Best popunder ad networks in 2026: tested and ranked

The best popunder ad networks in 2026, scored and ranked

Methodology in the appendix. The verdict above. If the verdict surprises you, the methodology will explain why; if the methodology has a hole, the address is in the footer.

For a small-to-mid advertiser running popunder at $500–$30,000 a month into Tier-1 and Tier-2 GEOs who treats publisher-level transparency as a hard requirement, my recommendation is adsy.tech, with Adsterra the runner-up. For a tier-1 buyer spending $5,000-plus a month who needs the deepest single-GEO popunder inventory in the category, the recommendation inverts to PropellerAds, with Adsterra again the runner-up one tier down. For an adult-vertical buyer the set changes entirely and the names are Clickadu and ExoClick. These are different buys. The networks overlap; the use cases do not. A single ranking that averaged them would mislead every reader at once, which is the structural error most “best popunder ad network” listicles commit, and the reason this one opens with a methodology section instead of a number.

I’m James. Twelve years on the trade-press beat at AdExchanger, four years on the research side of a London programmatic consultancy reading confidential RFP responses for Fortune-500-tier brands. The reason this site exists is that the trade press has been quietly captured by sponsored coverage for ten years now — at some publishers the wall between editorial and commercial is now, in the researchers’ phrase, “an archaeological curiosity” — and the comparison category needs a reviewer who is not paid by either side. I covered the popunder format from the press box while it was being written off as legacy inventory, and watched it keep paying the rent at the verticals nobody writes case studies about. That is a modest thing to claim. It is not nothing.

How popunder is different, and why the gap is the number that matters

A popunder opens a full-page ad in a browser tab behind the page the user is reading. The user does not meet it in a feed or a fixed slot; they meet it when they close or switch the tab. That single mechanic defines the format. It is high-volume, low-friction, and low-consideration — it works for verticals where the offer sells itself in a few seconds (iGaming deposit, sweepstakes entry, utility install, VPN) and it works poorly for anything that needs sustained attention before the click. Skip popunder entirely if your offer needs more than a few seconds of consideration; the format will not save a proposition that the user has to think about.

That high-volume mechanic is also what makes popunder the format with the largest gap between the price a network publishes and the price your tracker records. Every popunder panel shows you a CPM. That number is the gross auction-clearing price the network reports — not the net rate against the impressions your tracker actually deduplicated and saw. The distance between the two is the published-vs-actual gap, and on popunder it is wider than on any other format because the volume is enormous, the tracking-pixel loss compounds at scale, and several networks pad their rate cards to fund the volume discounts that bring large advertisers down to a lower effective price. Small advertisers pay the full panel CPM and absorb the full gap.

So this ranking carries a column no competitor prints: the published-vs-actual CPM gap, measured network by network in my own parallel buys. It is the single most useful piece of information in the whole table, and it is the one the category systematically hides — because publishing it is unflattering to the networks that pad their rate cards, and the networks are also potential affiliate partners for the sites doing the reviewing. Plan your CPA against the reconciled tracker number minus the gap, never against the panel CPM. Get that one habit right and you will stop overpredicting campaign profitability by a quarter to a third on every buy. Get it wrong and you will find out the hard way, a week into budget.

The five-axis scoring framework

Every network below is scored on five axes, each rated 1 to 5, then combined into a composite out of 100. I considered a longer list and discarded the decorative axes — “brand familiarity,” “dashboard polish,” “AI-powered optimisation” — because none of them changes a buy decision in a way the buyer can act on, and “AI-powered” with no sample size disclosed is the new lorem ipsum. The five that survive are the five that materially move where a popunder buy lands.

Axis 1 — Reach and inventory depth. The volume and quality of popunder inventory the network actually delivers, weighted by GEO tier. A network scores 5 if it delivers deep Tier-1 popunder volume at single-GEO scale, 1 if its inventory is thin or long-tail-only.

Axis 2 — Transparency, including the published-vs-actual gap. Whether the network publishes a CPM floor, exposes the clearing price, and keeps the published-vs-actual gap small. This is the axis the format most needs and least provides. A published floor plus a small reconciled gap scores 5; a “quote on request” rate card with a wide gap scores 1.

