Estimate the number of uber rides?

estimation estimation market sizing Market sizing
New answer on Sep 07, 2020
9 Answers
3.1 k Views
PK asked on Apr 10, 2020
Director of Product at PayPal. Based in SF. Worked at Twitter, Yelp, Salesforce, Houzz in the past.

Assuming:

  • Talking about uber car-based rides and not scooters, public transit, or helicopter.

  • This is # of rides per day.

  • US only.

Approach:

Estimate the # of uber riders x their average number of rides per day. --- (a)

Splitting (a) by power users, normal users dormant users.

Estimate the volume of each of the above user types and count their rides/day separately.

US population = 300M

Target Uber audience in the range 15-65, with average life expectance of 80 years.

Thus, the target user count 50/80x300M = 190M

Assuming 70% of these users actually use Uber → 130M users

Based on my experience seeing users, 20% of users, in general, tend to be power ridesharing user, they use it atleast 1 time a week.

60% are normal users - Use it atleast 1 time a month

20% are dormant users use it once every quarter,

  1. 26M power → Daily rides → 26M/7 rides per day → 3.7M rides/day

  2. 26M casual users → 26M/90 rides per day → 0.2M rides/day

  3. 78M normal users → 78M/30 rides per day → 2.6 M rides/day

Thus total rides per day → 3.7M + 0.2M + 2.6M ~6.5M rides per day

Any feedback on my approach and assumptions?

(edited)

Overview of answers

Upvotes
  • Upvotes
  • Date ascending
  • Date descending
Best answer
Emily
Expert
Content Creator
replied on Apr 10, 2020
BCG Project Leader | 3+ years interview experience for BCG SEA recruiting | Kellogg MBA, NTU, Peking University

Hi PK,

Overall a good and logical approach!

Just some small points on assumptions.

(1) 70% penetration of ride-hailing seems high. While it could be possible in the big metropolitan cities like NY or SF, outside the handful of big cities the penetration rate would be rather low.

(2) Don't forget there is Lyft also. So besides the 70% penetration rate, you probably need to think about market share as well. At the moment Lyft probably is comparable or just a bit smaller than Uber?

Cheers,

Emily

Was this answer helpful?
Thomas
CoachingPlus Expert
Content Creator
replied on Apr 11, 2020
150+ interviews | 5+ years experience | Kearney & Accenture | Sold consulting startup| London Business School

Hi PK,

Generally a very solid approach. What would give you plus points is to sanity check the numbers throughout. Complementary to what others have said the numbers are high.

1) The general consensus is that this 70% is way too high and this might a bias that you want to be aware of (given your tech background). What would have scored you extra points are to think about alternative modes of transportation and that would have prompted you that 70% is likely to be high. Having said that, your segmentation somewhat corrects for this.

2) Second, clear reasoning on the 70% is missing as well. For these high impact numbers, it is smart to mention why you made that assumption (even if it is based on your own experience).

3) Important with market sizing questions is to do a quick sensitivity check at the end. E.g. acknowledging that your analysis is based on a few sensitive decisions/data points.

Was this answer helpful?
Axel
Expert
Content Creator
replied on Apr 10, 2020
Bain Consultant | Interviewer for 3 years at Bain |Passionate about coaching |I will make you a case interview Rockstar

Hi PK,

Overall I think your approach makes sense. I have a few ideas which I think could improve the market sizing even further:

1. Remove U.S. Population that does not live in urban areas or outside the 300 or so cities in which Uber has a presence. Around 50% of people in developed countries live in urban areas so that could be an easy assumption to use.

2. Reduce overall % of people that are Uber as 70% seems a little bit high to me.

3. Account for Uber's market share in the market. You have other competitors such as Lyft to account for.

Another way to think of this market sizing problem would be to start with the US population and try to estimate how many trips of various kinds that people would need to undertake per day, then what share Uber rides what take from the total.

Was this answer helpful?
Sanjeev
Expert
replied on Apr 10, 2020
PwC/Strategy&/GT/Chicago Booth - 2nd Session Complementary till June 1st

Hi PK,

A reasonably good and detailed approach....I like your segmentation. Few points that could make it stronger:

1. Market share - The addressable market for ride sharing - 70% assumption seems high considering so may alternatives (own car, public transport in arge cities, etc). In addition, you have to add a "market share" factor for Uber which would reduce the potential Uber rider population

2. Reality check - Once you come up with market estimation, always do a gut check. Your 6.5M rides per day seems v. large (2% of US population or 1 in 50 people will take Uber). A lot of people forget to do this - a good way to differentiate yourself (also a desirable characteristics for a Consultant).

Good luck!

Sanjeev

Was this answer helpful?
Anonymous replied on Apr 10, 2020

Hello PK,

I can see at least two different approachs for this estimation.

You can either think about (i) the # of users, like you did, or (ii) the # of drivers. To choose one, think about which one has a lower level of uncertainty on your assumptions.For instance, why do you believe 70% of target users actually use Uber? How can you anchor and justify this number?

