Push traffic is one of the cheapest paid sources in affiliate marketing: a click costs a fraction of a cent, a push network deposit starts at $50, and you can launch a campaign in one evening without accounts, proxies or an antidetect browser. But the golden age of push with 200%+ ROI is long gone. We collected fresh cases and threads from CPA.rip, Partnerkin, AffLift and BlackHatWorld, push network terms and official browser changes, and broke down which networks to pick, which offers to run and how to set up and optimize a campaign so you don't burn your first deposit.
In short: is push worth running now
- Yes, if your goal is a cheap and fast entry into paid traffic. A $50–150 deposit, bids from $0.001–0.005 per click, and easier moderation than Facebook or Google.
- Dating, sweepstakes, utilities and nutra work best in Tier 2–3. In the CPA.rip case base, out of 141 push cases, dating (30) and sweepstakes (25) lead.
- Real ROI in fresh cases is 8–50%, not 200%. In-Page dating in Japan made +39% on $200; a US sweepstakes test made +8.6% with CPA Goal and lost money on manual CPC.
- The subscriber base is shrinking: since October 2025 Chrome automatically revokes notification permissions from sites users don't engage with. Traffic quality now matters more than volume.
If you are just starting, first read the beginner's guide to traffic arbitrage and get comfortable with CTR, CR, EPC and ROI — you will need them below.
What push traffic is: classic, In-Page and other formats
Push traffic is clicks on ad notifications. A push network buys access to publishers' audiences and sells impressions to advertisers, usually on CPC (pay per click) or CPM (per 1,000 impressions). The ad looks like a regular notification: an icon, an image, a title and a short text.
- Classic web push. A user clicks "Allow" on a website, and the browser starts showing them notifications — even when the site is closed. This is subscription traffic: the network sends ads to its subscriber base. Classic web push barely exists on iPhone: Apple enabled it in iOS 16.4 only for web apps added to the Home Screen.
- In-Page push (IPP). A banner that looks like a notification but is shown right on the page — no subscription needed. That is why the format doesn't depend on browser restrictions and reaches iOS and macOS, where classic push is nearly absent. Most fresh CPA.rip cases are In-Page launches.
- Calendar push and other variations. Some networks, such as RichAds, sell notifications via calendar and similar formats. You don't need them to start — begin with classic or In-Page.
Inside a network, subscribers are segmented by activity and freshness — PropellerAds, for example, has User Activity segments. In a nutra case in Nigeria the author split Medium and Low activity into separate campaigns and excluded old Android versions — this is standard practice.
Is push alive: what changed and what forums say
The main change in recent years is that browsers started protecting users from notification spam. According to Google, less than 1% of all notifications get any interaction, so Chrome on Android and desktop now revokes the permission from sites that send many notifications and get no engagement. Users are notified and can re-grant the permission; installed web apps are not affected. Google notes that sites which started sending fewer notifications saw higher click rates.
For an affiliate this means two things: classic push bases are slowly shrinking, and networks that over-squeeze subscribers lose quality. That matches what forums say:
- In a March 2026 BlackHatWorld thread about which sources people scale now, members say push still works for dating and sweepstakes, but quality has dropped on most networks.
- In a "where to buy push traffic now" thread members recommend PropellerAds, Zeropark, AdMaven, RichAds and other networks, but agree on one thing: without bot filtering and quickly cutting weak placements, push doesn't pay off.
- The yearly PropellerAds Follow-Along contest on AffLift gathered 40+ public journals in 2025: mVAS, dating, CPI, VPN, casino. The fact that people are willing to show their launches with a $100 minimum spend is a good sign: the source is alive, but requires hands-on work.
Bottom line: push is not a money button but a cheap testing source where it's convenient to learn optimization and find setups for Tier 2–3 and mobile traffic.
Push ad networks: where to buy push traffic
We collected the networks forums recommend most often, with terms from their sites and the RichAds and ROIads reviews updated in 2026.
- PropellerAds — headquartered in Limassol, Cyprus. Push, In-Page, popunder; automatic SmartCPM and CPA Goal models. $100 deposit. The most frequent source in the CPA.rip case base and the sponsor of AffLift contests — so it has the most open breakdowns.
- RichAds — Cyprus, $150 deposit, push from $0.005 per click. Push, In-Page, Calendar push, pops. BHW members suggest it for Tier 3.
- RollerAds — Cyprus, the lowest entry at $50. Has SmartCPC and CPA Goal; the bidding model comparison case we cover below was run here.
- EvaDav — team in Warsaw and Prague, $100 deposit, Push, In-Page and native. One of the best-known archive cases — dating in France with 255% ROI — was run here.
- Zeropark — $200 deposit, push + pop + domain redirect. BHW suggests it for testing.
- AdMaven — $50 deposit, cheap traffic, but forum members warn directly: without aggressive bot filtering it burns budget.
