How do you easily track and visualize early ROAS Metrics (d3, d7)?

There are a few companies that can do this for you well by integrating into your specific set up. I work at one that does this for example (tenjin.io) that can offer really good prices. We functionally can give you a mobile marketing data warehouse as a service so you don't need to build your own BI setup internally or from scratch. If you want, I can also tell you how to build this on your own.

If you don't want to invest in a DW/BI set up, you can do this manually in Excel with some level of accuracy on aggregate data. You generally only need to invest in this process once, and then daily you can just copy paste new data from your ad network and attribution partner to update the math.

If you download the d3/d7 LTV aggregates from your attribution provider for a specific daily cohort, pull in the daily cost data for that cohort, then you can line up those estimates pretty easily to calculate your ROAS.

you would have the following column set up:

cohort date | cost | ad network | campaign | country | dimension 1 | dimension 2 | dimension x | d3 LTV | d7 LTV
Every day you would just update all the necessary rows for each dimension.

You would then pull all these aggregates into a pivot table which you could build graphs on within Excel. Not elegant, but doable without building a data pipeline.

The advantage a data pipline would give you is all these steps could get automated into one SQL query that you push a button every morning and get the up to date results without any of the manual daily work.

You'd have to manually map it out for each day's cohorted revenue by source. It's labour intensive and definately not "free" in terms of time spent.

Basically on the 3rd and 7th day of each day's campaign you'd have to manually track that day's cohorted revenue.

Example: lets say you started a campaign with Unity on May 1st.

Then on May 3rd you'd have to measure May 1's life time revenue from Unity (restrict the time frame to only May 1 on appsflyer) and figure out D3 ROAS by:

May 1's Unity D3 revenue/May 1's Unity Spend to get D 3 ROAS.

You'd have to do the same for May 1's D7 revenue on May 7th.

Now do this for every network, every day the campaign is live

Ultimately the easiest way is to get BI involved and just have daily cohorted revenue tracked.

Mobile attribition partners (like Appsflyer, Adjust, Branch, ...) help you not only to differentiate organic from non organic users. If you increase UA spend and do UA via different networks (facebook, UAC, ....) you want to splice your non organics so that you can see where they did come from and towards which ad they acted. But this is not for free. You typically will be charged for non organic installs. In fact this cost will reduce your ROAS. If you do UA in territories with low CPI / LTV this cost can be significant in relation to your CPI/LTV (so ROAS might be lowered substantially).

But for the start you could take a different approach:

1) concentrate on ONE UA source AT A TIME (I would recommend to just stick with UA via facebook as a start) for a given interval

2) deduct the installs that the UA source claims in any given interval from all installs (signups) that you track in your database

Now you have a rough ratio between organics and non-organics.

Turn on and off the paid UA for an appropriate interval (say 1 or 2 weeks) so that you can estimate the ARPU/LTV/KPI of a pure organic user of your app.

Of course as you progress and optimize your app the KPIs will change (hopefully for the better). So you need to re-calibrate from time to time.

I'm not sure if the free Appsflyer plan supports it but there is an af_purchase event which is mandatory to at least see the revenue generated by a given user (organic vs non organic). You can use Appsflyer's automatic validation (against your Apple/Google account) or you may send the af_purchase manually (be aware that in that case you should do the validation on your server before your app calls af_purchase or you will see inflated, fraudulent revenues).