Travel planning can be chaotic at a personal level, especially with the uncertainties of Covid thrown in. NASDAQ-listed Online Travel Aggregator (OTA) (MMT), which offers flight, train and bus ticketing, cab and hotel booking, and vacation planning services, has a team of nearly 100 people trying to find order in this chaos to customise the platform for its users.
Right from homepage and recommendation engine to suggesting add-on services, MMT’s data team has been working at improving its platform for the end user.
“Compared to other ecommerce services, travel requires context, it is a less frequent purchase and is a perishable inventory. Pricing is dynamic in this case and it is different from what a user might be looking for. That is why data science around travel is an interesting problem to solve,” Sanjay Mohan, Group CTO at MMT tells YourStory.
As a case in point, a solo traveller from Delhi to Bengaluru might have different needs in terms of cab and hotel booking as compared to someone traveling with four family members, depending on their age, amount of luggage, and other factors.
It is to improve personalisation and data-driven products for various business verticals that MMT set up its data team across Bengaluru and Gurugram nearly six years ago, though it had a handful of experts earlier.
Better data leads to better outcomes
The data team is divided into the data platform team and data sciences team, headed by Group CTO Sanjay. The data platform team is responsible for building the data pipeline -- right from how the data is logged to storage of quality data in accessible form.
“Data cleansing or data ranking is an industry-wide problem and takes up nearly 80 percent time of any data scientist. With the help of the data platform team, we have managed to bring that down and the time spent cleansing the data for the data sciences team is now negligible,” says Sanjay, adding that the initial two years were invested in setting up the data platform team.
The data sciences team has varying degrees of expertise in Machine Learning (ML), Artificial Intelligence (AI), and statistical and neural network models for data analysis. This team solves use-case specific business domain problems.
This has led the platform to come up with data-led features for various business domains across flight, hotel, and ground transportation.
“Even in my everyday business teams we need people with some understanding of data sciences so that they can take ownership of new models. So we train a lot of our engineering teams as well as product managers through certification to be data enabled and data aware,” says Sanjay.
Deep Kalra, Founder, MakeMyTrip
Product improvement and new features
With the help of this set up, MMT has been able to come up with new features and improved experience for customers.
Group company Goibibo was the first to announce the price-lock feature for flight tickets in December 2020, which was followed by the MMT platform. The feature allows a buyer to lock-in the price of a flight ticket for a period of up to seven days by paying a token amount. If the price is lower at the time of booking, the user can book it at a lower price, while in case of a higher price, they pay the lock-in price.
MMT also announced a partnership with Goldman Sachs-backed global travel booking app Hopper in October 2021 to use the app’s Price Freeze Technology to start a similar feature on MMT platform.
“The fare lock-in feature helps reduce customer anxiety around flight booking and reduces the friction of buying. These are still early days and we have opened up the feature only on select segments,” says Sanjay.
He added that MMT platform has data on ticket pricing from nearly seven years ago and can predict how much the price is likely to change, so that the customer is happy and the platform does not bear the brunt of a poor probabilistic model.
MMT also introduced its adtech feature during the Covid months, though it was in the pipeline for nearly a year and half.
“Lot of users come to our platform to check flight or train ticket schedules or cancellations, Covid related travel guidelines for states, and other information. They don’t transact on the platform and the ad-tech feature helps us monetise the traffic effectively,” says Sanjay.
While the adtech feature has been launched with a third-party technology provider, Sanjay adds that the analytics and dashboard for better targeting has been built by MMT’s data teams.
Other improvements on the platform include better customisation and recommendation for hotel bookings to ensure the customer is able to find the right kind of hotel in the top 15 search options.
“Nearly 80-85 percent of our users now find relevant hotel searches in a few clicks among the top 15 searches from 60-65 percent previously. This has definitely added to better conversions,” he adds.
The platform also introduced a travel fintech feature called TripMoney earlier this year to cross-sell and up-sell products, including BNPL solution, bite-sized insurance, and forex cards. The data team also works on developing risk models and personalisation for consumers availing of these features.
In the midst of a talent war in the Indian technology sector, MMT has managed to retain its data team as it focuses on hiring employees with domain expertise with a few years’ experience, says Sanjay.
“This is a cyclical process and we hope sanity prevails over a period of time,” he adds.
Edited by Megha Reddy