Sunday, August 5, 2012

Startup opportunity to disrupt the Singapore taxi industry


The Singapore taxi industry is good by most standards, but is inefficient. This much is clear from the public gripe and news coverage on the woes of taxi passengers. Two quick stats to drive this point through. The number of taxis per population is high compared to benchmarks, and increased growth in taxis on the roads in the past 5 years hasn't done anything to increase total passenger trips. In fact, taxis in Singapore served more trips in 2006 than 2011!

Clearly, what's needed isn't more taxis, but higher efficiency. While there are 7 competitive licensed taxi operators in Singapore, there's been very little impact beyond keeping prices relatively low and introducing some marginal innovations with new booking channels (SMS, apps). There's very little incentive for taxi companies to boost utilization since they all generate revenue through leasing out taxis to individual taxi drivers. They maximize revenues by making sure as many taxis are rented out daily, not by ensuring taxis are fetching as many passengers as possible. 

Taxi industry inefficiency in Singapore


With such inefficiency, slack incumbents and smartphone penetration at 90%, I believe Singapore is a unique market where a startup could disrupt this space and improve the industry's efficiency.  The general idea would be something like a Lyft or SideCar: a platform to link drivers with passengers in real-time, except for Singapore this would be initially focused on actual taxi-drivers instead of regular car-owners.

Here are some features of this hypothetical "Singapore Sidecar" system and how it could make a difference.

Rating and choosing driver quality 
Currently the LTA tries to enforce service quality through QoS ratings and satisfaction surveys across the taxi companies. This really is not particularly useful. Ratings need to be determined for drivers, not taxi  companies.

The "Singapore Sidecar" application would allow passengers to rate the specific booked drivers at the end of each trip. Over time, passengers can then filter for drivers based on historical ratings and feedback.
This also means an opportunity for drivers to rate passengers, allowing them to turn down poorly rated passengers too.

Dynamic pricing
Taxi drivers optimize their driving schedules around most profitable times and regions of Singapore. This leaves 'black-holes', areas in Singapore at which point getting a taxi is like trying to find a unicorn.

A "Singapore Sidecar" app could allow for passenger determined booking fee and provide suggested booking fees for the given time and location. So if the app figures you're in a 'black-hole' zone, it could suggest a higher booking fee to incentivize taxi drivers to serve the area.

Truly centralized dispatch system across taxi companies
Each taxi operator runs their own booking and dispatch systems. This results in inefficient overall use of taxis. Instead of dispatching the best available taxi across the 27,000 taxis across Singapore, at best you get one optimized from the pool of 15,000+  from Comfort DelGro, the largest operator.

A "Singapore SideCar" system could integrate across a larger pool of taxi drivers over time, and should then be able to optimize much better. In practice, this means that instead of sending a Comfort cab 10 mins away, the system could send an SMRT one that's 3 mins away, and assign the Comfort cab to a subsequent, nearer passenger.


Size of the prize : SGD$20Mn annually
ComfortDelGro reported 2.4Mn bookings a month in 2011. Extending that to the full industry, and assuming a notional SGD$0.50 per booking puts this opportunity at around SGD$20Mn annually. So, the question is, who's best placed to try capture this opportunity? LTA setting it up? a couple of undergrads from NUS? Rocket Internet?