Analytics - the single biggest competitive advantage companies have but not many leverage

(Notes from an interaction with Mr Raghunandan G Co-Founder and CEO of TaxiForSure.)

Raghunandan G (@raghugnandan) started off with why one should pay importance to analytics in any stage of the company form starting up to well established behemoths of the world. The main reason why it's so important is that while many other competitive advantages can be copied by competitors and hence can be short-lived, analytics will only give increasing compounding returns. More data you gather over time, the better insights you start getting for your business.

Next he shared examples and stories from his TaxiForSure journey. So they started around June 2011 and soon with some efforts were able to have 8 to 10 bookings per day. Another interesting insight was around how most important metrics in VC conversations were around repeat customers and referral customers back in those days.

It was at this point that they noticed something strange while analysing data. One of their customers booked 50 rides in a month, when they were overall averaging just about 8–10 bookings a day. When they asked the driver they found out that the customer was a pregnant working lady and it was taxi was used as her daily commute to and from office on weekdays and in aerobics/yoga classes on weekends.

This was their first "aha!" moment. They realised that no husband would mind the increased spends of a repeat cab bookings when safety of their pregnant wife is at stake. Soon they marketed at all these aerobics and yoga classes, the likely places where their pregnant customer visited. This resulted in an increase to about 40 bookings per day by month end. They soon secured funding by January next year and then began the next phase of their journey.

Given how they got their first growth spurt, they focused on a strong analytics culture in the company. The ownership of the analytics had to be at founder level, else the various divisions of the company might not agree at times on the way forward even when the data says so. He illustrated this with some examples:

When in initial phase of the company they had to decide if they need focus more on new users vs. existing users. People in the company were divided on where to focus more resources between these two segments. It is here that they had another one of their "aha!" insights. They discovered that users who have done more than 5 rides became loyal customers and were more much tolerant even if they missed a few calls, but they would still stay with them for their next booking (He also mentioned a caveat that there were no better fallback competitors like Ola or Uber in Bangalore at that time). So they decided to dedicate more resources to new users and help more of them cross this threshold of 5 bookings. This strategy resulted in high growth over the next few months.

It was somewhat similar to facebook's famous "get a user to reach 7 friends in 10 days" insight which led them to grow phenomenally with high engagement rates.

Next he described how during their initial phase driver's incentives were not exactly aligned to how they wanted them to behave and how a data driven strategy helped them deal with it. The problem they had was how drivers dealt with the two peaks in demand during the mornings and evenings. Most drivers would drive all through the morning peak hours, then during the lull hours go home to sleep and not come back for evening peak hours. This led to severe shortage during evening peak demand hours, and the main reason for it was the afternoon lull in demand which took many drivers away and their inertia to come back to work again.

They noticed most of their booking were done 1–2 hours prior to the time of pickup. So they next invested heavily to advertise on google in the morning hours so they have enough demand even post morning peak hours to have the drivers continue and stay incentivised through the day till evening. Next they also advertised heavily in the evening owing to the insight that there were a lot bookings done prior evenings for earlier in the next morning. So they flattened the demand curve throughout the day a little instead of sharp peaks. What it effectively did was that it increased the overall working time of drivers and helped them earn more, which automatically acted as an incentive enough to keep them motivated not go off. Thus they were successful in somewhat flattening both the supply side and demand side curves and their overall number of bookings doubled from about 600 per day to about 1200 per day within a month.

Another example he shared was how the data helped them optimise the experience for both the driver and users. He started with an example of how they get two requests from Domlur to Yeshwantpur and other to Whitefield, but had only say one cab available which one would they choose. The data said that it will be easier to get another ride back from Whitefield while driver will probably be stranded for another ride in Yeshwantpur. So they accepted the Whitefield booking as this resulted in 2 happy riders and 1 happy driver.

And it is here that he shared how earlier when most of the apps did not ask for a destination that studying the data of cancelled bookings they discovered how destinations like Yeshwantpur in above case were unfavourable destination for drivers based on time of day, day of week etc. And how even post bookings drivers would cancel when they discovered user were going to some of these unfavourable destinations. It is these insights which led them to ask for destinations while booking itself to optimise these experiences, which has become a regular feature of all taxi apps now.

Overall it was a good session knowing how they leveraged data to grow and optimise their business, and how it can be replicated for most companies irrespective of which stage of growth they are in.