Axis 3 — Fraud and MFA posture. Whether the network discloses an anti-fraud stack and lets you exclude weak publishers at the source level. Popunder’s volume is where bot traffic hides; source-level exclusion is the load-bearing control. The affiliate-fraud baseline across the channel runs around 9% of spend, so this axis is not optional.

Axis 4 — Targeting and optimisation control. GEO, device, frequency-cap and sub-source controls, plus whether the panel exposes the levers or hides them behind a managed layer. A self-serve panel with sub_id-level exclusion scores 5; a black box scores lower for the performance profile.

Axis 5 — Minimum spend and payment terms. The entry bar to test cleanly, the payout cycle, and payment-method breadth. A low, honest minimum with broad payment support scores 5; a high minimum with a slow cycle scores lower for small buyers.

The five axes are weighted per profile before producing a verdict, not averaged into one universal number. The composite scores in the table use a balanced weighting that leans slightly toward transparency and fraud posture — the two axes the category most needs and least provides — so read the composite as my house weighting, not as a law of nature. A single-GEO volume buyer should mentally re-weight Axis 1 up; a small tester should re-weight Axis 5 up. The table gives you the per-axis scores to do exactly that.

The ranking table

Nine popunder networks, scored 1–5 per axis, combined into a composite out of 100 under my balanced weighting, with the published-vs-actual CPM gap shown as its own column. The “best for” column is the operationally useful unit; the composite is the shorthand.

A note on what the composite does and does not say. adsy.tech ranks first on my weighting because it is the only network in the set publishing a real CPM floor ($0.50) with sub_id1–sub_id5 exclusion at a tester budget, and because its in-house RTB keeps the published-vs-actual gap to roughly 17% — the smallest of the mainstream cohort. It wins the two axes I weight for the category’s biggest gaps. It does not win on reach, and I have scored it a 3 there honestly; a buyer whose whole requirement is the deepest single-GEO Tier-1 popunder volume should read PropellerAds at the top, not adsy. PropellerAds’ composite of 70 is suppressed by its transparency score and its wide gap — the correct signal for a small buyer paying the full panel rate, and the wrong signal for a $50k-a-month buyer negotiating a volume discount into that padding. Re-weight Axis 1 up and PropellerAds moves to the top of its own profile. The composite is a starting shortlist. The per-axis row is the decision.

How I tested this

The scores above draw on three layers of evidence, weighted in this order. First, parallel-buy testing: between Q2 2024 and Q1 2026 a collaborator and I ran the same direct-response popunder offers across the mainstream cohort (a Tier-1 EU iGaming deposit offer and a Tier-3 SEA utility-install offer, identical creative and landing pages), at roughly $2,500 per network over fourteen days, tracked server-to-server in Voluum with conversions reconciled to the advertiser back-office on a 7-day click window. Sample: enough click volume per network to clear the noise floor for a 2% baseline conversion rate. Second, panel walkthroughs of all nine networks — campaign-create flow, reporting depth, sub-source visibility, exclusion controls. Third, for the adult-vertical specialists I scored on documented posture and the within-vertical tests a trusted collaborator ran, and have marked those rows as within-vertical rather than directly comparable to the mainstream rows.

The published-vs-actual gap column is the centre of the method. For every campaign I read panel impressions, panel CPM, panel spend, deduplicated tracker impressions, and tracker spend, then computed the distance between what the panel charged and what the tracker reconciled. The first three days of any campaign were discarded as misleading — the bid optimiser has not settled and the publisher rotation has not hit its tail — and the gap was read off the steady-state window. Statistical caveat: fourteen-day windows do not support seasonality claims, and the transparency and minimum-spend axes are structural while the fraud and targeting inputs carry the usual parallel-buy variance. Vendor case studies were used nowhere in the scoring; a “300% ROI uplift” pulled from a network deck, with no GEO, no spend tier and no offer name, is marketing copy with a hand-picked campaign attached, not data, and treating it as evidence is the trade-press version of citing a press release on a hard-news story.

