Step 1: clarify

• What part of lyft are we going to focus on? Shared, economy, Transit, luxury?
• What is the overall goal of Lyft right now?
• Are we Lyft as it is today or when Lyft was starting out years ago?

Step 2: assumptions

• Lyft’s main goal is to grow usage and grow profitability. However, since the threat of Uber is high, growth of usage should be of higher concern at this moment in time.
• Lyft only operates in the united states and we are only taking into account US usage
• I’m going to assume that when you say “Dashboard” that you want me to build dashboard that tracks one primary metric and several secondary metrics.

Step 3: let’s first define what we mean by ‘growth’

• Growth can mean growing revenue or growing the number of rides taken. Since growing number of rides taken will lead to increased revenue, let’s establish Lyft’s goal as “increasing number of rides taken on Lyft’s platform”

Step 4:Framework of user actions, specific metrics, supporting metrics and pitfalls

To grow # of rides taken on Lyft’s platform we need to persuade:

• Users to open the app
  • Specific metric to track: average # of times app opened per user per week
    • Pitfall: this metric could be inflated because there could be accidental app opens or opening the app for other reasons like checking on your account or promotions. Notifications would also throw off this metric.
• Users to enter a destination
  • Specific metric to track: percent of app opens that result in at least 1 destination search by user on a per user basis
    • Supporting metric to track: average # of cars visible within screen’s display per user
• Users to book a ride
  • Specific metric to track: percent of app opens that results in a booked ride per user
    • Pitfall: this metric can be misleading because a user could be booking rides if there are no other options available but not booking rides when there are other options available. Lyft needs to ensure that book rate is high even when options are available and one way to find out is by segmenting this metric by time of day.

Lyft at its core is a marketplace with drivers and cars on the supply side, and passengers on the demand side. I would begin by dividing the goal of the dashboard into two categories :

• Health of the marketplace
• Health of the app/ reliability from a technical perspective

The second one is more straightforward so I’ll briefly mention some overarching metrics I would look at and come back to it later if time allows.

Health or reliability of the app can be further divided into the following

• Crash rates
• ANR (app not responding)
• Uninstalls
• Network error rates
• Battery and network usage by the apps
• App ratings on the app or play store due to instability

Coming back to monitoring the health of the ecosystem, we need to drill down further into the two sides of the marketplace.

As a driver, the goal is to maximize on time by making the most money while I am on the app. As the platform that means I can look at the following :

• The amount of time a driver indicates available for Lyft and the time driver spends on a Lyft ride. The delta is what we will track.
• How much money on average per hour or per ride did the driver make. This would directly correlate with how much money Lyft as a business makes.

As a passenger, the goal is to get from point A to point B in the most efficient way. Efficient travel can mean two things

• I spend the least amount of money possible
• I spend the least amount of time between point A and B

A big part of this marketplace also relies on two-way feedback between drivers and passengers. I wouldn’t want that on a primary dashboard however, unless there is an unusual amount of negative reviews.

From a metrics perspective for passengers, we can track

• Time between request and arrival of the car
• Efficiency of route – delta between what a third party API, for eg. Google maps shows vs how long it took
• Number of lyft users, repeat users i.e. D7, D30 users
• Average driver ratings

Distilling to the top 7-8 metrics one can realistically look at on a dashboard :

• Average price of a trip per mile or per hour
• Fill rate : Number of rides taken/Number of rides requested
• Ratio of number of rides to number of available drivers
• Ratio of driver’s engaged time vs time available
• Retention : repeat riders within 7, 15 and 30 day time periods
• Average wait times for passengers and average unexpected delays en route
• With the recent addition of public transit info in the app : # of searches for a Lyft ride vs lookup of public transit option
• App reliability : Uninstall rates, network error rates and crash rates