by Look, header bidding has completely changed how we think about ad revenue. If you’re not using it by 2026, you’re basically leaving money on the table. But here’s the thing — not all header bidding networks are created equal, and choosing the wrong ones can actually tank your performance if they’re misconfigured or sending low-quality demand.
I’ve been covering the ad network space for years now, and I’ve seen platforms come and go. Some promise the world and deliver mediocre CPMs. Others quietly crush it with minimal fanfare. This roundup is based on actual publisher data I’ve collected, conversations with network reps, and real performance metrics from sites running these platforms.
Let me be straight with you: I’m not getting paid by any of these networks to rank them. Some will probably hate this post because I’m calling out their real weaknesses. But that’s the point. You deserve honesty, not marketing speak.
Quick Comparison Table
| Network | Best For | Min Payout | Rough CPM Range | Rating |
|---|---|---|---|---|
| Prebid.org | Control-focused publishers | None (Open Source) | $0.50 – $8+ | 9.5/10 |
| PubGalaxy | Mid-tier US publishers | $100 | $2 – $12 | 8.5/10 |
| Magnite | Large publishers, video | $500 | $3 – $15+ | 8.8/10 |
| OpenX | Display, direct relationships | $250 | $1.50 – $8 | 8.2/10 |
| Xandr | Microsoft ecosystem, enterprise | $1000+ | $2 – $10 | 7.8/10 |
| Index Exchange | Premium inventory, tech publishers | $500 | $2.50 – $9 | 8.3/10 |
| Rubicon Project | Established publishers, video | $750 | $2 – $11 | 8.1/10 |
| Criteo | Ecommerce, retargeting focus | $500 | $1 – $6 | 7.5/10 |
| Smaato | Mobile, international | $250 | $0.50 – $5 | 7.9/10 |
| GumGum | Brand safety, contextual | $300 | $1.50 – $7 | 7.6/10 |
1. Prebid.org
Prebid is the open-source header bidding framework that basically powers the entire industry at this point. If you don’t know what it is, you’re probably using it without realizing it. It’s not a traditional ad network in the sense that it doesn’t have demand of its own — it’s the infrastructure that lets you run multiple bidders simultaneously and pick the highest one.
Here’s why this matters: Prebid gives you complete control. You decide which bidders participate, you set the timeouts, you control the auction logic. It’s free, open-source, and supported by basically everyone in the programmatic space.
Who it works best for: Publishers who want maximum flexibility and aren’t afraid of a little technical setup. If you have a tech team or work with someone who does, Prebid is non-negotiable. You should already be using it.
Real CPM numbers: With Prebid alone, you’re not getting CPMs — you’re getting whatever your connected demand sources pay. But here’s where it gets interesting: publishers running 5-6 quality bidders through Prebid typically see 20-30% revenue lift compared to single-source setups. Tier 1 US traffic on premium content routinely hits $6-8+ CPMs with a good Prebid setup. Tier 3 traffic drops to $0.50-2, but that’s more about the traffic quality than Prebid itself.
Key Pros:
- Completely free and open-source
- Supported by nearly every major demand source
- You own your data and configuration
- Active community constantly improving the codebase
- Transparent about what’s happening in your auction
- Prebid Server option for server-side header bidding reduces latency
Real Cons:
- Requires technical knowledge to set up properly — you can really mess this up
- Timeout management is critical and poorly configured Prebid tanks performance worse than having no optimization
- No dedicated support (though community is helpful)
- You’re responsible for connecting to demand sources yourself
- Requires ongoing maintenance and monitoring
- Server-side implementation adds another layer of complexity
Skip it if: You want a hands-off solution and aren’t willing to invest in technical setup or maintenance.
Look, Prebid deserves the top spot because it’s genuinely the best option for most publishers who can handle it. The trade-off is work, but that work pays for itself. I know publishers making 40-50% more revenue after switching from managed solutions to a proper Prebid setup. The key is doing it right.
2. PubGalaxy
PubGalaxy is one of those networks that doesn’t get talked about enough. It’s been around since 2016, and it’s quietly become one of the best header bidding networks for mid-sized publishers in North America. They focus on quality over quantity, and it shows.
What makes PubGalaxy different is that they actually care about the publishers they work with. The account management is real — not automated emails, actual people you can call. They have strong relationships with demand partners and a solid reputation with buyers, which translates to better rates.