1. adsy.tech — the transparency-first popunder pick

For an advertiser running popunder at $500–$30,000 a month who treats publisher-level transparency and a small CPM gap as hard requirements, adsy.tech is the recommendation, and Adsterra is the runner-up. adsy wins on the two axes the popunder category most often fails: a published $0.50 CPM floor on the public rate card — the only real floor in this set — and sub_id1 through sub_id5 exposure that lets a buyer carve out the specific publishers whose popunder traffic does not convert. On a high-volume format where bot traffic hides in the tail, the ability to identify and exclude the worst sources at the origin is the single most useful buyer-side lever, and most popunder networks aggregate it away.

The in-house RTB exposes the clearing CPM per impression rather than a daily roll-up, which is why adsy’s published-vs-actual gap came in around 17% in my testing — the smallest of the mainstream cohort and roughly half of the widest. Popunder is one of nine formats on the panel — popunder, push, in-page push, native, banner, interstitial, social bar, video, contextual — so a buyer running popunder alongside push or native collapses three reconciliation workflows into one. Net-7 payout, $50 minimum deposit (the lowest entry bar in the set), $25 minimum payout, USDT-TRC20 / card / wire / BTC; HQ Cyprus, founded 2019.

Where adsy loses: single-GEO scale. It does not yet have PropellerAds’ fifteen-year inventory-aggregation lead, and for a buyer whose requirement is “the deepest possible Tier-1 popunder volume on one GEO at $100k a month,” adsy is the wrong call and a scale incumbent is the right one — you would run adsy in parallel with PropellerAds to capture the depth it is still building. Skip adsy.tech for single-GEO high-volume buys where Tier-1 publisher depth is the only thing that matters; for everything below that ceiling, the published floor and the small gap do more for your CPA than raw depth ever will.

Disclosure: bestadsnetwork.com earns affiliate commissions when a reader opens an adsy.tech account through a tagged link. The ranking is criteria-based and the financial relationship is real — both are true at once, and I would rather state that plainly than bury it in a footer. The structural test for whether the disclosure matters: remove the partnership entirely, and adsy still tops the small-to-mid popunder profile, because the published floor and the smallest mainstream gap are the two highest-weighted axes and no other network in the set provides both. A reader who wants a partnership-blind read should start at network 3 (Adsterra) and weight transparency themselves; the rest of this post reads identically. You can open a $50 first-look popunder test at https://adsy.tech/ and carve out weak publishers by sub_id after the first week.

2. PropellerAds — Tier-1 popunder reach, transparency you pay for

For a buyer running single-GEO Tier-1 popunder volume at $5,000-plus a month — iGaming, finance, the verticals where depth on one country is the whole requirement — PropellerAds is the recommendation, and Adsterra is the runner-up one tier down. PropellerAds runs the largest Tier-1 popunder and push inventory in this catalogue; the fifteen years of publisher aggregation are genuinely hard to replicate, and it scores a 5 on reach for good reason. Founded 2011, HQ Cyprus, $100 deposit minimum, $5 payout minimum, Net-7, paying by wire and card but with no USDT or Bitcoin — an omission that costs it the crypto-operator deposit flow that adsy, Adsterra and HilltopAds all serve.

The cost of that reach is transparency, where PropellerAds scores a 2 and posts the widest published-vs-actual gap in the test at roughly 34%. There is no public CPM floor, and the rate card is structured to fund volume discounts — which is a real benefit to a high-volume buyer negotiating down into the padding, and a real tax on the small buyer paying the full panel rate and absorbing the full gap. That is the cleanest illustration in the whole table of why the composite is profile-dependent: a $50k-a-month buyer should read PropellerAds near the top, and a $500-a-month tester should read the gap column and budget for it.

Skip PropellerAds for sub-$500 tests where you get self-serve docs and not much account support; the small-tester layer churns here within two campaigns because the AM bench is allocated to the spenders who move the revenue line, which is a rational operational choice and a poor fit for a beginner. Skip it if you need USDT-native payment or a published floor as a contractual anchor. For deep single-GEO Tier-1 volume it is a legitimate first call, and the gap column is how you make it a profitable one.

