How to optimize the Waterfall in monetization to fullfill the maximum revenue?

I don't think there is a one-size fits all answer here. In the context of Rewarded Video, here are a few general tips:

• When it comes to how many price floors you have... Think big. For the big networks especially, they are going to have a ton of value to add at 3-5 different floors. 
• Assuming you are managing ad requests well on device, you should be able to support at least 20 floors without running into fill rate issues.
• For many games, Facebook and Google will be able to provide meaningful fill at $100 CPMs. The impact of adding floors in the $60-100 CPM ranges can be very impactful for some games.
• When deciding where to add price floors, look for areas with high fill. For example, a floor at $15 with a fill of 8-10% would almost certainly need more competition above it.
• Ask networks what they're willing to do for the top position in your waterfall. Depending on your scale, their answer may suprise you.

Hi Wendy,

I think I can give you some advice, but it will vary by what mediation platform you are using. Are you able to share that?

You need a mediation solution that supports AB testing the waterfall (MAX, Ironsource, and I think Mopub support this now).   

You use one waterfall set up on the A side and one on the B side.    If you have a decent size DAU this should be an ongoing process.  Like in most AB testing it's best to tackle it incrementatlly and not change too many variables at once.     A most simple example is on the A side you have your current waterfall which does not include ironsource and on the B side you add in several ironsource line items.   If B is a good lift in overal revenue for that placement that tells you to keep ironsource in.     Then you can put that new waterfall with ironsource as A and then for example you double you number of facebook line items on the B side.  Promote the winning side..repeat often.  If you have high scale you also want to be in contact with the network so they can evaluate their supply and work together to agree on price points to test that they expect to do well.

This is simplified as you add more in-app bidding partners to your mix.  Since they effectively compete at all price points you don't have to decide on a handful of fixed price points.

Josh's comment is all great detailed advice and reminded me of basics to reitierate.   As he said you should have at least 3-5 spots for every network.   In banner you probably want more than that.    Because of this you also need to make sure your mediation provider can handle 20-30+ lines in a waterfall.  

Key metrics to monitor include:

• Ad ARPDAU  (Ad rev/DAU)
  • This is what you care about.  Average CPMs and Fill rates are components that drive ARPDAU and not end goals themselves
• Impressions per DAU
  • Monitoring changes here (Assuming no product changes) will give you an idea of the technical effectiveness of the mediation and your networks.  If you ad a network and impressions per DAU go down it indicates something is not working well in that mediation black box and is probably an issue with that newtorks tech.   If you add a bunch of new lines and you see impressions going down you probably have found where the mediation performance drops.
  • If you are ever testing mediation solutions against each other this is an indicator of their techincal efficiency.   If you set up the same basic waterall in two different mediator..does one deliver a different impressions per DAU?