Discussion 2: Case Study on Innovation at Uber: The Launch of Express POOL

1. How does Uber innovate? What is the role of data science in innovation at Uber?

The company implements their innovation strategies by putting the minimum viable products in
market and iterate accordingly based on the outcome or experience received. It’s focused to
execute the rapid product iteration and capable of retrieving the statistically significant findings
about the effects of a product tweak within a short amount of time and immediately change the
product accordingly. It is quite daring whereby they enter a new city and lands a standard
package in the market. They let the riders to get used to the company’s core products then they
proceed to iterate followed by altering their findings depending on the market’s reaction be it
acceptance or rejection. The company has also invested a lot of in the efforts of developing and
iterating its mobile app to make it seamless and responsive to market needs and demands. This
approach affirms its steady steadfast commitment to make its operations user friendly and
adaptive to different market dynamic surge and pricing strategies. This advantage has partly
spurred its cost leadership advantages. It has also provided it with scalability advantages and
eliminated barriers to growth.

A key element of Uber’s innovation strategy was its substantial investment in data science. The
data scientist at Uber plays a major role in the company’s operation and assisting the engineers to
develop the efficient algorithms in order to capture or identify the most suitable route, prospect
passengers waiting and the minimum travel time estimation. It is about extracting, analyzing,
visualizing, managing and storing data to create insights. These insights help the companies to
make powerful data-driven decisions. Data Science requires the usage of both unstructured and
structured data. With the help of the efficient data analysis, a more accurate and reliable heat
map can be created. With the rapid development of the mobile Internet, the virtual world and the
real world merge with each other. Based on mobile Internet technology, people, machines, and
information can be interconnected and integrated through computing, autonomous control, and
the Internet of Things. Mobile Internet has had a profound impact on the development of
multiple industries, and Uber is not an exception too.

2. Why does Uber have so many different ways to run experiments? What are the
pros and cons of each type?
Uber has so many ways to run it’s experiments to improve its products, depending on the type of
the improvements to be tested. The most common and mostly used by Uber were User A/B
experiments, Switchbacks an also the synthetic controls. The standard user level A/B
experiments compared the behavior of app users to test the effects of platform decisions.
Switchbacks were another type of study designed to evaluate the effects of a product tweak on
some outcome variable of interest. The issue with this is that it can only be ran one time in a city
to prevent them from interacting with each other. The Synthetic Control Experiments is used to
create treatment and control cities to study the effects of a product tweak on a set of outcome
variables of interest. In order to detect any changes, they need to be fairly insignificant but lots of
Uber’s experiments generate effects of smaller magnitudes. The setbacks faced by Uber is that
their team ran several experiments simultaneously and results in the risk of facing the
contamination and spillovers effects, whereby the result of on experiments were misinterpreted
as the results of another. There is also a challenge in figuring out how to launch new products
and test their effects without interfering with other peoples experiments. It is also found that its
very hard to measure the effects of their product tweaks due to lots of external factors to take into
account plus there are network effects to be taken into consideration as well whereby the teams
presented results that overstated effects.

3. Evaluate the Express development project. What did Uber do well in this project?
What could have been improved?

In the EXPRESS development project, there were much better focus and rethink the strategy of
shared rides it has in place and to improve further compared to what POOL has gained. One of
the best moves in the project is the decision to create the Task Force by combining the
marketplace team and shared ride team. The team adopted 2 key strategies in this project to
improve the share rides. One is by requesting the riders to wait up to 2 minutes while the
algorithm matched them to co-riders. And the second one would be for the riders to walk to/from
a pickup or drop points. These strategies were done in the anticipation of solving the
backtracking/ detouring issue as well as getting more suitable matches for the particular ride
intended. A new algorithm was developed catering to the above strategies and this later proves to
be more efficient compared to the greedy algorithm of POOL. Implementing dynamics and
flexibility in the pick-up points helped Uber to gain more seat utilization per mile as well as
providing the riders options to choose the best, nearest and most convenient pick-up points
which resulted in Win-Win situation. As and added value or improvement options, Uber could
also have focused in the van or mpv options when it comes to car pooling or sharing rides as it
could be more comfortable than a sedan vehicle. This would help to reduce or eliminate the
hesitation of sharing the space in the vehicle as there are ample comfort available and not
cramped. This who do not prefer sharing the normal sedan could also choose this option and the
Uber can retain their customer base and attract more. Provided the clients who choose this option
should be willing to adjust to the vehicle seat maximum utilization or reasonable accommodating
algorithm so that the cost can be covered accordingly by Uber. Cost effective based on more
passenger volume achieved.

4. Based on the data available to you, other insights from the dataset, and additional
qualitative considerations, what would you recommend that Stock do? Should he increase
wait times from two to five minutes in the six treatment cities of the experimental launch?
If so, when should he do so?

Based on the data and insights received, I would recommend Stock not to increase the waiting
time to fixed 5 minutes instead set a waiting range from 2-5 minutes waiting time. This should be
applied by taking consideration of the peak/non peak hours, the number of riders
availability/traffic in the vicinity. Maximum waiting time for a client should not exceeding 5
mins especially if it’s a 1st customer pick up. Because as for the App in place, the possible pickup
rote/passengers along the route can be mapped accordingly in a fast pace by the algorithm
developed by Uber and minimum waiting time targeted can be achieved efficiently. Another
option is that, as for the cities where the traffic of riders can be increased where it demands,
Stock could strategize a plan for partnership with the idle cab/taxi companies and utilize their
transports so that once the sufficient traffic is available based on the heat map then the waiting
time issues can be sorted.