How would you triage a 10% drop in rider cancellations?

Calrifying the problem statement

"10% drop in Rider Cancellations" might sound like a good thing. Before we assume that we should understand what "Drop" means here, lets assume - Rider cancellation rate % = (Ride requested + Cancelled) / (Ride Requested)

It's possinle thay there has been a decrease in overall ride request made because of whihc the overall rider cancelation rate has spiked. 

To investigate the decrease in Ride Request let's follow the following structure

1. External Factors - Have there been any competitor or gov / regulatory changes causing decrease in ride request

• Competitor has launched a new promo / discount for that week cauing riders to switch to competition 
• New competitors - new market entrants 
• Bad Media / PR cauing riders to choose other options
• OS level changes made that is cauding riders to nt be able to request ride
• Is this a market trend?
• Regulatory / Goverment changes preventing riders to request ride ---- Airport ride req ban etc.

2. Internal Factors

• Investigate the data logging instrumentation and confirm if the data is logged correctly for ride requests
• Have there been any product chnages made in that week - Feature or UI that is casuing riders to not be able to make a ride request
• iOs or Android users?

3. Rider experience - Funnel breakdown - Have there been any changes in the ride request flow? 

• Rider open the app
• Rider can see cars in their location in the map based UI
• Rider sets detination location 
  • Rider is not seeing ride options  ----- possible shortage of drivers demand x supply issue
  • Ride options are Priced high ----- possible shortage of drivers demand x supply 
  • Estimated ETA displayed is high ---- possible shortage of drivers demand x supply 
• Metrics to investigate 
  • % Ride Requested = (Ride searched + requested) / Ride Searched
  • % App open to ride search = (App Opened + rider searched) / App opened
  • Driver Acceptance Rate % = (Rider made request for ride but rider cancelled by driver before ride request was confimed ---- possible that drivers are cancelling ride request recieved more than usual in the window alloted for accepting and confirming ride request

Assumning that competiris have launched a promo that week casuing riders to choose competing and hence making less ride request

Promo and discount have been a strategy to do some of the following. Lets think through strategy / solution to tackle these

• Increase market share 
• Bring dormant users back 
• convert casual user to more active users

Can we work with out proudct and marketing team to build staregy to increase Revenue, Market Share and Rider Expereience.

• Revenue: Can we Increase ARPU? --- it will be importan to understand what is
  • Avg rev per trip per rider 
  • Avg trip per rider 
• Market Share
  • Can we build promotion and discount strategies to attact lyft users
  • Can be optimise ride pricing and make it more competitive
  • Can we optimize our demand and supply model
  • Can we increase driver market share - get new drivers
• Work with Marketing to build solution to increase Lyft Awareness, interest, trial or registration