Should ROAS be calculated before or after the 30% platform fee?

Question

We currently log the revenue total we get from the Apple and Google APIs to attribution to track revenue.  I can see an argument for ROAS reporting dashboard to take 30% off the amount before logging.  That way ROAS is based on net revenue.  Anyone try this before?

Comments

Is the 30% a monthly expense? i.e. does Apple make you pay 30% every month or at the end of year?

Answer 1

After! Always! That 30% isn't your money, so it isn't ROAS.

Comments

So to be clear: you would log IAP revenue *.7 from the client code to the attribution API so the reporting dashboard ROAS number is "correct"?

---

I might be misunderstanding: you're talking about the revenue amount you send to an attribution provider? In that case, no, I'd send them the actual amount, but in my internal analytics I always use the 0.7x number to define ROAS.

---

That's the reason for the question.  A reporting tool like facebook is platform agnostic, so the ROAS column is based on the what ever revenue it receives.  I was hoping for a nice way to not always have to do mental math when looking at the dash.

---

I had a similar question actually, we currently send the entire revenue (i.e. $100 of Gross IAP, are sent as $100) however have wondered if I am better off only sending 70% of that amount. With a move towards automated campaign optimisation, I wonder if sending the gross amounts possibly gives FB and other networks a "false positive" on the ROAS front in some circumstances where you may have made a slight positive gain, but not enough to offset the store fee.

Example:

Spend $100 on FB ads
Send FB Events with $120 in Revenue (Gross, before store fees)
FB Reports as +20% ROAS and their algorithm theoretically deems it a success.
Reality is, after taking off store fees you have only made $84, so actually negative ROAS relative to the original $100 cost.

The alternate is, we only send $84 (the net amount) of revenue to begin with, and FB now has the signal that the campaign is not yet ROAS positive and either optimises accordingly, or just ends up failing to deliver enough volume as the ROAS goal is too aggressive for that particular product.

---

@stephen you could use Facebook ad reporting to avoid mental math. That’s a lesser known feature that allows you to use custom formulas (they introduced it at some point this year)

You can either go to ads reporting from the top menu or select some items and use “export - ads reporting”.

---

@jdask In theory it shouldn’t matter much. If you’re using autobid with value, the algorithm’s goal is to give you the highest value (D1 or D7 revenue) which ostensibly should result in the highest roas. So it shouldn’t use positive returns as a criterion (although “ostensibly” is of course the key word here)  

And if you’re setting roas goals, you can adjust it to whichever revenue you have in there

So in short I doubt it matters, but it feels like there are no downsides with using 100% revenue (aside from the bad numbers in ads manager)

Answer 2

I have done this before. So besides sending the actual revenue amount, I also sent "estimated lifetime value" (ELV) which took into account net revenue, and expected retention based on country, platform & early user behaviour.

IMO, if you're going to make ad adjustment to the revenue metric, you might as well add more information about the value of the user.

We sent the "Subscribe" event and value with gross revenue and "Purchase" event with ELV.