There's a 50% drop in the number of Uber rides per day. Diagnose the problem.

Question

There's a 50% drop in the number of Uber rides per day. Diagnose the problem.

Asked at Google over a year ago

5.8k views

Problem Solving category

Asked at

5.8k views

How to answer Problem Solving questions

Interview Guide

Top Problem Solving interview questions

Invite members
Invite by email

Add another

beetle62 · Answer 1 · 2021-01-03T00:01:20+0000

Candidate ("C"): The first thing I would try to do is to dig further to understand the problem and its root cause. Has the drop in rides been a gradual drop or has it been sudden?

Interviewer ("I"): The drop has happened suddenly over the last 2 days.

C: Is the drop specific to any geographic location or global across all users?

I: The drop is limited to New York City.

C: This is very interesting. I would like to probe a little further on this and ask if the drop in rides is across all ride types or is it specific to certain ride types such as Uber X, Uber Pool, or Uber Select?

I: The drop is specific to Uber X rides.

C: Okay, to re-iterate the problem, we have seen a 50% drop in UberX rides over the last 2 days in the city of New York. I would like to spend some time first exploring external factors that may be causing the drop in riders. Have there been any events in the city such as a marathon, a Covid-19 lock down, or anything else in the city that might be impacting transportation?

I: Nothing out of the ordinary is happening across the city.

C: And what about our competitors, do we know if they have pushed out any promotions or if they are doing anything different in the last few days in New York?

I: Yes, our competitor is running a promotion.

C: Do we know what the promotion is?

I: They are offering $7 off of your next ride across the city this wee

C: Okay, so the root cause of the 50% drop in Uber X rides in New York for the last 2 days is a promotion being run by our competitor. Now that we know the problem, I would explore next steps.

I assume that our mission at Uber is currently growth and given ride sharing is fairly commoditized, I would recommend exploring a retaliatory tactic by offering a more aggressive discount to convert our users and drivers back to Uber in New York.

I would also review the impact of our competitor's promotion on our users in the city as they may have similarly lured over some of our drivers, leaving our users to face longer wait times and more expensive rides due to a lack or drivers. If this is the case, I would explore increasing the driver surge pricing delta to lure drivers back faster to high demand areas to ensure that our users needs are being met (particularly if we expect an increase in demand with our own promotion).

hjc · Answer 2 · 2020-08-26T05:06:35+0000

My approach to this Google problem solving interview question:

Clarifications:

Is this the number of successful uber rides? or ride requests? [Successful rides]
Is it a sudden drop? or a gradual one?[Sudden]
Can I know how long the metrics has been at this low level? [It just dropped yesterday]
Does this drop happend in one city, one country or involving multiple countries? [One city]

Analysis:

Let's take a look at the user journey of a successful uber rides:

1. Passenger opens Uber App

2. Passenger input the destination address

3. Uber allocate the request to a nearby Uber driver

4. Uber driver accepts the ride

5. Uber driver pick up the passenger (if the driver couldn't meet the passenger, it will cancel the ride)

6. Uber driver arrives at the destination and mark the ride as completed.

7. Some passengers might leave a review and /or tip the driver.

First, we need to rule out possibilities of seasonal change. Since Uber has been around for quite a while, and based on the date, history data and comparison with adjascent cities, it's fairly easy to figure out. [Let's assume this is not seasonal]

Then I will move on to determine whether the drop is caused by internal factors or external factors.

Internal factors might include:

Metrics Error
Feature release causing interruption
Bug, such as application crush or billing failure causing passenger or drivers not able to use the app.
Driver Incentive Program or Passenger Incentive Program stoped in this city.

External factors might include:

Various reasons causing interruption to the transportation, such as weather, public event, road maintenance, etc.
Competitor, for example Lyft launched its service in the area or started promotion.
Bad PR, such as Uninstall Uber
Google Map outage
Regulation or court order
Cyber Attacks
Android or IOS bugs or even mobile carrier outage

Most Internal/External factors we probably already get headsups from PR, Legal, Support teams. We can also proactively narrow down the root cause by check the metrics at different stages of the user journey, and conversion rate between stages.

For example, # of unique passengers for this day, # of active drivers, # of ride requests, # of passenger cancellations, # of driver cancellations, # of billing failures, # of trips that driver drove to close to destination. It's always useful to split the metrics by location, sw version, carrier, OS, phone brand, model, We can also reference data from support call center, billing partner data, App crush report, etc.

If the metrics of daily successful rides (DSR) is wrong due to a bug, we probably will see data from billing partners, such as revenue and # of charges, didn't match the metrics at our end.

If the root cause is a feature push causing system not able to allocate task properly to drivers, we will probably see the # of Unique passengers opening the app didn't decrease, the # of ride request didn't decrease, or even increase, # of active driver didn't decrease, but # of successfully matched rides decrease.

By looking at those data and pin pointing the root cause, it's much more straight forward to propose and implement tailored solution, such as rolling back the feature sw, push fix in a new client version, or adjusting the driver/passenger incentive program or pricing. I wouldn't go to full length here unless you would like me to. Thanks!

Kamal Mihani · Answer 3 · 2020-07-10T22:07:18+0000

I would approach the diagnosis with a 3 step approach to isolate the root cause,

First, I would try to segment the user properties to see if a particlaur type of user is leading to the drop.

Second, I would segment the product prpoperties to see if Uber platform is leading to the drop.

Third, I would see if the users are switching to another platform to book their rides due to a competitor like Lyft dropping their prices significantly.

To segment user properties, I would look at the following things:

Geography
Device
App Installs - Success/Failure
Traffic Source
Admin tools Uber uses to collect this data

To segment Product Properties, I would look at the following things

Session Length
Session Frequency
# users successfully login
Login issues
% successful completion of user flow on confirm ride page
# viistors on the confirm page/ # total logins
Recent change in UI for "Confirm Ride" Page
Error codes coming from the apps/database

Finally, I would look at recent news articles, talk to the marketing and customer success teams, and perform a high level market analysis to see if user are shifting from Uber to Lyft because of things that Lyft might have launched.