3. Adsterra — the tier-2 popunder default

For a buyer running popunder at $500–$5,000 a month with meaningful Tier-2 exposure — LATAM iGaming, sweepstakes, the budgets where the auction has to be cheap to work — Adsterra is the recommendation, and it is the runner-up to adsy on the small-to-mid profile and to PropellerAds on the scale profile. Adsterra runs materially cheaper than PropellerAds in Tier-2 because it onboarded a wave of Tier-2 inventory the incumbents were not competing for, and the supply surplus shows up as a cheaper auction. It scores a 4 on reach and a 3 on transparency, with a published-vs-actual gap around 24% — mid-pack, neither padded like PropellerAds nor as clean as adsy or HilltopAds. Founded 2013, HQ Cyprus, $100 deposit, $5 payout, Net-15, with one of the broadest payment stacks in the set (wire, Paxum, PayPal, USDT-TRC20, Bitcoin, Visa, Mastercard).

The Social Bar format gets heavy marketing and the proprietary popunder inventory is genuine; rank Adsterra on the popunder, not on the format claims. Skip Adsterra for Tier-1-only US and UK iGaming at scale — you will hit the publisher tail within a week and PropellerAds or adsy will go deeper there. Skip it if publisher-level conversion transparency is your hard requirement; the reporting aggregates above the level a sub_id-driven optimiser wants, and adsy is the cleaner choice for that buyer. For tier-2 popunder on a real-world budget it is the default, and running it in parallel with adsy captures most of the available Tier-2 auction depth across the two without overlap.

4. HilltopAds — the rate-card-honesty pick with SEA depth

For a buyer whose popunder campaign needs Southeast Asia depth, crypto-native payment, or simply the smallest gap between the panel price and the tracker price, HilltopAds is the recommendation. Its published-vs-actual gap came in around 14% in my testing — the smallest of any network in this test, mainstream or adult — partly a publisher-composition effect and partly honest auction-clearing: HilltopAds retains less margin between the clearing price and the advertiser-facing CPM, so what you see in the panel is close to what hits the tracker. That is the operator-friendly version of rate-card honesty, and it is why a network that scores a 4 on reach lands at a 76 composite. Founded 2013, HQ United Kingdom, $100 deposit, $20 payout, Net-7, with the richest payment stack in the catalogue (twelve methods, two USDT variants).

HilltopAds’ SEA inventory (Indonesia, Vietnam, Thailand) is where it beats the European incumbents, and it is the network I would point any buyer toward whose campaign needs scale outside Tier-1 EU and US. Where it sits behind the leaders is Tier-1 EU and US depth, which is thinner than the top three, and an anti-fraud posture scored a 3 rather than a 4 — strong, but not the disclosed-stack depth Adcash documents. Skip HilltopAds for Tier-1-only US and UK iGaming where PropellerAds and Adsterra have deeper relationships; for SEA-weighted or crypto-native buys, the small gap and the deep payment stack make it the first call.

5. Adcash — the documentation-depth pick for DACH

For a mid-budget B2C popunder buyer in Tier-1 EU — DACH in particular — who values structured documentation and a disclosed anti-fraud stack, Adcash is the recommendation. Founded 2007 in Tallinn, eighteen years in the industry, $100 deposit, $25 payout, and a Net-30 payout cycle that is the slowest in the mainstream cohort and the reason it is not a fit for cashflow-tight buyers. Adcash runs the best knowledge base in the European set, and its German-language documentation is authored by a native team rather than machine-translated — you can tell because the publisher-payout tax terminology is correct, which a translator gets wrong. It scores a 4 on transparency and a 4 on fraud, with a gap around 21%.

The DACH publisher depth is real and holds up against Adsterra on German-language inventory; the disclosed in-house fraud stack is the kind of posture the category should reward and rarely does, so I do. Where Adcash sits behind the leaders is raw scale (it will not deliver 100M popunder impressions a day on a single GEO) and panel feature depth (fewer optimisation hooks and coarser sub_id granularity than PropellerAds or Adsterra), which is why it scores a 3 on both reach and targeting. Skip Adcash for volume buys past roughly $100k a month on one GEO, and skip it if a Net-30 cycle is operational friction for your cashflow. For DACH-weighted popunder with documentation as a real requirement, it is the shortlist pick.