Personally, I would choose the second options because I believe it is easier to justify the assumptions needed:

* # drivers in the US

* Drivers segmentation: full-time vs part-time

* Average working hours for both full-time and part-time

* Average rides per hour: average ride (min) + average waiting time (min)

Please note that we can add as many layers of complexity as we wish. For example, we could have split working hours into peak and non-peak.

Try this approach to do the estimation and, then, we can discuss your assumptions as well.

Also, use it to double check your first answer!

Cheers,

Was this answer helpful?
4
RUBEN on Apr 20, 2020
Hi Raphael, how would you make a solid assumption for the # of drivers in the US? I know is a common practice for some consultants but sometimes it seems like getting one's finger in the air. I assume we do not do another market sizing to get the number, we just make an assumption. Ruben
Anonymous on Apr 21, 2020
Ruben, you are right. Usually, a candidate will not have time to do this "extra" estimation. So, you will have to make an assumption and here is important to show your business sense. Nobody has to know the exact number of Uber drivers in the US, but you are supposed to know the order of magnitude of this number. Are we talking about thousands or millions of drivers? Here, we cannot commit this kind of mistake.
Anonymous on Apr 21, 2020
If a candidates says "Let's assume there are 1 MM Uber drivers in the US", the interviewer could say "Humm, I think this number is a little bit low. Let's use 2 MM instead" and then you continue the case.
Clara
Expert
Content Creator
replied on Apr 10, 2020
McKinsey | Awarded professor at Master in Management @ IE | MBA at MIT |+180 students coached | Integrated FIT Guide aut

Again, good logic but too high estimations

Was this answer helpful?
Dhiraj replied on Sep 07, 2020

Question - Estimate the number of Uber rides?

Clarifying questions and Assumptions -

Considering only the Delhi NCR region, India

Calculating # no rides per day

Will consider all cab rides (Uber go, Sedan, premier, auto, bike, share, etc) but not food delivery rides

Approach: To come to a reasonable number I will user below equation

# number of Uber rides in Delhi NCR per day = #of uber users * avg. no of rides

Now I will calculate the number of uber users in the Delhi NCR region

The total population of Delhi NCR = 20M

Internet and smartphone users = ~50% so the remaining number is 10M

Assuming the target audience which might use online cab services is 60% which makes it 6M.

Assuming Uber and OLA being 2 major players in the NCR region with equal market share so uber’s market share is approximately 50% which implies 3M users of UBER in Delhi NCR

Now to calculate avg no of rides per day I will break the Uber users in the following segments -

Power users = 10% of 3M = 0.3M and takes 10 rides per week making ~0.42 M rides per day

Daily users = 20% of 3M = 0.6M and takes 5 rides per week making ~0.42 M rides per day

Frequent users = 30% of 3M = 0.9M and takes 2 rides per week making ~0.25 M rides per day

Casual users = 20% of 3M = 0.6M and takes 1 ride per month making ~0.02 M rides per day

Rare users = 20% of 3M = 0.6M and takes 1 ride per quarter making ~0.007 M rides per day

Total uber rides per day = Total no of rides by all different types of users

Total uber rides in Delhi NCR per day is around 1.1 M rides

Was this answer helpful?
0
Anton
Expert
replied on Jun 29, 2020
FREE 1st session in November | From Lawyer to MBB | Top in FIT | 10x your structuring skills | Message to get Free Prep Checklist

Hi,

Your approach is indeed a good one.

If you want to master your market sizing skills I suggest you focus on segmentations patterns. You can use the following segmentation for market sizes:

B2C:
-Demographics (Age, education, income, family size, race, gender, occupation, nationality)
-Behavioral (Purchasing behavior, customer journey stage, occasion & timing,
customer loyalty & interest, risk tolerance, user status)
-Psychographic (Lifestyle, personality traits, values, opinions, interests of consumers)
-Geographic (Geographical boundaries)

B2B:
-Company characteristics (Industry, company size, number of employees)
-Geography (Geographical boundaries)
-Purchasing Approach (Occasion & timing, customer capabilities, nature of existing relationship)
-Personal Characteristics (Loyalty, risk attitude, user status)

B2G:
-Demographics (Type of agency, size of budget, the amount of autonomy)
-Geographic (Geographical boundaries)
-Government Tier (Federal , State, Local, Quasi-governmental, International)
-Bid type (Closed, Open)

But sometimes you don’t need to segmentation. Here is an example of case that could be solved with high level top down approach - estimate the size of credit card market in the US:
https://www.preplounge.com/en/consulting-forum/how-should-i-approach-the-following-question-estimate-the-market-size-of-credit-cards-in-the-us-6695

Was this answer helpful?
Antonello
Expert
Content Creator
updated an answer on Apr 30, 2020
McKinsey | MBA professor for consulting interviews

Hi, I think it is a good approach. Always try to stay in 20 minutes and - if needed - accelerate some estimations by asking to the interviewer

Best,
Antonello

(edited)

Was this answer helpful?
Emily gave the best answer

Emily

Content Creator
BCG Project Leader | 3+ years interview experience for BCG SEA recruiting | Kellogg MBA, NTU, Peking University
113
Meetings
2,156
Q&A Upvotes
14
Awards
37 Reviews