Decoding: The Maths behind Uber [And On-Demand Businesses]April 10, 2015 2015-04-10 15:06
Decoding: The Maths behind Uber [And On-Demand Businesses]
Decoding: The Maths behind Uber [And On-Demand Businesses]
Logistics = Efficiency + Reliability
Uber has built a $40 billion company based on these core principles. At the heart of this valuation lies the success Uber has showcased in making a previously inaccessible service cheaper and more accessible. Not surprisingly the disruptive innovation has gone on to impact many other industry value chains. The term ‘Uber for X’ is now being used to describe the model of increasing number of startups looking to deliver/service practically everything on demand.
But how do you convert efficiency and reliability into quantifiable insights that can help you take business decisions for your On-Demand startup?
Lessons drawn come from unlikeliest of places – heading Jugnoo, an On Demand Logistics startup out of India. The experience over the past 6 months competing with Uber and Ola (Uber’s Indian clone) and creating a niche for Jugnoo resulted in insights that I believe are at the core of Uber’s success.
A brief background to give context to further analysis, Jugnoo is building a hyperlocal marketplace to provide everything on demand using auto rickshaws (most common means to go from Point A to Point B in India) for providing both rides and deliveries. Even Uber has been unable to achieve this elusive dream but Jugnoo is built around the notion of multiple use case from ground up. Having started 6 months ago, we are already doing 1500 transactions a day including auto rides, meal and grocery deliveries.
Reliability is the ability of the platform to service any demand while ensuring ETA for the request is below a certain acceptable threshold, x. For the purpose of this discussion we will use ETA as a quantifiable metric to represent reliability.
Efficiency, Q equals the number of minutes per working hour the average logistics unit stays engaged servicing the demand. As a platform you need to cross a certain threshold to become profitable.
x depends on a lot of factors – geographic, vertical that you are operating in, cultural, competition, etc. Assuming that you know Q and x at every point during your growth we can proceed to draw key insights based on these metrics.
A. Critical Mass
Critical mass for a certain region is the amount of total supply and total demand where you are ensuring both reliability and efficiency, thus requirements of both the ETA and supply being engaged above a threshold are met.
Insights into total supply and total demand, required to attain liquidity, is a quantifiable way to make many business decisions.
|Here is the analysis I used to pull off these numbers –
Available Supply, M, required at any point that can service any incoming requests;
Engaged Supply or Total Demand, N, that is busy servicing particular requests at any point.
Total Supply = M + N
M can be back calculated by increasing the available supply to a point in different demand centers such that ETA requirements are met. At any point you can improve your supply matching and demand prediction capabilities to get better ETA’s for the same available supply.
N can then be calculated from this equation –
Sum of N/M+N>Q
Actually, it should be integral on both sides over the duration of the day but we are just simplifying it.
In this era of startups on steroids where startups backed by VC war chests can burn money to create artificial demand and supply, starting with the analysis of a Critical Mass to sustain might seem counterintuitive. But the concept of liquidity in the marketplace still remains the most important metric that any marketplace should strive to attain.
- Knowing these numbers helps you plan your organic or VC backed growth for a particular city.
- You can also base the calculations of investment required off these numbers.
- Reaching these numbers shields you from any established competitor/ startup entering the area with a similar concept.
[B] Dynamic Pricing
Uber’s surge pricing has been one of the most controversial aspects about Uber for a long time. Many authors have gone on to offer deeply passionate views but despite all the negative publicity surge pricing has remained a constant for Uber for the simple reason that it makes business sense and goes with the base requirement of ensuring efficiency and reliability (Reliability here being equated to the ability to get a ride no matter how much the pressure on demand and supply).
|Pricing of a service comes from matching the demand and supply curves. Higher the demand for a given supply, higher the price.
When applied to a logistics network analysis flows directly from our calculation above. There is a limit to the elasticity in total supply, while the demand fluctuates a lot on a daily and weekly basis.
If M decreases due to significant increase in N, you need to artificially take the prices up so that less number of customers start demanding the service. Thus you maximize revenues per transaction while keeping service quality high.
The difficulty to staff or enable the platform for the highest demand is a problem almost every industry value chain that deals with fluctuating demand has to deal with. While dynamic pricing is a logical answer in the On Demand logistics vertical the decision on its introduction should also weigh in the price sensitivity of your audience and the competition.
[C] Multiple Services
A logistics network by definition can be used to do a lot more than taking the users from Point A to Point B. Uber has been experimenting with things as diverse as bringing users ice creams and kittens. When I started conceptualizing and creating a logistics network based on Auto Rickshaws, I had set my eyes on creating an Uber for Everything in India.
Although many people have advised me that spreading yourself thin and concentrating on multiple verticals is not a wise strategy. But here’s the thought process –
|Our “product”is essentially a collection of a tech platform, logistics fleet and consumer data. Rides and deliveries are just different use cases following exactly same process. The unit economics makes sense for both use cases. There is no logic behind choosing one use case over another. These multiple use cases smoothen the demand.
Using Jugnoo Autos we were able to reach 5-10 additional rides every day. Getting additional demand during the off peak hours via the meal delivery and hyperlocal delivery service was an attempt at leveraging the underutilized inventory during the off peak hours.
[D] Choice of the logistics platform
When I started there was no one who was providing an Uber like seamless service for people who want to hail autos. Around 46 million auto rides happen every day in India and its safe to claim that India moves on autos and not taxis. What I didn’t know at that time was how this ties in with providing multiple services.
|The choice of auto rickshaws as a base unit to create a logistics network comes with opportunities and challenges. What’s important from the standpoint of this discussion, is that
An average ride in an auto rickshaw in India costs around $1.2. Compare this with around $4-5 average for taxis in India. This is the reason why Uber has not chosen to do multiple services as a coherent business.
On the other hand, an average meal and grocery transaction also ranges between $6-8. So everything on-demand is just a natural fit for Jugnoo.
This unit economics is thus a challenge as well as an opportunity. The analysis of this unit economics for any On-Demand business is important before hand as challenges from realizations later can sometimes make the complete business unviable. For instance Ola was generating around 160 orders a day for some of the restaurants when it started a month ago and according to reports these orders have reduced to 15 a day now just because they are using taxis to deliver these meals – difficult from both the delivery time and cost perspective.
[Author- Samar Singla is the cofounder of Jugnoo, which competes with Uber/Ola.]