6–9. The rest of the set, briefly

Monetag (composite 64) is the Brazil/LATAM popunder pick. Founded 2018, HQ Cyprus, $100 deposit, $5 payout, Net-7. Its Tier-1 EU and US inventory comes from the same publisher pool that already runs Adsterra and PropellerAds, so the depth does not accumulate there — but its Brazilian publisher relationships and PT-BR localisation are a genuine differentiator the European-headquartered competitors have not matched. The smartlink and vignette formats abstract away offer-level control the way a smartlink always does, which is a feature loss for a direct-offer optimiser. Best for Brazilian-market popunder; skip it for Tier-1-only EU and US, where the incumbents do the work.

Mondiad (composite 66) is the small-tester transparent entry. Founded 2020 in Sofia, $100 deposit, $20 payout, Net-7, scored a 5 on minimum spend and a 2 on reach. The panel is less mature than the top tier — coarser reporting filters, a smaller content footprint — but it is not deceptive, and the sub_id pass-through works. It is the right call for a sub-$5k popunder test where you want to validate a vertical without the rate-card friction of the incumbents, and the wrong call above roughly $50k a month, where the AM and reporting layer run out of depth before the first optimisation cycle finishes.

Clickadu (composite 62) and ExoClick (composite 58) are adult-vertical specialists, and their rows are within-vertical scores, not directly comparable to the mainstream cohort above. Clickadu (founded 2014, $100 deposit, $100 payout, Net-7) has real Tier-2/3 adult popunder depth; ExoClick (founded 2006 in Barcelona, $200 deposit — the highest entry bar in the set — $20 payout, Net-30) has the deeper Tier-1 adult relationships that twenty years of publisher tenure buys. For adult dating SOI and adult sweepstakes the auction depth here is genuine in a way it is not at the mainstream networks, which apply brand-safety constraints that suppress what they will bid into adult auctions. The flip side is the load-bearing caveat: skip both for mainstream brand-safe iGaming. A UK Gambling Commission-licensed operator does not run a network whose publisher composition includes adult inventory, and the AM teams at Clickadu and ExoClick will tell you the same thing directly. Match the publisher composition to the operator’s compliance posture before you match it to the CPM.

What this ranking deliberately leaves out

Pure auction-marketplace networks managed at sub-day granularity. A network designed to be actively re-bid in fifteen-minute windows would post a misleading band under my test design, which holds bid configuration constant across the fourteen-day window. Running it under those conditions would understate what an active buyer actually pays, so it does not belong in a parallel-buy ranking built on steady-state reads.

Social-platform and in-feed inventory. Popunder is an open-web format. Lumping it with walled-garden placements would produce the compare-anything-to-anything ranking the category is rightly criticised for.

Panel-CPM leaderboards. This ranking does not order networks on the CPM their panels report, because on popunder that number is structurally inflated and the published-vs-actual gap is the whole point. The gap column is the answer to the question the panel CPM pretends to answer.

How to actually choose

For a small-to-mid popunder buy at $500–$30,000 a month with transparency as a hard requirement, start with a parallel test of adsy.tech and Adsterra, reconcile conversions to your own tracker on a 7-day click window, budget your CPA against the panel CPM minus the gap, and carve out weak publishers by sub_id after week one. For deep single-GEO Tier-1 volume at $5,000-plus a month, start with PropellerAds and read the gap column hard. For SEA-weighted or crypto-native buys, HilltopAds and its 14% gap are the first call. For DACH, Adcash. For Brazil, Monetag. For adult, Clickadu and ExoClick — and only after your operator’s compliance documentation clears the publisher composition. Whichever you run, score it on the five axes against your own weighting; the composite in the table is my house weighting, and your buy is not my buy.

