Api: Analytics/Reporting

[kristopolous]

From an email by Greg Tabak:

Draft of data points to capture, measure and evaluate - with the goal of creating a more efficient operating playbook for growth and expansion. The challenge historically has been tying inputs to outputs: e.g. employees at one location as an input, to the output of more cars being on the road/higher performance.

The below is list of potential data that we can use to tie inputs to outputs, also these would remain internal only:

time between bookings, per car
time car is registered below 30% charge, rotated to base to charged
time cars spend offline vs. available to rent
% fleet available to rent, per day
% fleet actively rented, per day
$$ earned per car, per fleet, per day
miles per car, per day
impressions per car, per day
reservations completed per day v. app logins
app downloads, and trends of usage 
abandoned reservations for reservations that are driven off
number of users in a car before a WaiveCar employee/contract touches the car (charges it, cleans it, rotates, etc)

Let's continue this discussion on what to track and what works. I think tracking more data is better than less, and we can utilize it in various ways - but more business intelligence (BI) is better than analysis paralysis, or making decisions based on "feelings" rather than data-driven business rules. We can use a lot of this BI to also back into a more robust ad-tech solution, that can be more comprehensive that ads on top.

[kristopolous]

Idle time / available time / charging time / rental time, this is #620

[kristopolous]

Really what I ought to do here is constantly log inverse data.

[kristopolous]

map out start/stop charge times

[kristopolous]

also see #627

[kristopolous]

/stats?date=XXX may be a good EP for this.

[kristopolous]

The "reservations completed per day v. app logins" is trying to get at how many people look for cars versus how many people actually get them.

[kristopolous]

The "app downloads, and trends of usage " could be communicated as registered users.

[kristopolous]

The "abandoned reservations for reservations that are driven off" is a reporting of total number of reservations versus total number of started bookings.

[kristopolous]

The "number of users in a car before a WaiveCar employee/contract touches the car (charges it, cleans it, rotates, etc)" can be detected by where the bookings end

[kristopolous]

A mockup for the web interface.

[kristopolous]

Also the report is a CSV that has a per-car breakdown table followed by a stats table.

[kristopolous]

I'll probably be coming back here in the future so for those purposes this is what I'm looking for: https://api.waivecar.com/report/2018-03/{normal,points}