Starter tip: pick one network with an automatic bidding model (CPA Goal or SmartCPC) and make the minimum deposit. Spreading $300 across three networks is a sure way to get statistics nowhere.
Which offers to run on push traffic
Push is impulse traffic: a person clicks a notification without planning a purchase. So offers with a simple conversion event work best — registration, e-mail, app install, cash-on-delivery order. The starting budgets by vertical below come from the AffMoment push guide.
- Dating (SOI/DOI, smartlinks). The main push vertical. Test budget — $600–900 in Tier 2–3.
- Sweepstakes and CPL. $0.2–3 payout for an e-mail or a form — in the AffLift contest breakdown the author worked with $0.20–1 payouts. Low payout is compensated by volume and automatic bidding.
- Utilities and apps (CPI). VPNs, cleaners, games. Test budget — $500–800. In a Partnerkin case a utility in India on In-Page with $0.001 CPC made $1,223 from $421 — 190% ROI.
- Nutra (COD). Test budget from $2,000, because of approval and buyout rates. In an AffTimes case the author notes the headline matters more than the image, and approval was 22.6%.
- Gambling and betting. Test budget from $5,000, licensed brands only and only in GEOs where it is legal.
Find offers in our list of CPA networks and in the article on how to choose a CPA network. Ask your manager which offers already convert from push in your GEO: in the AffTimes case the landing page was chosen on the manager's recommendation.
Smartlink or a specific offer? AffMoment calls smartlinks a beginner mistake: you don't see what actually converts. Yet two of the three fresh CPA.rip dating cases ran on smartlinks. A compromise: a smartlink for the first GEO test, a specific offer once you've found working placements.
How much money you need to start
- Deposit: $50–200 depending on the network.
- Testing one offer: per AffMoment — 5–10 offer payouts. With a $3 payout that's $15–30 per setup, with $10 — $50–100.
- Campaign daily budget: AdMaven's advertiser guidelines recommend at least $35 or three offer payouts — whichever is higher.
- A realistic first budget: $200–300 — as in the Japan case ($200) and the sweepstakes case (about $100 per bidding model).
If you have no budget at all, push is not the best start: begin with organic traffic using our guide to affiliate marketing with no investment, and move to push with your first profit.
How to run push traffic: step-by-step campaign setup
- Tracker and postback. Neither CPA Goal nor auto rules work without passing conversions back to the network. If conversions go missing, go through the postback error checklist.
- One campaign — one GEO, one OS, one device type. That's what PropellerAds experts advise in the AffLift contest breakdown, and AffMoment lists mixing desktop and mobile in one campaign as a typical mistake.
- Bidding model. Start with CPA Goal or SmartCPC. In the RollerAds test manual CPC made −6%, SmartCPC on test −33%, and CPA Goal +8.6% plus access to expensive zones SmartCPC doesn't reach.
- Bid. Use the recommended bid in the dashboard as your anchor. In the Japan case the author bid slightly above the recommended rate to get better placements.
- Targeting — broad at the start. AdMaven advises running on the whole available inventory (RON): narrow targeting gives too little traffic for statistics. The exception is obviously junk traffic such as old Android versions: they were excluded both in the Nigeria case and in the India case.
- Creatives. 2–4 variants for the test, a rotation of 5–12 in production (AffMoment). Title and text — up to 30 characters on mobile, up to 60 on desktop (AdMaven). No more than 1–2 emojis. Write in the GEO's language: in the AffLift contest lessons localized landers clearly beat English ones. More ideas in our article on creatives for affiliate marketing, and spy tools help you see other people's setups.
- Frequency cap. The In-Page Japan case used 1 impression per user per day; AffMoment suggests testing a frequency of 2–6.
- Daily cap. Always set a campaign cap and spread the budget over at least 3 days — that way you see different audience segments, not just morning traffic.
Optimization: how to cut zones without burning budget
Most of the work with push is cleaning placements (zones, sources). Rules from cases and breakdowns:
- Day 3–4: block zones that spent two offer payouts with zero conversions (PropellerAds × AffLift). With a $3 payout the threshold is $6.
- Day 7: block zones that convert but have a strongly negative ROI. Contest participant Sorrowz used a −20% threshold and blocked zones that ate 50% of the CPA with no conversions even earlier.
- Per-zone cap. In the Japan case — $8–10 per zone before evaluating it.
- Bot signal: high CTR with no conversions. In the same case, zones with 3.5%+ CTR and zero conversions after 150–200 clicks went to the blacklist.
- Auto rules. In the India case placements that spent more than $0.10 without a lead were blacklisted automatically every day, and sites with leads went to the whitelist.
- After stabilizing keep two campaigns: a whitelist on proven zones and a "everything except the blacklist" test to catch new sources — Sorrowz's advice.