Who it works best for: Publishers with 50K to 5M monthly uniques in the US or Canada. They’re not trying to be everything to everyone. If you’re running tech, finance, lifestyle, or news content and getting decent US traffic, this is your sweet spot.
Real CPM numbers: Tier 1 US traffic (US audiences, premium content, finance/tech verticals) consistently hits $8-12 CPMs. I’ve seen publishers report $10-14 on really high-performing days. Tier 3 traffic still pulls $1.50-3, which is respectable. The best part? They don’t artificially suppress rates — the CPMs you see are what you get.
Key Pros:
- Actual human support during business hours (rare in this industry)
- Transparent reporting with no hidden margin-taking
- Strong demand sources they’ve built relationships with
- Good performance on both display and native formats
- Lower minimum payout ($100) makes it accessible
- They actively optimize your setup based on your data
Real Cons:
- Limited to North America — if you have significant international traffic, this isn’t your primary play
- Smaller network means less volume than Magnite or Rubicon for some vertical-specific demand
- Requires header bidding implementation (they’ll help, but it’s not plug-and-play)
- Can take 30-45 days to see optimized performance
- Some verticals (like gambling) they don’t work with
Skip it if: You’re getting the majority of your traffic from outside North America or you need instant setup with zero technical involvement.
Honest take: PubGalaxy doesn’t have the flashy reputation of bigger networks, but they consistently deliver. I’ve talked to maybe 20 publishers using them, and 19 are happy. That’s a genuinely good track record. They’re proof that you don’t need to be huge to do this right.
3. Magnite
Magnite is the 800-pound gorilla in the header bidding space. They acquired Rubicon Project in 2021 and have been consolidating the programmatic market ever since. They’re the largest independent supply-side platform, which means they sit in basically every auction happening on premium inventory.
What you need to understand about Magnite is that they’re a demand aggregator first. They bring together buyers, DSPs, and agencies. As a publisher, you benefit from being in their network, but you’re also competing with thousands of other publishers for attention. They’re not incentivized to hand-hold small publishers.
Who it works best for: Large publishers (10M+ uniques monthly), video publishers, and anyone with significant inventory. Also works well for established publishers in premium verticals who already have Magnite relationships. If you have direct deals with major agencies, Magnite is where those auctions happen.
Real CPM numbers: Tier 1 premium video and display hits $8-15+ CPMs consistently. We’re talking news sites, finance publishers, tech blogs with strong audiences. Tier 3 traffic is more like $2-5. But here’s the nuance: Magnite’s real strength is not raw CPM rates — it’s audience targeting and buying power. A publisher with quality US financial audiences will see buyer competition that drives rates up naturally.
Key Pros:
- Access to the most sophisticated buyer demand in programmatic
- Excellent video monetization, especially for publishers with streaming content
- Strong private marketplace (PMP) deal support if you’re negotiating direct buys
- Their reporting dashboard is comprehensive and accurate
- Native format optimization
- International reach if you need it
Real Cons:
- High minimum payout ($500) and account minimums make small publishers unviable
- Account management is sparse unless you’re a large publisher
- Integration can be complex, especially for video
- They take a margin on top of what buyers pay (like most networks)
- Performance heavily depends on your traffic quality — garbage traffic gets garbage rates
- Can take 60+ days to see full optimization
Skip it if: You’re below 5M monthly uniques or getting significant international traffic you want to monetize — you’ll struggle to meet minimums and their focus on large publishers will show.
Real story: I talked to a publisher with 500K monthly US uniques who went through Magnite onboarding. They were excited to access “premium demand.” Six months in, CPMs were mediocre and support was non-existent. They switched to PubGalaxy and made more money with better service. Size matters here — if you don’t have scale, you’re a low priority for Magnite. Nothing personal, just economics.
4. OpenX
OpenX has been in the market since 2000 and positioned themselves as the “publisher-first” alternative to bigger networks. Their claim is that they take smaller margins and focus more on direct relationships. Whether that’s completely true is debatable, but their reputation is solid.
What OpenX does really well is display and native formats. Their header bidding wrapper is clean, their integration is straightforward, and they’re good at explaining what’s happening. They also have a real partnership approach — they want to understand your content and audience.
