In May, as air travel bookings were in freefall due to the novel coronavirus, IATA and McKinsey & Co. shifted focus on their existing data partnership to attempt to see where the bottom was and where signs of recovery might appear.
Initially focused on more general airline retailing issues, IATA and McKinsey quickly crafted a database, dubbed Air Travel Pulse, to synthesize the enormous and wide-ranging array of travel-demand indicators (such as air-travel search and ticketing data, flight schedules, and cancellations).
Air Travel Pulse also includes publicly available data sources (such as travel restrictions, online search behavior, and other traveler activity).
The over-arching goal of Air Travel Pulse is to assist airlines, as well as other interested parts of the travel and hospitality industry, to get a better sense of the progression of the booking funnel, deploy capacity, allocate marketing resources, and adjust pricing in the months ahead, says Nina Wittkamp, a McKinsey associate partner based in the global consultancy’s Munich office.
Kambr Media: How did McKinsey & Co. initiate Air Travel Pulse?
Nina Wittkamp: It started with discussions we had with IATA at the end of last year about understanding the value and the prospects for retailing initiatives. At the beginning of this year, we got together again and discussed, "What's the agenda for this year?"
Then Covid-19 happened. We spoke with IATA again and asked, "How can we help the industry to get on its feet again and what can we do to help everyone now make the right decisions, whether it's in distribution or commercial more broadly?"
What we realized very quickly was that there was actually no data source available to do this. There was no information that could help you get an understanding of how demand is developing, or how quickly it was slowing down. Also, there was no way to find where demand is recovering. We needed a tool that could tell us demand is recovering before we actually see the bookings, because all the planning processes need a little bit of lead time.
For example, for pricing and revenue management, you would typically look at the past year’s booking behavior, which would allow you to calculate booking curves. That would simply not be valid this time around. First of all, there was no baseline anymore and second the current algorithms would just not apply. Automatically, the systems would simply just price down to stimulate demand, because that's what they learned.
We realized we also need to provide transparency to the industry to make sure we're all making the right decisions. Given that IATA has a number of data partnerships, as well as the DDS database, which is very comprehensive, and all the data from the booking systems, they're able to see the entire booking pipeline and the transactions that are going through the systems.
We said, "Why don't we create an understanding of how the demand development funnel would apply in this situation?" We started to model the entire demand development funnel from the onset of the disease, the various pandemic states, how the virus spreads in each geography, and how it develops. We looked at what travel restrictions are put in place because in many places we do see demand, but then you can't really go anywhere.
How does Air Travel Pulse work?
We calculate traveler sentiment by looking at web and search sentiment. We also look at search volumes with regards to mobility, the travel industry broadly, and then air travel specifically.
Looking at search behavior tells us how far out people are thinking about travel. We then look at how it converts to bookings. Then last but not least, we're also looking at how is airline capacity trending based on the demand development that we're seeing. Is it tracking? Is it tracking faster or slower?
What is the data telling you right now about the state of air travel demand?
What's really interesting is that we now see a couple of geographies with significant demand uptick.
For example, in Europe, we see the geographies that have been less affected by Covid-19, such as Germany and some of the Scandinavian markets, develop demand for Mediterranean destinations now that the travel restrictions have been lifted. Germany announced a lift of Europe-wide restrictions as of mid-June, and almost on the day of the announcement, we saw the search volumes for international travel go up significantly.
Last week, in Germany, most searches were for flights to Spain and Turkey, followed by domestic flights and flights to Greece, Italy, and Portugal.
We already see that German travelers represent some of the biggest inbound markets for Spain, Greece, and Italy.
It shows that the summer travel markets are coming back for many destinations. Daily bookings with destinations for Spain, for example, increased by 33 percent within one week. For Greece and Spain, we've already reached more than 50 percent of the booking volumes for July to September compared to last year. Our data indicates that many tourists decided to delay their travel to October, where booking volumes for Greece and Spain reach around 70 percent compared to last year.
What specifically is the search data telling you about the prospects for an uptick in air travel bookings?
For most geographies, we see that searches for travel and bookings are still very much decoupled. It seems that travel is top-of-mind for a couple of markets and people search again on the different travel websites, but the bookings are still very, very cautious.
For the markets that are maturing a bit more on the recovery curve, at some point we do see the conversion and it happens with very short lead time because people really take the time to be 100 percent sure that they can actually go.
It also seems that any change in search and booking behavior is immediately linked towards government announcements on travel restrictions and lifting restrictions. That seems to be a big influence for how people are booking. In some countries, after travel restrictions are announced to be lifted, air travel sentiment increases by 50 percent.
