Scribd Logo
Upload

Welcome to Scribd!

  • 3K views

    Innovation at Uber: The Launch of Express POOL: Business Intelligence

    Uploaded by

    Rhishabh
    This document summarizes an individual assignment on Uber's launch of Express Pool. It discusses how Uber tested the innovation through experiments and analyzed data from a switchback experiment in Boston to determine the optimal wait time for customers. The analysis showed more cancellations and lower driver payout with a 5-minute wait versus a 2-minute wait. The recommendation is for Uber to use a dynamic wait time model and initially prefer the 2-minute window to prioritize customer experience and market share over profits as the service is still growing.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Flag for inappropriate content

    You might also like

    Innovation at Uber: The Launch of Express POOL: Business Intelligence

    Uploaded by

    Rhishabh
    3K views7 pages
    This document summarizes an individual assignment on Uber's launch of Express Pool. It discusses how Uber tested the innovation through experiments and analyzed data from a switchback experiment in Boston to determine the optimal wait time for customers. The analysis showed more cancellations and lower driver payout with a 5-minute wait versus a 2-minute wait. The recommendation is for Uber to use a dynamic wait time model and initially prefer the 2-minute window to prioritize customer experience and market share over profits as the service is still growing.

    Original Description:

    Original Title

    191229_1

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Read this document in other languages

    Did you find this document useful?

    Is this content inappropriate?

    Report this Document
    This document summarizes an individual assignment on Uber's launch of Express Pool. It discusses how Uber tested the innovation through experiments and analyzed data from a switchback experiment in Boston to determine the optimal wait time for customers. The analysis showed more cancellations and lower driver payout with a 5-minute wait versus a 2-minute wait. The recommendation is for Uber to use a dynamic wait time model and initially prefer the 2-minute window to prioritize customer experience and market share over profits as the service is still growing.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Flag for inappropriate content
    Download as pdf or txt
    3K views7 pages

    Innovation at Uber: The Launch of Express POOL: Business Intelligence

    Uploaded by

    Rhishabh
    This document summarizes an individual assignment on Uber's launch of Express Pool. It discusses how Uber tested the innovation through experiments and analyzed data from a switchback experiment in Boston to determine the optimal wait time for customers. The analysis showed more cancellations and lower driver payout with a 5-minute wait versus a 2-minute wait. The recommendation is for Uber to use a dynamic wait time model and initially prefer the 2-minute window to prioritize customer experience and market share over profits as the service is still growing.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Flag for inappropriate content
    Download as pdf or txt
    of 7

    Individual Assignment

    BUSINESS INTELLIGENCE
    Innovation at Uber: The Launch of Express POOL

    Submitted By

    K Parikshith Nair
    191229
    MBA FT 19-21

    Submitted To

    Prof. Balakrishnan Unny

    Date of Submission

    21 September 2020
    Summary and Key Points:

    ● T​he case deals with how the decisions on the innovation of service-the express pool-
    is done and tested along with the challenges they faced while implementing it.
    ● Express Pool is basically Uber pool + walking to a specific point and drop at a
    specific point (as expressed in their blog it can be considered as a on demand bus
    service)
    ● Some of the main competitors would include Lyft , Ola, Taxify etc...
    ● The range of services it offers are UberPool, UberX ,UberXL, UberSelect ,
    UberBLACK, UberSUV and UberLUX.
    ● Under the CEO three broad categories can be understood in organizational structure
    the Tech segment, Operations and Human resources.
    ● Tech division is divided into three verticals namely drivers ( recruitment training etc
    of the drivers), the Riders ( customers service etc) and marketplace ( substitution
    patterns, back end support etc).
    ● In many cases there is a need for cross functional integration and does play a role in
    the decisions made.
    ● Uber performs many experiments like user level a/b, switch back tests and synthetic
    control experiments to aid in its innovation to attain competitive advantage.
    ● One of the strongest strengths of Uber is its strong data analytics investment and
    analytical team.
    ● 2014 Uber launched Uber pool. ( algorithm used : Greedy algorithm) - customers
    going in the same direction in which a passenger is already going so that the ride can
    be shared among them increasing occupied seats - reducing cost for customers and
    increasing efficiency and profit for Uber.
    ● Extra time taken to reach the location of a customer and delaying rides was an issue
    with this version. To overcome this they came up with an idea of letting customers
    walk a bit so they can be picked up on the route of the ride. (express pool project)
    ● In 2017 they started working on it and in 2018 they started a 5 week test in 2 minute
    wait time for a synthetic control test across 6 cities in a controlled manner. The results
    varied from region to region.
    ● 2018 switchback experiments in boston started.
    ● they wanted to decide what wait time is better 2 min or 5 min . Each option has its
    trade off as explained in analysis before based on Boston switchback experiment data.
    Problem Statement:

    The crucial decision here is to decide on what the wait time should be - 2 minutes or 5
    minutes , the latter option being the one that has better optimised algorithms and better
    profits, however the tradeoff being that there might be more number of cancellations.
    Problem is to decide the optimal wait time as per time of day for the customers such that
    profit can be maximised by analysing their selection priority of uberpool or expresspool or
    cancelling from the switch back experiment at boston.

    Analysis:

    Data has different variables and is classified into two clusters - 2minute cluster and 5 minute
    cluster which is again classified into rush hour and non rush hour. By analysing the data we
    see that on an average the variables involved are as follows:

    trips_express rider_cancell total_driver_ total_double


    wait time trips_pool _pool ations payout total_matches _matches

    2 1377.177419 2653.790323 168.016129 29763.67826 2675.677419 1307.919355

    5 1460.698413 2419.380952 190.1269841 27183.20239 2438.301587 1357.730159

    ● We notice that the express pool still has a considerable demand in both cases.
    ● We notice that more cancelations are there as a part of the 5 minute cluster however
    the payout is considerably low.
    ● Increase in double matches is nearly the same as the number of rides lost.
    ● Further we notice the following trends:
    In non rush time we see that there is very little difference between the preference of the
    customers.
    In the 2 min cluster we see that there is more difference when compared to the 5 minute wait
    time in the customer preference.
    Double matches are lesser in number and pay out increases when the number of rides
    increases however the trend and number of customers are irregular so a dynamic model to
    predict and implement the wait time would be better. Also notice more matches, more
    customers are lost and when they are lost to competitors it is a bad thing.

    Suggestions and Recommendations / Solution:

    Uber should come up with a dynamic model of wait time which would also incorporate the
    history of wait time of the customers as a factor to see that algorithm is optimised at the same
    time the customer is not lost due to the same. As of now, as the business is still in the ​growth
    phase of this particular product based service, we need to see that gathering market share and
    making customers more comfortable in using to be a bigger priority focusing on profits and
    not making them used to this extra walking would result in loss of market share and
    unnecessary increase in competition. So the 2 minute window should be preferred.

    However in cases where the number of drivers available are less the 5 minute window would
    be better as the customer knows due to the UI of Uber that there are less drivers so would be
    mentally prepared to wait more.

    From Uber’s blog we know it is a process with a lot of potential and at specific geographic
    locations it has more impact like eg: airports .

    You might also like