Who it works best for: Mid-to-large publishers (500K to 10M uniques) with strong display inventory. News sites, blogs, lifestyle publications. Also solid for publishers who want to run direct deals alongside their programmatic — OpenX handles that integration well.
Real CPM numbers: Tier 1 US traffic on display content typically sees $3-8 CPMs. Not as high as specialized networks for premium video, but more than respectable for standard display. Tier 3 traffic is $1.50-3. The sweet spot is established publishers with known audiences — that’s where OpenX demand is concentrated.
Key Pros:
- Genuinely fast integration — 2-3 weeks to deployment vs. 2-3 months elsewhere
- Transparent about their cut — they’re taking 15% not 30%+
- Actual publisher partnerships with named account managers
- Works well with PMP and direct deals
- Good mobile support
- Solid reporting and real-time alerts
Real Cons:
- Demand is weighted toward certain verticals (news, finance, tech) — other verticals get lower competition
- CPMs can be lower than specialized networks for premium inventory
- International demand is there but not as strong as competitors
- Requires minimum $250 payout which isn’t huge but filters some publishers
- Not the best for video-first publishers
Skip it if: Video is your primary revenue source or you’re primarily focused on non-English speaking markets.
OpenX is the comfortable middle ground. Not the highest CPMs, not the most cutting-edge tech, but reliable and well-run. If you want to spend your time building content instead of optimizing ad tech, OpenX is a solid choice.
5. Xandr (Microsoft)
Xandr is Microsoft’s programmatic advertising platform, which tells you something important: they have massive buyer demand (hello, Microsoft’s own ad network, LinkedIn integration, etc.). But they’re also enterprise-focused and complex, which has both benefits and drawbacks.
Xandr actually has smart people building it. Their machine learning for demand forecasting is genuinely sophisticated. But the platform is complicated, and the minimum account sizes keep out smaller publishers.
Who it works best for: Large publishers (20M+ uniques) who are already in the Microsoft ecosystem or have significant B2B audiences. Also relevant for publishers running direct deals with large brands that buy through Microsoft.
Real CPM numbers: Tier 1 traffic in the right verticals (B2B, finance, tech) pushes $5-10 CPMs. Tier 3 is $1.50-3. But honestly, the CPM numbers are less interesting than the fill rate and consistency. Xandr’s real strength is steady demand — your floors don’t drop as much during slow periods.
Key Pros:
- Microsoft’s buying power means consistent demand
- Sophisticated audience targeting and buying algorithms
- Good video and native support
- Strong in B2B and tech verticals
- LinkedIn integration adds unique targeting capabilities
Real Cons:
- Minimum $1000+ monthly payout keeps out most publishers
- Platform is genuinely complicated — need dedicated ad ops person
- Learning curve is steep
- Support is focused on enterprise customers
- Integration takes 2-3 months typically
- Not ideal if you’re trying to keep things simple
Skip it if: You don’t have a dedicated ad ops resource or you’re below 10M monthly uniques — you’ll be frustrated with the minimums and the complexity.
Xandr is powerful for the right publisher, but wrong for most. It’s like buying a Ferrari when you want a reliable sedan. The performance is there if you can drive it.
6. Index Exchange
Index Exchange has positioned themselves as the “premium tech platform” for header bidding. They focus on quality over volume and have built strong relationships with sophisticated buyers. Their demand sources skew toward real-time bidders and agencies that actually care about what they’re buying.
What’s interesting about Index is that they’re very transparent about removing low-quality demand. They actively audit their bidders and kick people off if the demand is sketchy. This is good for your inventory quality, but it means lower overall auction participation in some cases.
Who it works best for: Tech publishers, publishers with high-engagement audiences, and established publishers in premium verticals (finance, news, lifestyle) with quality traffic.
Real CPM numbers: Tier 1 US traffic on high-engagement content regularly hits $5-9 CPMs. Premium tech content sometimes breaks $12. Tier 3 traffic is more like $1.50-3. The differentiator here is that Index’s floor CPMs are higher than some competitors — they won’t participate in garbage auctions.