You mentioned that sentiment analysis is a key part of Travel and its ability to see changes in demand right when it starts. How do you measure sentiment? Is it primarily search?
Sentiment is a step in between, so we're looking at positive sentiment towards travel. Web searches, for example, as well as social media chatter about travel tells us a lot. Travel searches are a clear form of sentiment, as people entering a booking request in a travel website or in a travel agency is how they tell us their travel intent.
We do see that, first of all, the sentiment towards the disease typically changes. Then we see a positive uptick in terms of mobility and then travel more broadly and then, typically, air travel follows. It's quite interesting for some domestic market, e.g. Finland, air travel hasn't followed and those are the places where we expect a modal shift to happen, where people take the train and other modes of transport instead of flying.
How does McKinsey incorporate other data sources besides IATA’s information into Air Travel Pulse? Or do you mostly rely on internal data sources?
It's a combination of things. There were a couple of sources that we've already been subscribing to, including IATA sources and third-party sources like schedule data, which is, of course, one of the fundamental data sources in the industry. We’ve also pulled in a few more proprietary databases on the development of the Covid-19, models that we developed ourselves. Then there's also a lot of, essentially, public information that is out there on travel restrictions.
The magic here is really pulling all these sources together into one integrated view and deriving insights, looking at it holistically. The “big unlock” in the insights from air travel polls was the ability to truly connect the dots between the different sources and the steps along the funnel.
In terms of Air Travel Pulse’s mission, who are the intended users, specifically? How do you expect the user profile to evolve?
It was initially designed to help the airline industry get on its feet again. In the first release we made it available to IATA members. The organization would share a summary update of the insights to the broad audience or the general public.
We've gotten a lot of requests from airlines, of course. But we’ve also gotten requests from travel agencies, intermediaries, and hotel chains. We’re hearing from many tourism destinations, which are trying to understand how the demand pipeline is looking with respect to their properties. They ask: “Where's the demand coming from? Is it changing? What are the geographies I should be marketing to? Should I maybe make any adjustments in my product?”
I would say that while it was most initially designed for airlines, it looks like the entire travel value chain naturally finds value in understanding how demand is recovering.
Do you expect Air Travel Pulse to be diversified into other verticals to serve specific parts of the travel and hospitality industry?
We're developing Air Travel Pulse as we speak. There's an active intention to understand other parts of the value chain better, because we know that a flight decision is typically not an isolated decision. It's done in context with a certain travel purpose, such as a vacation or trip where you also make other kinds of bookings.
We've gotten questions from a lot of airlines such as "I'd love to understand how the hotel sector is opening up in the destinations I'm flying to." We've gotten the same question from the hotels who ask, "Should I open up because the flights are coming back – or is that not the case?"
There is definitely value in extending it along the value chain. And then also we now work with a couple of airlines and other travel companies in going a lot deeper on their specific O&Ds to understand how the demand is developing specifically in the markets that they're serving.
Looking beyond pandemic and the potential recovery in demand, how do you think the data needs of airlines and the rest of the travel sector will change?
I hope that some of these approaches that we’ve started taking will remain. We think there is an opportunity for pricing revenue management in particular. There’s a realization about the importance of looking a more intently as what the real drivers for travel are and how travel demand is impacted. We see the value in seeing how individual micro-segments are developing versus forecasting based on last year's behavior and last year's volume. This is a first step to show what additional insights you can find when you include third-party data, alternative data sets, and just more contextual data on who your customers are, where they're from, where they want to go and what their sentiment is.
As you said, Air Travel Pulse is still a work in progress. How should airlines and other travel entities contribute and receive Air Travel Pulse information?
Generally, IATA’s airline members should have received access by now. We're happy to discuss with any interested players to see whether they want to contribute to this emerging data platform to just enrich the data set.
On a personal level, have you been flying lately? Do you have any air travel plans coming up?
I've been flying already. I'm looking at a summer holiday, which this time around will be a driving holiday. But I've generally been flying again, and I find, personally, this whole information about where I can go and what the restrictions are in those destinations quite helpful. It’s crucial to know if there's still a quarantine in place, whether all the surroundings are open and available and so forth for traveling.
If I can't spend time at a beach or can't go to a restaurant then, of course, the destination becomes less attractive. That's definitely helpful to follow and look at. We're also reflecting some of that in our dashboards, including travel restrictions from a country perspective, but then also airport travel restrictions.