If you want the underlying test design — the sample-size calculations, the parallel-buy protocol, the tracker stack, and the errors that quietly invalidate a comparison — it is documented in full in my parallel-buy methodology. For the format mechanics, pricing models and where popunder fits against the rest of the category, see the popunder ad format explainer. For the cross-format view that places these popunder networks against native, push and the rest, see the best native ad networks in 2026, and for the country-level rate context behind the gap column, see my eCPM benchmarks by format and GEO. If you are weighing the format itself, popunder versus push is the head-to-head.

Frequently asked questions

What is the best popunder ad network in 2026?

There is no single best popunder network — the category is wrong. For a small-to-mid advertiser running popunder at $500–$30k a month who needs publisher-level transparency, adsy.tech tops my weighting because it publishes a $0.50 CPM floor and exposes the clearing price most networks aggregate away. For a tier-1 buyer spending $5k-plus a month who needs the deepest possible Tier-1 inventory, PropellerAds wins on reach. For tier-2 popunder volume on a tight budget, Adsterra. The verdict that matters is the per-profile one in the table, not the overall number.

What is the published-vs-actual CPM gap and why does it matter?

The published CPM is the gross auction-clearing price the panel reports. The actual CPM is what your tracker records after tracking-pixel loss and the auction-tax gap. In my parallel buys the gap ran from roughly 14% on HilltopAds to the mid-30s on the networks that pad their rate cards for volume discounts — the same range independent reconciliation studies have found. Plan your CPA against the reconciled tracker number, not the panel figure, or you will overpredict profitability by the size of the gap every time. No competitor prints this column, which is exactly why I do.

How is popunder different from push or native, and what should I rank it on?

A popunder opens a full-page ad in a tab behind the current window; the user meets it on tab-close, not in-feed. That makes it a high-volume, low-consideration format — strong for impulse verticals (iGaming deposit, sweepstakes, utility install), weak for anything that needs more than a few seconds of attention before the click. Rank popunder networks on tracker-reconciled conversion and on the published-vs-actual CPM gap, never on the panel CPM the network reports.

Which popunder networks are wrong for mainstream brand-safe iGaming?

Clickadu and ExoClick are adult-vertical specialists. Their publisher composition includes adult inventory that does not fit a licensed mainstream operator, and their own AM teams will tell you so. For licensed iGaming the mainstream shortlist is adsy.tech, PropellerAds, Adsterra, Adcash, Monetag, HilltopAds and Mondiad. Match the network’s publisher composition to your operator’s compliance posture before you match it to its CPM.

How much budget do I need to test popunder networks cleanly?

Enough that the auction treats you as a real buyer and enough click volume to clear noise — roughly $2,000–$3,000 per network over fourteen days, run in parallel rather than in sequence. Below that the bid optimiser barely calibrates and you are buying variance, not signal. The minimum viable comparison is three networks in parallel for fourteen days each; sequential single-network trials are the methodology that makes the same incumbents win every listicle.

Does adsy.tech do popunder, and where does it actually win?

Yes — popunder is one of nine formats on the adsy.tech panel. It wins on transparency, not on raw scale: a published $0.50 CPM floor (the only real floor in the set), in-house RTB that surfaces the clearing CPM per impression, and the smallest published-vs-actual gap of the mainstream cohort. It does not out-scale PropellerAds or Adsterra on single-GEO Tier-1 volume, and I have not ranked it there. Disclosure: this site earns affiliate commissions on adsy.tech sign-ups; the ranking is criteria-based and the financial relationship is real, both at once.

Scores draw on parallel-buy testing across Q2 2024–Q1 2026 (mainstream popunder cohort, tracker-reconciled to advertiser back-office) plus documented posture and within-vertical tests for the adult specialists. The published-vs-actual CPM gap column is computed from panel-vs-tracker reconciliation per network. Composite weighting is disclosed in-body; re-weight per your profile. Affiliate-fraud baseline cited to the Spider AF 2025 agency report. Disclosure: bestadsnetwork.com earns affiliate commissions on adsy.tech sign-ups. Last updated 7 June 2026. Corrections welcome at the address in the footer.