Key Pros:
- Quality demand means better bid competition at your actual floor
- Less bot traffic and fraud compared to larger networks
- Good tech publisher relationships
- Native and video support
- Transparent about demand and bidder quality
- Better than industry average for payment reliability
Real Cons:
- Smaller demand pool means fewer bidders in some auctions
- CPMs can be lower on low-engagement content
- Account management is decent but not best-in-class
- International demand is weaker than competitors
- Requires minimum $500 setup
Skip it if: Your traffic is low-engagement or you’re in verticals they don’t have strong demand for (gambling, adult content, etc.).
Index Exchange is where quality-conscious publishers should be. You’re not getting the absolute highest CPMs, but you’re getting smart money and less fraud. That’s worth something.
7. Rubicon Project
Rubicon Project (now part of Magnite, but still operating under the brand) is the OG supply-side platform. If you’ve been in programmatic advertising for more than five years, you’ve been in Rubicon auctions. They were one of the first platforms to do header bidding well.
Here’s what’s important to know: Magnite owns Rubicon now, so there’s some consolidation happening. But Rubicon still operates with its own account team and brand. They’ve actually kept the publisher-focused approach even after the acquisition, which is surprising.
Who it works best for: Established publishers who’ve been in programmatic for a while. Video publishers especially. If you have existing Rubicon relationships, staying with them makes sense. Also good for publishers with video inventory and live streaming.
Real CPM numbers: Tier 1 video content hits $7-11 CPMs. Display is more like $3-8. Tier 3 traffic is $1.50-4. The video performance here is particularly strong — they’ve invested heavily in video demand.
Key Pros:
- Strong video demand and optimization
- Established relationships with major buyers
- Separate account teams from Magnite (still getting personal attention)
- Good reporting and transparency
- Flexible private marketplace deal support
- Can handle high-volume inventory without issues
Real Cons:
- Minimum $750 monthly payout keeps out small publishers
- Setup can be slow (60+ days to optimization)
- Owned by Magnite, so some uncertainty about long-term independence
- Account management quality varies depending on publisher size
- Can be overwhelming if you’re new to programmatic
Skip it if: You’re below 5M monthly uniques or don’t have video inventory — they’re optimized for established publishers with video.
Rubicon is the reliable veteran. Not flashy, but proven. If they offer you an account, take it unless you have a specific reason not to.
8. Criteo
Criteo is the retargeting and ecommerce specialist. If you have product-focused or ecommerce content, they’re worth considering because their demand is specialized for that audience segment. Their whole platform is built around intent and purchase behavior.
But here’s the thing: Criteo’s header bidding product is good, but it’s not their main value proposition. Their real strength is their first-party data and retargeting capabilities. If you’re using them just for general header bidding, you’re not getting their full advantage.
Who it works best for: Ecommerce publishers, product review sites, publishers with shopping-focused content, and anyone trying to monetize intent-rich audiences.
Real CPM numbers: On ecommerce and product content, you’re looking at $3-6 CPMs on average. Tier 1 audiences can hit $6-8. Tier 3 is more like $1-2. The key here is that Criteo has buyers specifically looking for ecommerce audiences, so their bids are relevant.
Key Pros:
- Access to ecommerce-specific demand not available elsewhere
- Good retargeting integration if you want to layer that in
- Works well with shopping content and product reviews
- Reliable payment and reporting
- Good for international ecommerce verticals
Real Cons:
- CPMs are generally lower than general demand networks
- Ecommerce focus means poor performance on non-product content
- Minimum $500 payout
- Account management is relatively hands-off
- Not the best for premium branding content
- Their margin is higher than some competitors
Skip it if: You’re not in ecommerce, shopping, or intent-based content verticals — you’ll see lower CPMs than general networks.
Criteo works great in their lane. Don’t use them as your primary demand source unless ecommerce is actually your focus. They’re excellent as a secondary network for that vertical.
9. Smaato
Smaato is the mobile-first and international specialist. If you’re getting significant traffic from mobile or international audiences, this is worth taking seriously. They’ve built their entire platform around mobile-first monetization and have strong demand from international buyers.
What’s interesting about Smaato is that they’re relatively transparent about their margins and they actively work with smaller publishers. They don’t have the 10M minimum that the big networks do.
Who it works best for: Mobile publishers, publishers with significant international traffic (especially Europe, Asia), and smaller publishers (500K-5M uniques) who want header bidding without the enterprise requirements.
