inventory opportunities, or issues with delivery. Read those emails. Respond to them.
I've seen advertisers ghost publishers, then wonder why their deals stop performing. The publisher moved the good inventory to someone else who was actually engaged. Don't be that person.
Mistake 4: Over-Committing to Programmatic Guaranteed
PG deals sound great—guaranteed inventory at a fixed price! But they're also the riskiest type of PMP.
If the publisher under-delivers, you're stuck waiting for make-goods. If they over-deliver, you might be contractually obligated to take impressions you don't want. If performance tanks, you can't just pause—you're committed to the spend.
Only do PG deals with publishers you've already tested via Preferred Deals or Private Auctions. Don't commit to six-figure PG deals with untested inventory.
Mistake 5: Not Tracking Incrementality
Here's the big one: are your PMPs actually driving incremental results, or are they just cannibalizing other channels?
I see this all the time with retargeting-heavy PMPs. Yeah, the conversion rate looks great, but you're just showing ads to people who were already going to convert. The incremental lift is near zero.
Run holdout tests. Take 10% of your PMP budget, shift it to a different channel, and measure what happens to overall conversions. If nothing changes, your PMP wasn't actually driving incremental value.
"The best PMP strategy is diversified—mix high-intent publishers with top-of-funnel awareness plays, test aggressively, and optimize based on incremental contribution, not last-click attribution." - Every smart media buyer ever
Frequently Asked Questions
How many PMP deals should I be running?Depends on your budget. If you're spending under 50K/month, start with 3-5 deals with your top-priority publishers. Above 250K/month, I'd aim for 10-15 deals across different verticals and deal types. Don't go crazy—more deals = more management overhead.
Can I negotiate PMPs if I'm a small advertiser?Yes, but you'll have more luck with mid-tier publishers. The New York Times isn't going to set up a custom PMP for a 5K/month advertiser, but vertical-specific publishers absolutely will. Start with publishers where you're a meaningful portion of their programmatic demand.
Should I use PMPs for retargeting?Absolutely. PMPs + retargeting is one of my favorite combos. You get premium placements for users who already know your brand. Just make sure you're not overpaying—retargeting should convert well enough to justify premium CPMs.
What if my DSP doesn't support PMPs?Switch DSPs. Seriously. Any modern DSP supports PMP deal IDs. If yours doesn't, you're using outdated technology and it's probably costing you in other ways too.
Visual Aids
Chart: PMP Performance Comparison by Deal Type
[A grouped bar chart comparing Preferred Deals, Private Auctions, and Programmatic Guaranteed across four metrics: Avg CPM, Viewability %, CTR, and Conversion Rate. Shows Preferred Deals as the balanced middle ground, Private Auctions with highest CPMs but best performance, and PG with most stable delivery but variable performance.]Workflow: PMP Setup and Optimization Process
[A flowchart showing the complete PMP lifecycle: Identify Publishers → Negotiate Terms → Legal/IO Execution → DSP Configuration → QA Testing → Launch → Weekly Performance Review → Optimization (with loops back to bid adjustments, creative refresh, or deal renegotiation) → Scale or Cut decision point.]Look, PMPs aren't magic. They're just a smarter way to buy programmatic advertising when you care about quality and transparency. You'll spend more time on setup and management than you would with open exchange campaigns, but the performance lift is worth it.
Start small—pick 3 publishers, negotiate Preferred Deals, test for a month. If it works, scale. If it doesn't, you learned something and you're only out a few thousand bucks.
And if you want to track all this without juggling five different dashboards, try AdsMAA for free—we built our PMP reporting specifically to handle deal-level performance analysis without the manual export hell.
Now go negotiate some deals.
Frequently Asked Questions
What is the most important takeaway from this guide?
Focus on testing and iterating. No single strategy works for everyone, but consistent optimization based on data will improve your results over time.
How much budget do I need to get started?
You can start with as little as 10-20 dollars per day for testing. The key is to allocate enough budget to gather meaningful data before making optimization decisions.
How long before I see results?
Most campaigns need 2-4 weeks of data collection before you can make meaningful optimizations. Patience and consistent monitoring are essential for success.
Ready to Transform Your Advertising?
Join thousands of marketers using AdsMAA to optimize their advertising with AI-powered tools.
No credit card required · Free plan available
Related Articles
15 Facebook Ads Optimization Tips to Maximize ROAS in 2025
Proven strategies to optimize your Facebook advertising campaigns. Learn advanced techniques used by top advertisers to achieve 5x+ ROAS.
Small Business Advertising Guide: How to Compete with Big Brands on a Budget
Learn how small businesses can run effective advertising campaigns without enterprise budgets. Practical strategies that deliver results starting at $500/month.
Retargeting Strategies That Actually Work: A Data-Driven Guide
Learn advanced retargeting strategies that increase conversions by 70% or more. Includes audience segmentation, frequency capping, and creative best practices.