Traffic SourcesPeter Founder中文原文

PropellerAds Paid Social Case Study: $128.95 CPM, 42,000 Payments

Cover: PropellerAds Paid Social Case Study: $128.95 CPM, 42,000 Payments

Case Breakdown · Issue 4

This one is a case study the official PropellerAds blog published in July: a media buyer running a gaming offer bought Paid Social traffic in Southeast Asia at a CPM as high as $128.95.

In 30 days that produced 42,000 payments at $13 each — 9% cheaper than Pop traffic over the same period.

Is traffic this expensive actually worth it? This case gives a pretty blunt answer. Let’s take it apart step by step.

01 The numbers

Here is the data for the full 30 days from June 10 to July 9:

MetricValue
GEOSoutheast Asia
Traffic sourcePropellerAds · Paid Social Traffic
Period2026/06/10 – 2026/07/09 (30 days)
Pricing modelCPA Goal
Spend (test budget)$547,268
Impressions4.2 million
Payments42,097
Average cost per payment$13.00
CPM$128.95
Offer typeGaming (PWA funnel)

Targeting: mainly Android, mobile only.

02 Why a $128.95 CPM didn’t rattle him

Start with the funnel. He was running a PWA funnel: the user sees the ad → lands on the landing page → installs the PWA → starts using the product.

The key is what he watched: across the whole funnel, the only number he cared about was the cost of the first payment.

How many clicks, how many installs — he didn’t much care. What he checked every day was how many people actually paid and what each payment cost.

Bidding was handed entirely to CPA Goal, so the system chased cheap payments, not cheap impressions. Seen that way, a $128.95 CPM was just an intermediate number along the road, and it didn’t scare him.

03 Three GEOs pulled apart: how he moved the budget

At the start he tested three countries at the same time. After a month it looked like this:

GEOPaymentsCost per paymentShare of volume
GEO 136,504$11.8979%
GEO 24,831$18.7716%
GEO 3762$29.735%

GEO 1 jumped ahead in the first week and stayed in first place the whole month.

GEO 2 was weaker: 4,831 payments at $18.77 each.

GEO 3 looked the worst: only 762 payments at $29.73 each — almost 3 times GEO 1.

The gap was already there in week one, and how he handled it is worth noting: he didn’t simply switch GEO 3 off, he just stopped adding budget to it and fed the savings to GEO 1 and GEO 2.

04 Result: head-to-head with Pop traffic

30 days, 4.2 million impressions, in exchange for 42,097 payments at an average of $13 each.

Put it next to the Pop traffic running over the same period and it gets more interesting:

MetricPaid SocialPopDifference
CPM$128.95$72.261.8x higher
Cost per payment$13.00$14.289% cheaper
Payments (30 days)42,09737,002+14%
Payments per 1,000 impressions9.95.1about 2x

The expensive CPM didn’t drag things down. Look at it the other way: behind the pricier impressions is most likely a more active audience, more willing to pay — for the same thousand impressions, Paid Social produced twice as many payments as Pop.

“A high CPM can still be a great deal. This campaign’s CPM was 1.8x higher than the Pop traffic benchmark, yet the cost per payment was 9% lower. On top of that, one strong GEO is enough to carry the whole campaign — it took the lead in the first week and kept it through the end of the month.”

— Oleksandr Shovkun, PropellerAds Sales Team Leader

05 Four lessons learned

  • Don’t be scared off by intermediate metrics. CPM is only the price along the way; what a buyer is really paying for is that final step, the paid conversion. Anchor your bid on the final payment event and expensive impressions can’t scare you.
  • Expensive traffic is often good traffic. The CPM was 1.8x higher, yet payment density was twice Pop’s — you’re not the only one willing to pay more for these users.
  • Budget follows the data; no need to wait a full month. The gap between GEOs shows in the first week, so move money from the weak ones to the strong ones promptly.
  • Don’t switch it off, just stop feeding it. GEO 3 wasn’t cut in one stroke; the extra budget simply stopped. Leaving a tail running to keep observing is calmer than cutting it outright.

06 Wrapping up

Paid Social traffic can perform, and the reason isn’t complicated: these users are already immersed in a social setting, not just idly scrolling, so they’re more willing to click and convert.

Gaming works in Southeast Asia, but that doesn’t mean only gaming can run.

This format is worth a round of testing with your own offer — the scaling logic is the same as above: anchor the bid on the final conversion and let CPA Goal absorb the expensive middle.

Income and ROI figures here are past results of one person or case, not typical outcomes. Paid traffic can lose money; start with a small test budget.