Real CPM numbers: Mobile Tier 1 (US, high engagement) is usually $1.50-3. International Tier 1 is $0.80-2. Tier 3 is often under $0.50. I know these seem low, but that’s the reality of mobile and international demand. What Smaato does well is maximizing what’s available in those markets.
Key Pros:
- Excellent mobile optimization and demand sources
- Strong in Asia and European markets
- Lower minimums than competitors ($250)
- Works well with smaller publishers
- Good mobile SDK integration if you have apps
- International demand that actually converts
Real Cons:
- CPMs are generally lower than US-focused networks
- International demand, while strong, pays less per impression
- Not the best for premium desktop display content
- Account management is okay but not exceptional
- Reporting tools are functional but not cutting-edge
Skip it if: Most of your traffic is US desktop or you’re in a premium vertical where US demand is your primary revenue source.
Smaato is the right choice if you’re actually mobile and international. Don’t use them as a general play for US desktop traffic.
10. GumGum
GumGum is the contextual advertising and brand safety specialist. They’ve built their reputation on computer vision technology that understands content context without relying on cookies or personal data. For publishers concerned about brand safety and privacy, they’re interesting.
But here’s the reality: contextual advertising pays less than behavioral targeting. GumGum’s buyers are paying for placement safety and quality context, not intent. So you’re trading CPM for brand safety.
Who it works best for: Publishers in sensitive verticals (news, politics, finance), publishers who want cookie-free monetization, and publishers whose advertisers care more about context than targeting.
Real CPM numbers: Tier 1 contextual content (news, finance, tech) hits $2-4 CPMs typically. Premium news content sometimes breaks $6. Tier 3 is $0.50-1.50. The story here is consistency — you’re not getting wild CPM swings because they’re not chasing cheap behavioral demand.
Key Pros:
- Privacy-first approach (no cookie dependency)
- Excellent brand safety features
- Good for publishers concerned about user privacy
- Contextual targeting actually works (surprisingly)
- Minimum $300 payout is reasonable
- Works well in sensitive verticals
Real Cons:
- CPMs are generally lower than general demand networks
- Cookie-free approach limits advertiser targeting options
- Works best in specific verticals (news, politics)
- Not ideal for entertainment or lifestyle content
- Smaller demand pool than competitors
- Account management is decent but not exceptional
Skip it if: You’re trying to maximize raw revenue and don’t care about privacy or brand safety — you’ll leave money on the table versus behavioral networks.
GumGum is philosophy as much as platform. If you actually believe in privacy-first advertising, they’re the right choice. If you just want maximum CPMs, look elsewhere.
How to Actually Pick the Right Network(s) for Your Situation
Okay, I’ve given you the details on 10 networks. Now let’s talk about how to actually choose, because the answer is different for everyone.
Step 1: Know Your Traffic Profile
Before you even look at networks, understand your traffic. Specifically:
- How much monthly traffic do you get? (This filters out 60% of networks immediately — if you’re under 1M uniques, you can’t use Magnite or Xandr)
- What percentage is US vs. international? (This matters hugely for which networks will actually compete for your inventory)
- What percentage is desktop vs. mobile? (Mobile traffic changes your network strategy)
- What’s your content vertical? (Some networks specialize in certain categories)
- What’s your traffic quality like? (Be honest — premium finance sites and mass-market entertainment sites need different networks)
Step 2: Start with Your Foundation
If you’re not using Prebid, that’s your first priority. Prebid is infrastructure, not a network replacement. It’s like asking “should I use electricity?” — you should. Period. Get Prebid implemented first, then layer networks on top.
Step 3: Pick Based on Your Size Segment
If you’re under 1M monthly uniques: PubGalaxy, OpenX, Smaato, or Prebid with individual SSP connections. Magnite, Xandr, and Rubicon won’t work for you.
If you’re 1-10M monthly uniques: PubGalaxy, OpenX, Index Exchange, Rubicon (if video), or Magnite (if premium content).
If you’re 10M+ monthly uniques: Magnite, Xandr, Rubicon, Index Exchange, OpenX.
