Indian Railway Catering & Tourism Corporation's Flexi-Fares strategy jolts passenger occupancy and premium trains revenues – can dynamic pricing engines solve the revenue optimization challenge?
25th November 2016, Profisor Team makes a day trip from Delhi (The Capital City) to Ludhiana (Major City in North Western State of Punjab). An early start to the day . . all preparations done for a two-fold agenda 1. Conduct regional-level research for analysing a political party’s penetration using social media data and 2. Meet Professor Doc. K. S. Mann, Dean of Student Training & Placement Cell and Heads of IT Department at the Guru Nanak Dev Engineering College in Ludhiana to establishing an internship program for graduate and master level students.
Week before this planned travel, we booked our train tickets for Shatabdi Express, a premium Indian train that runs during the day usually with two types of passenger coaches – An Executive Class and A Chair Car. Quite recently, on 9th of September 2016, Indian Railways Minister, Suresh Prabhu, had introduced Flexi-Fares (or Dynamic Fares) claiming to rake-in additional INR 500 Crores (USD 77 Million) by 31st March 2017. Much less to our surprise we paid 40% additional fare compared to our past travel experience and discovered a lower passenger occupancy in our coach. Is this the result of Flex-Fare System? Or Surge Pricing? The Uber Ola Debate! Upon further enquiry, from fellow travelers and train staff, we concluded with mixed, yet divided, responses. Passengers opined that the new fare system was negatively impacting the train occupancy rates, while the train staff appeared more optimistic hoping for improvements sooner than expected. As the calendar turns into December 2016, a big question looms on the face of Railways Minister, who has a strong bend towards technology, especially the use of social media for listening to train passengers and improving passenger experience. Does the Flexi-Fare system work? Owing to further decline in occupancy rates, on December 19, 2016, Railways tweaked the Flex-Fare Structure, offering 10% rebate on seats left vacant after chart preparation. Today we are in the month of March 2017 which marks the end of a financial year. Economic Times reports, “Flexi-Fare system in premier trains to be revised again: Indian Railways”. Railways earned about an additional INR 260 Crores from the Flexi-Fare System, nearly 48% short of their planned target of INR 500 Crores, and continued to face lower occupancy rates. What that means is fewer people travelled between 9th Sep 2016 and 28th March 2017, on trains operating with the Flexi-Fare System, paying a higher fare for the same journey, which in result contributed to the additional INR 260 Crore earnings. That’s quite a dent in the passenger’s pocket! Unfair! While that’s an opening statement for a big long news hour debate, let’s explore the dynamic pricing challenge for Indian Railways. Can Dynamic Pricing Engines Help Accomplish Higher Passenger Occupancy and Optimize Revenues? Here are some informational facts about Indian Railways and the Flex-Fare System or The Dynamic Pricing Engine.
How does the Indian Railways Flexi-Fare System Or Dynamic Fares Work?
What was the impact of Flex-Fare System on Revenues and Occupancy Rates?
Delhi to Mumbai travel is one of the most heavily booked journey and an equally relevant travel route for both domestic and international Airliners in India. A 10-week analysis of the route presents us with the following stats:
Delhi – Mumbai Rajdhani Express Train
Period Between 9th September 2016 to 19th November 2016
Delhi – Mumbai Rajdhani Express Train
Outside Festive Season Between 9th September 2016 to 15th October 2016
AKR, Mumbai – Hazarat Nizamuddin, Delhi Train
Period Between 9th September 2016 to 19th November 2016
AKR, Mumbai – Hazarat Nizamuddin, Delhi Train
Outside Festive Season Between 9th September 2016 to 15th October 2016
Against the principles of Dynamic Pricing, the current Indian Railways Flexi-Fare System changes fares based on Supply. Such an analysis must begin with a Segmentation exercise, a study of the Demand Supply factors and the total inventory of train seats pan-India, to arrive at trains and categories of seats that offer revenue enhancement opportunities. Opportunities that reflect operations research and data-oriented statistically prudent methods for successful implementation of Dynamic Pricing, Upselling, Resource Optimization, Meals Cost & Sales Optimization, Route-wise and Segment-wise Meals Personalization, Train Schedule Optimization (although politically driven . . ) . . and many more . . . .
Are You Flying High Enough? - Offers insight into how such studies are performed in the Airlines industry.
Data is rich. Technologies are emerging. In today’s times where Data Scientists are joining hands with business to solve computational problems, Indian Railways must make a genuine effort to evaluate all factors that contribute to the study of Dynamic Pricing of train passengers, and exercise caution when analysing the impact of fare changes on the travelers’ pockets.
Why Technology needs a special focus when aiming for revenue or inventory cost optimization?
Before we explain the role of technology, let’s remember the problem statement. Case in perspective is Revenue Optimization – a challenge to manage perishable inventor sales within a defined lifetime of the inventory to improve revenue. We discussed the imperatives of Factor Analysis and Statistical Models in our previous viewpoint – Statistics Can Be Majestic If The Factors Have Characteristic. Multiple factors, rising counts of data and studying multiple data relationship requires intense analysis. Make it a real-time problem and you need a robust compute-intensive solution, ready to deliver at speed, at scale and with highest precision at scale. One miscalculation or wrong choice of factor that contributes to the problem of revenue optimization can sink the ship.
Are You Flying High Enough? A use case that offers insight on how dynamic pricing helps optimize revenue.
Achieving revenue optimization through a technology based solution is an act where machines prescribe different price points for a product, based on a real-time analysis of all the factors that contribute to the demand and supply analysis of such product. Hence, technology architecture plays a significant role in delivering these accurate price prescriptions. Let’s break this further into a study of different aspects of technology.
