# Poseidon
*"May the god of the seas help us to get benifit out of our data lake!"*

## Goals
* Identify **Power Users** (below called **PU**) by the frequency of identical uuid parameters
* Identify **sell out rate** (=>**SOR**) (not-fallback-impression-rate)
* Get a picture about **monetization**
* Force a defined uuid (by browser plugin) and track the journey for it
* Get a database to make decisions about ad density

In order to reach these goals we need to get data (impressions, adunits, lineitems), sanitize and store them in a database.

## Reporting Page with Data Views
In phase1 this will be only for *one site (Kabel1)* because data size and performance is a challenge.

### Segments of power users
* Pie chart with segments of PUs <br>(eg. 0, 1-9, 10-49, 49-100, 100-499, >=500 impressions by uuid)<br>The segments can by dynamic in the future
* Show percentage per segment
* Filter by date
* Filter per distribution channel (CTV - inApp- Online)

### Segments of device categories
* Pie chart with device categories
* Show percentage per category
* Filter by date

### Sell out rate
* Table with sell out rate
* One line per PU segment with columns:<br>fallback - programmatic (list different SSPs) - IO
* Overall

## User journey
* number of impressions - money - sell-out-rate ( - viewability in phase2)
* per sizes
* overall

## Plugin to provoke custom uuid
* create one uuid to be used
* change uuid
* de/activate uuid
* see uuid
* functionality
* replace add this uuid to every network request to https://securepubads.....
* Link to the report site, passing the current uuid -> go to "user journey"
* Question: How long will manipulation of request urls be possible via Plugin API??

## Endpoint for targeting module in adtec-core
* Retrieve data for current uuid
* Details TBD

### Loss rate => phase2
* Table of loss rate
* Per PU segment
* Overall

### Most popular referrers => phase 3
* List of 20 most frequent referrers with their
* Sell out rate
* Loss rate

## Glossary

* `Loss Rate`: NetworkImpressions / NetworkRequests
* `Sell out Rate`: not-fallback impressions / impressions

## Fragen Horst
* brauchen wir die creative sizes?
* welche targetings möchtest du nachher betrachten?
* wie erkenne ich ein fallback in einem impression data set?

## Issues
* Infos about revenue would be in the impression's `Revenue` field. To get the field we would have to enable this in GAM > Admin > Data transfer.
* Would that have side effects on other services using these export files?

* On some display impressions `chreativesizedelivered` has the value `Intersticial` instead of eg. `300x250`