Step 4: Add Specialists for Your Vertical
After your primary network, add specialists:
- Video content: Add Magnite or Rubicon (even if you’re using OpenX as primary)
- Ecommerce: Add Criteo as secondary
- Mobile-heavy: Add Smaato
- International: Add Smaato
- Privacy-focused: Add GumGum as secondary
- Tech/premium: Index Exchange as secondary
Step 5: Test, Don’t Commit
Set up test account with your primary network. Run it for 2-3 weeks before fully migrating. Look at:
- Are they actually hitting the CPMs they promised? (Usually they won’t, by 10-15%)
- Is their account management helpful?
- How complicated is the integration? (This matters for ongoing maintenance)
- What’s your fill rate like?
- Are they easy to communicate with?
Step 6: Implement a Monitoring Process
After you go live, you need to monitor continuously. Once a month, look at:
- CPM trends (up or down compared to last month)
- Fill rates (dropped bidder?)
- Payment timing (got paid on time?)
- Any account issues or support responsiveness
This is the part people skip, and it’s why they leave money on the table. Networks will deprioritize you if you don’t pay attention.
5 Common Questions About Header Bidding Networks
Q: Should I use multiple networks or just one primary?
A: Multiple is better, but with caveats. You want 2-3 networks, not 10. Too many networks create timeout issues and weird auction dynamics. I typically recommend one primary (based on your size/type) plus one specialist. That’s it. More than that is just adding complexity and fragmentation. If you’re using Prebid, you can connect to multiple SSPs directly, which is better than stacking networks anyway.
Q: How long does it take to see good CPMs after switching networks?
A: Plan for 30-60 days minimum. The networks need data to understand your traffic and audience. Buyers need to figure out what you are. For Magnite, Xandr, or Rubicon, expect 60-90 days. For smaller networks like PubGalaxy or OpenX, 2-4 weeks is more realistic. Don’t panic if CPMs dip in week 1-2. That’s normal as the network figures out your demand.
Q: What’s a “good” CPM I should expect?
A: This depends entirely on your traffic. US desktop display content in finance/tech/news: $4-8 CPMs is good. International or mobile: $0.50-2. Entertainment/lifestyle: $2-5. Niche or low-traffic verticals: $0.50-2. The only real benchmark is your own historical data. If you were averaging $3 and now average $2.50, you’re doing worse. If you went from $2 to $3, you’re winning. Compare your own trends, not industry stats.
Q: Can I use the same network for display and video?
A: Technically yes, but don’t. Video demand and display demand are completely different. Video buyers have different budgets and preferences. Use your primary network for display, then add a video specialist (Magnite, Rubicon, or Spotx). Or better yet, use a video ad server that connects to multiple video networks simultaneously. This is one area where you should have specialized infrastructure.
Q: What happens if a network underperforms — can I switch quickly?
A: Kind of. You can pause them and test another network in parallel, which takes 2-3 weeks. Full migration takes 30-45 days to stabilize at the new network. This is why choosing right the first time matters — switching costs time and potentially revenue. You want to get this decision 80% right initially, then optimize, not jump ship constantly. That said, if a network is actively underperforming or not responsive, yeah, switch. Life’s too short for bad partnerships.
Final Recommendation: What I’d Actually Do
If I were starting from scratch with a mid-sized publisher (1-10M uniques) in the US with mixed content, here’s what I’d implement:
Foundation: Prebid.org as the header bidding framework. This is non-negotiable. Yes, it requires technical setup. Yes, it’s worth it. You’ll recoup the setup cost in a month.
Primary network: PubGalaxy if you want personal service and solid rates. OpenX if you want faster implementation. Magnite if you have premium inventory and can hit their size minimums.
Secondary network: Index Exchange if you want quality demand, or Smaato if you have significant mobile/international traffic.
Specialty: Rubicon if you have video, Criteo if you have ecommerce, GumGum if you care about privacy and brand safety.
That’s it. Don’t over-complicate it. 2-3 networks + Prebid gets you to 90% of optimal revenue without the complexity.
The real secret is that you’ll see more revenue improvement from implementing one network correctly than you’ll see from using five networks poorly. Focus on doing one thing well. Then optimize. Then add the next piece.
And honestly? The differences between these networks are smaller than the differences between good implementation and bad implementation. A publisher using OpenX well will make more money than a publisher using Magnite poorly. The fundamentals of traffic quality, audience understanding, and technical setup matter more than which network you pick.
Pick one, implement it right, monitor it monthly, and optimize. That’s the 80/20 play. Everything else is just noise.
