What do you do when your crystal ball cracks?
In a panel discussion led by Sinead Finn, Founder, Affinity Ltd., Bryan Porter, Head of Commercial — EMEA, Accelya Group, Jason Coverston, Director, Office Domain, Navitaire, an Amadeus Company, Sophie Dekkers, CCO easyJet, Achim Tyler, Vice President of Global Sales, Infare Solutions, and Krassimir Tanev, Chief Commercial Officer, Blue Air each shared their views on changes in the marketplace since 2019 and the rapid adaptation required to stay ahead, particularly in revenue management systems.
“In the last 20 months or so, is it’s very much a switch towards away from looking at historical behaviour and historical data into much more forecasting and forward-looking; which is a challenge, because who knows whose crystal ball is right,” said Sophie Dekkers, CCO, easyJet. “But we are having to switch and use much many external data sources to help make sure the points that we’re looking at in terms of forecasting are correct… In terms of the revenue management system itself, it’s trained to price at certain points with certain low matter. We’re in a very different time now, with a much later booking window. We were at about 75% of our sales were coming in the last three months [before travel]. Recently that changed to around 63% now [booking] in the last three months for travel next three months. That’s still very near-term [bookings] versus what the system had learned over the last 25 years, built to look much further out.”
“How do you stimulate demand? Once they get to the website, the conversion is there because the prices are attractive at the moment.”
Customer behaviour has also shifted somewhat in terms of apprehension to book, although Dekkers said the airline had seen a marked improvement in conversions, returning to 2019 levels. “But it’s the demand traffic in the first place that isn’t coming in. Once they come in, they see the prices, and they’re converting. So, our conversion rates are the same as in 2019. But it’s the demand that’s still depressed. If we look at searches on Google for flights, generically, they’re still down about 40% of what they were from 2019. That demand is the challenge—how do you stimulate demand? Once they get to the website, the conversion is there because the prices are attractive at the moment.”
Krassimir Tanev, Chief Commercial Officer, Blue Air, explained that the company as a whole had reshaped its business model from a hybrid budget airline to a true low-cost airline, with a greater focus on ancillaries.
“We have deployed a robust commercial plan over the last 12 months, focused on three main pillars that underpin our developments. The first one was network development, or rather, network enhancements, adding more value to our customers by focusing first on primary airports. We have tried our best to expand our operations rapidly and flexibly into the gaps that full-service carriers have opened up. We’ve opened up new markets like London Heathrow, Amsterdam Schiphol, Milan, Linate, and Frankfurt, just to name a few. The second priority for us was the customer experience elevation. Here, we also made significant progress. We used this time to upgrade our products. We have added a new fleet with a state-of-the-art, cost-efficient 737 Max 8 aircraft. And on top of that, we have also enhanced our customer experience capabilities and customer notification platforms–we’ve added new chatbots. Our digital systems, we’ve improved overall. We focused on adding more value to the customer experience. And last but not least, we focused a lot on actually driving growth in ancillaries. We know that most carriers have reported weaker or softer yields throughout the pandemic. But one thing that we delivered very well was that we provided more opportunities to our customers. We increased our ancillary yields by more than 50% thanks to our new ancillary strategy, where we have unbundled our products further. We have created new product features and new product bundles for our customers. And we have deployed dynamic ancillary pricing, just to name a few of our actions. But more importantly, we have enhanced, or we have increased our exposure to the VFR (visiting friends and family) markets, which has proven to be much more resilient compared to other market segments.”
“We work very quickly with existing customers to start consuming new types of datasets”
Bryan Porter, Head of Commercial — EMEA, Accelya Group, shared some examples of how the company’s airline customers had adjusted RM when historical data proved less useful. “Norwegian started using Accelya’s airRM revenue management solution back at the end of 2019. So we implemented a solution that effectively used historical behaviour to predict future behaviour. We built in all the forecast models, and, of course, along comes COVID–none of that is effective anymore. [How we deal with that is] we work very quickly with existing customers to start consuming new types of datasets. This is taking in market insight data and competitive fare data and utilising that to optimise prices. We also started pulling in data from the look-to-book activity that we’ve seen on some of our customers’ websites and other data that allowed us to focus on price optimal optimisation on an intraday basis. So if there’s a sudden, unscheduled event or a sudden fluctuation, our customers were able to cater for that. We effectively moved midstream with Norwegian. At the same time, we had customers coming to us—Iberia being one of them—who had barely been using their traditional RM system and suddenly found that it wasn’t fit for purpose. So they started working with us, implementing some of the solutions similar to what I’ve just described for Norwegian.