- A simple conversion event gives CPA Goal more volume: in the AffLift breakdown, a "lander click" goal produced 3x higher eCPM than a registration.
Forum cases: what push really brings
In-Page → dating, Japan: $200 → $278 (CPA.rip, 2026)
A fresh case: a dating smartlink, In-Page format, Smart CPM, frequency of 1 impression per day, $8–10 cap per zone, creatives in Japanese ("Want to talk to someone tonight?"). Result — $78 profit and +39% ROI. The main mistake fixed along the way: zones with 3.5%+ CTR and zero conversions.
Push → $3 sweepstakes, US: three bidding models compared (CPA.rip, 2025)
The author tested CPC, SmartCPC and CPA Goal with ~$100 each on one CPL offer. CPC: $102 → $96 (−6%). SmartCPC on test — −33%, but handy for building a whitelist. CPA Goal with a $3 target: $108 → $117 (+8.6%). This is the honest picture of sweepstakes: a small plus that scales with volume.
In-Page → dating, India: three campaigns, three results (CPA.rip, 2025)
Over 11 days three campaigns on a dating smartlink in India's major cities made +49%, +42% and −27% ROI. Volumes — millions of clicks at 0.5–0.9% CTR. The key was automatic black- and whitelist rules checked every day.
Push → game (CPI), PropellerAds: $213 → $408
A case from the AffMoment guide: $194 profit and +92% ROI on a game CPI offer. The same guide has a $4,478 → $7,645 example on a dating smartlink, but from a different source.
Archive: dating in France with 255% ROI (Partnerkin, 2019)
A classic case on EvaDav: $1,506 → $5,355, 351 thousand clicks at ~$0.004, 1,392 leads. We include it as a benchmark of how push worked at its peak. Today's numbers are closer to +8–50%.
Typical beginner mistakes on push
- Launching without a postback and counting conversions by eye.
- Mixing desktop and mobile, several GEOs or OSes in one campaign.
- Narrowing targeting to a couple of browsers and cities right away — not enough traffic for statistics.
- Not setting a daily cap and burning the deposit overnight.
- Cutting zones after 10–20 clicks instead of waiting for at least two payouts of spend.
- Celebrating high CTR: on push it is often bots, not interest.
- Running without localization: an English creative in Japan or India almost always loses to the native language.
What not to do with push traffic
Push has a "grey" reputation because of creatives that imitate system warnings: "Your phone is infected", "Account blocked", "Update your browser now". Such ads are banned by network rules, and advertisers increasingly refuse to pay for that traffic. We recommend running only white and grey verticals — dating, sweepstakes, utilities, nutra, gambling in legal GEOs — and staying away from crypto scams, phishing and fake antivirus offers.
FAQ
What is push traffic in affiliate marketing?
It's clicks on ad notifications in the browser (classic push) or on notification-style banners on websites (In-Page push). The traffic is bought in push networks on CPC or CPM and sent to affiliate offers.
How much money do you need to start with push?
A network deposit starts at $50 (RollerAds, AdMaven) and goes up to $150–200 (RichAds, Zeropark). A realistic first test is $200–300 for one GEO and one offer.
How is In-Page push different from regular push?
Classic push requires a subscription, and notifications arrive even when the site is closed. In-Page is shown right on the page without a subscription and works on iOS, where classic web push is only available for Home Screen web apps.
Which offers work best with push?
Dating, sweepstakes and CPL, utilities and apps on CPI, nutra in Tier 2–3. Gambling and betting work too, but need a big budget and a legal GEO.
Which bidding model should a beginner choose?
CPA Goal or SmartCPC: they cut inefficient placements automatically. In a CPA.rip case CPA Goal made +8.6% where manual CPC went negative.
Is push traffic dead?
No, but it has shrunk. Chrome revokes permissions from spammy sites, quality has dropped on many networks, and ROI in fresh cases is 8–50%, not 200%+. For dating, sweepstakes and utilities in Tier 2–3 it is still a working and the cheapest paid source.
More push cases and setups for other sources — in our Telegram channel.
Sources: CPA.rip (push case base; cases: In-Page dating in Japan, US sweepstakes on three bidding models, In-Page dating in India), Partnerkin (EvaDav × dating in France, In-Page push: utility and dating cases, nutra in Nigeria), AffMoment ("How to run push traffic" guide), AffTimes (nutra on push case), AffLift (PropellerAds Follow-Along contest 2025; contest breakdowns on the PropellerAds blog), BlackHatWorld (threads on push networks in 2025 and traffic sources in 2026), Google Chrome Blog (automatic notification permission revocation), WebKit Blog (web push on iOS 16.4), AdMaven Help (push campaign requirements), RichAds and ROIads (push network comparisons), PropellerAds and EvaDav websites. Data checked in October 2026; case results are the authors' own reports, network terms may change.
