So there are times where I work with extremely low payout offers. When I say “extreme” I mean things that have a payout of less than 1 cent. Typical these offers are push collection offers whereby we get paid based upon a visitor subbing onto the advertiser’s push list.

For example, looking at my Evadav CPA rate for India:

$0.0025 for mobile in India… As I said, extremely low payout.

As a result when you work with extremely low payout you want scale, since 1,000 conversions will only get you less than $10. As such you’ll want to make sure you get a lot, a lot of traffic. To do that you can either bid high and get a lot of traffic or create a lot of campaigns across different traffic sources. Given that we can’t really bid very high with such a low payout, I went with the latter:

(Proof that I actually created these campaigns)

Okay, so creating a lot of campaigns is painful but what is worse is managing a lot of campaigns especially when each of you campaign is tiny i.e. spend $1 a day. I mean, just look at this:

(A portion of my camps on a random day)

A camp that spends $1 a day is not worth your time to look at. However if the campaigns can make you on average $1 a day on autopilot then if you duplicate it 100 times, that is already $100 in profit.

As such we need a very efficient way to work and autopilot each campaign.

Enter: Auto Optimization

Nowadays most traffic sources come with auto optimization. They ask you for your CPA and then they will auto-optimize it for you. Problem is all of them have a minimum CPA number, take PropellerAds for example:

Remember our payout for India is only $0.0025 so targeting the suggested $0.01 is out of the question. So what do we do?

 Binom’s Postback % Feature

To deal with this issue you’ll need Binom’s “Postback %” feature. Not sure if you know where it is as its hidden underneath “Create Campaign” → “Advance Settings”:

I’m assuming most affiliates won’t have used it. But it is actually a pretty useful feature in this case. By controlling the postback % you can effectively control the CPA. So for our case here:

India’s Payout = $0.0025

Let’s say we want ROI of 30%

So our Target CPA = $0.0025 / 1.3 = $0.0019

But because Propeller’s minimum CPA is $0.01 and we can’t change that we need this formula:

Minimum CPA x Postback % = Target CPA

Rearranging to figure out how to set postback %:

Postback % = Target CPA / Minimum CPA

Putting in the numbers:

Postback % = $0.0019 / $0.01 = 19%

You then just put the number in like this:

And then just set up your campaign as you would normally do. Presto!

The Magic: Explanation of What Happens

So what actually happens is that instead of sending every single conversion back like you would normally do Binom will only send back a portion of the conversions (in this case 19%). 

This is by no means as accurate as working with sending back the exact values to the optimizer, but given that you don’t have an option (because traffic sources force you to work with their minimum bids) I’d gladly use this option instead of manually optimizing myself.

The inaccuracy is also less of an issue when you’re working with 1000+ conversions daily (for people that studied stats: remember “law of large numbers”). So I wouldn’t worry about it too much.


Everyone loves seeing results so I thought I’d just pull up whatever I had in my tracker that still uses this method (it is currently managed by one of my virtual assistants):

Lo and behold, it seems there are some diamonds in the rough if you test enough (we have a few bigger spending campaigns, and I can assure you we are using the same method described above – it just took a ton of testing).

Additional Point

You’ll notice that I have a lot of events (since I mostly run pop traffic) – 2.5 million clicks. These costs can add up if you use a hosted tracking solution that charges per event; as such, I would recommend a self hosted tracking solution such as Binom. 

For any curious I use Hetzner’s PX92 that was recommended by Binom support and have had moments where I’ve run 10,000 clicks/minute (for a brief period of 1-2 hours) and things were still functioning.