“Another example is we have seen the rise of new low-cost carriers that have effectively started during the pandemic. They don’t have any huge debt burden, and the cost of entry into the market is lower than it’s ever been. So we’ve had airlines coming in, who have no historical data to optimise their prices effectively, and are looking for new tools that can suddenly get them into the market. It’s been an interesting time for us. We’ve managed to onboard 11 new customers during the pandemic—our team’s been very hard at work. We’ve effectively introduced 280 discrete additional value points to our product set. We focused on new data integration, better visualisation to allow revenue managers to identify trends and start making more active decisions, as well as to look at automation and intraday optimisation.”
“One of the things that we’re working on with some of our newer customers is total offer optimisation.”
Porter also discussed how the company’s acquisition of FareLogix has helped refine the pricing of both fares and ancillaries. “One of the things that we’re working on with some of our newer customers is total offer optimisation. Through our flex merchandising module, we’ve started revenue managing ancillaries. We can start using price optimisation across the ancillary products and start doing that with bundling as well. [We’re] working with a large European airline currently to effectively couple our airRM revenue management solution with our merchandising solution. We are introducing AI (artificial intelligence) and ML (machine learning) to enable ‘willingness to pay’ models. We’re not just looking for an entry bid price, but ultimately at what the customer is willing to pay and utilising that to optimise price.”
Achim Tyler, Vice President of Global Sales, Infare Solutions, said data supply to inform RM models has increased.
“We’re looking in the future, then looking into historical data. We don’t throw away any information we collect—it’s still available. If you want to go back to 10 years and further, if you want to look into the market, we still do that. But last year and this year was always about the future… We invented a product called Market Trends, which observed the scheduled flights versus the flown flights, how prices were changing how quickly they were changing, how [many flights were] cancelled. That was last year. This year… there’s not too much churn. We focused this year more on optimising costs for our customers in the way we source our data. We did a couple of projects regarding API access to data—NDC API projects… We have already about ten customers on NDC APIs. NDC is the most efficient way for airlines to access the information they have on their website, which is our quality aim… We also see shopping data and ancillary data as a newer trend, talking about dynamic pricing and continuous pricing. You can imagine it’s going to be much, much more data. We work together with airlines and RM vendors to make this most efficient for all of us. A lot of data needs to be collected and consumed. You have to run all the algorithms, on top of that, to get to the really good revenue results.”
“One of the things that we’ve found is that our crystal ball is not as good as the crystal balls that our customers have—and they each have a different one.
Jason Coverston, Director, Office Domain, Navitaire, an Amadeus Company, also mentioned the importance of working closely and collaboratively with airlines to collect the correct data and refine predictive models. “You’re just talking about how much data it takes to react,” he said. “One of the things that we’ve found is that our crystal ball is not as good as the crystal balls that our customers have—and they each have a different one. So we try and get out of the way of our customers. For example, [one of our] customers realised they
needed to react more quickly [to] data coming in from Infare. We have a team dedicated to helping customers operationalise the data they get. So, we came in, and we created a pipeline. They were able to implement exactly the logic and algorithms they wanted. We take a very non-structured approach to deliver value to the customer. When it came to the algorithm choosing fares and exactly how to write the logic around how to compete, that was completely up to the customer. But they needed a bit of help in operationalising that because it wasn’t part of their out of the box system. So, whether it’s [data sources like that], or maybe somebody’s got a new Google TensorFlow model that they need to get into the system. We don’t use our crystal ball to decide [for our customers what they need, such as] you need willingness to pay, you need probability, you need x, y, and z. Instead, we say, we’ve got some patterns that we’ve seen to be successful, and they’ll show you what other customers have done. But at the end of the day, you can invent entirely new steps in the process so that you can be agile as the world changes from month to month.”
By Marisa Garcia
Revenue Management and Pricing
RM, continuous pricing and Demand Forecasting Panel Discussion: Rebuilding and adapting forecasting models with different inputs and future strategies to ensure that we capture demand?