Emerging technologies have far superior capabilities compared to the ones we currently operate – data consumption, data processing, data cleansing, parallel computing and analysis. Their viability, flexibility, strength and scalability must be adjudged before putting any ideas to practice.
Next up, the last 100 meters, Products Or Solutions, Seek Advise To Plan Wise "Why is it important to invest time and effort into the study of factor analysis and statistical Science when designing a dynamic pricing model?"
It is a race against time and tough competition to reach the finish line, first. Planning and preparation is essential. Breaking into the revenue optimization problem and complexities associated with perishable inventory management, those discussed in our first viewpoint – Make A Move Or Choose To Lose, we concluded with a set of pertinent questions that the business must answer when they decisively undertake a journey towards devising a dynamic pricing engine.
The next 100-meter relay readiness is about shaping a study of Factors Analysis and Statistics. Factors that represent commercials, competitors, customers, seasonality and the differentiated business model. Followed by, a prudent choice of statistical models, ones with evidential accuracy, that will drive price prescriptions for revenue optimization.
Judicious selection of data sources forms the fundamental basis for Factors Analysis and Statistics. Taking events data into perspective, by way of example, 26th January marks the Republic Day celebration in India. Parades showcasing India’s defence capability and its cultural and social heritage take place at Rajpath in the Nation’s Capital – Delhi. Amidst high security alerts, the Nation’s President, the Governing Machinery and a Chief Guest travel into Delhi to celebrate and pay their tributes. Necessary security measures enforcing preventive checks impact the state border, city and inter-state trains, metro rails, air traffic and road transport. A quick look at the week of 20th Jan to 26th Jan offers the following information:
Changing trends, otherwise unavailable in seasonality plans or pricing plans of airliners and hoteliers, can be converted into the business opportunities, using statistical algorithms.
Are You Flying High Enough? A use case that offers insight on how dynamic pricing helps optimize revenue.
Information on events is publicly available which could be aggregated to build an events repository. Then, events that have a strong correlation with booking curves must be shortlisted for a comprehensive study of demand trends. An equally relevant source of real time traffic information is also useful to understand the impact of no-shows on changing booking curves. Speaking of which, Google, Bing and Yahoo Map APIs (and soon Uber Maps) provide intelligent traffic data insights.
To conclude succinctly, statistical algorithms are only as good as the quality of data sources, business assumptions and the quantification techniques. While linear regression remains the favourable choice for arriving at the dynamic price range, for accomplishing revenue optimization, developing and implementing a statistical model and validating its decay rates requisites feature engineering capabilities. Real-time model operability for pricing decisions is most successful when technology infrastructure is architected to engender statistical modelling at its highest precision levels.
The buck doesn’t stop here or with the development and implementation of a statistical model once. Frequent examination of the changing characteristics of factors and validation of the statistical models for change in decay rates is essential to find the new optimized levels.
Next up, Investment In New-age Technology Architecture Prevents A Fall & A Fracture How dynamic pricing and inventory management models can help optimize revenue and perishable inventory management costs?
Numerous businesses, across industries, deal with perishable inventory in everyday operations . . be it goods, like fresh food, meat, chemicals, composite materials, blood products, medicines, surgical instruments, clothing, etc. or service based products, like flight and train seats, hotel rooms, concert tickets, etc. . . managing shelf-life, shelf and or storage space, time bound consumption or utilization, and time bound sales necessitates finite decision making to optimize revenue and inventory costs. This challenge spreads further to logistics and supply chain businesses who have a significant role to play in managing the perishable inventory lifecycle.
Over the years, industries and businesses have made significant investments in tools and technology that deliver streamlined processes, generate the much-needed data and information, and help in day-to-day decision making. With advancements in business, operations, transaction size, data size, data type and evolving relationship of data to transactions, the number of factors that form part of the problem statement continue to evolve and expand. Changing and evolving factors have made the decision-making process more complex and raised the demand for comprehensive analysis of multiple factors.
This is where statistics and science are joining hands with technology to produce prudent predictions and prescriptions, and to offer intelligence that drives revenue growth and cost optimization. Revenue Optimization challenge is one such study of several variables that can be categorized into commercial factors, competition factors, seasonal factors and customer preference factors. 
Before one begins to study these factors, an equally bigger challenge is to define and manage the life of perishable inventory. While the shelf life is well defined for most goods, it is the service based products that demand a perishable life definition.
By way of example, what is the perishable life of a flight seat or a hotel room? . . and then more complex questions that require additional analysis, like does the perishable life of a flight seat change due to seasonal factors? or does it differ based on flight routes or flying distance? These and many more questions that contribute to the challenge at hand. Refer the example below of a short haul A 380 flight from Sydney to Melbourne that has 371 economy seats. 
Are You Flying High Enough? A use case that offers insight on how dynamic pricing helps optimize revenue.
Appreciating the challenge, we comprehensively define the problem statement with the following questions:
Many early stage movers have undertaken this challenge either by way of research and development and or M&A investment activity to build or own essential new-age technology capabilities that can solve for such problems and many more. Are you preparing to match the pace or are you already in the race? Continue to watch this space as we offer more insights into factor analysis, statistical science, role of technology and available capable solutions that can help you match pace with innovative data science techniques. Next up, Statistics Can Be Majestic If The Factors Have Characteristic |
AuthorProfisor Team - Proficient Advisors. Categories
All
|
RSS Feed