Data and its promise of making the travel industry less average

As human beings we tend to take some comfort from the ‘law of averages’.

As human beings we tend to take some comfort from the ‘law of averages’.

It gives us some level of assurance that even though many events appear to be random most are at least partially predictable because outcomes usually fall within a pre-definable range.

The problem is that the average is just a statistical entity and doesn’t necessarily exist in reality or guarantee that anything in particular is due to happen.

This is what the Travelport Evolve conference in Monaco was told this week by former Traveltainment boss Andy Owen-Jones.

He told delegates: “Averages lie. If you are marketing to your consumers on the basis of averages there is basically no one who is an average.”

Owen-Jones now runs his own Bid Data company called bd4travel which is taking the conventional practice of AB testing on websites – effectively an averages-based approach – to the next level.

“We try to drill it down so you are not testing averages,” he said. “It’s gone from hypothesis testing to now so that with artificial intelligence you can try lots of different things.”

So the promise of Big Data is much more qualitative than the rather quantitative-sounding buzzword would have you believe.

It opens up the possibility for retailers to create entirely new customer segmentation matrices based on observed behaviours but predicted by massive number crunching.

What a customer doesn’t buy becomes as important as what they end up buying, and deeper insight into that person’s preferences will enables firms to offer a highly personal service.

Or if little is known about the customer – a scenario more likely in leisure travel where transactions are infrequent – learning algorithms can draw on massive amounts of data to predict need.

Owen-Jones illustrated how travel needs to progress to Netflix’s algorithm that pushes relevant content up the search results.

Or, taking things a step further, what about Amazon’s predictive analytics which, apparently, will ship products to customers before they’ve actually been ordered.

“What you want to do is start to identify the behaviours customers exhibit.

“Think about them as patterns buying patterns, identify the patterns that sell to other patterns you can mathematically work out,” said Owen-Jones.

All this takes a huge amount of control of data and, as was pointed out, travel is a sector in which online retailers often don’t even know the attributes of what they are selling until the inquiry is made of the warehouse.

So content becomes vital – content about the product, right through to content about the end consumer and all the factors, events and behaviours in the middle that unite the two.

And this is where a company like Travelport comes in with its vast amount of data generated by the 77 million daily search requests and 1.2 billion results it handles every day.

Reg Warlop, Travelport vice-president of search and transaction processing, said Big Data is opening up this sort of closely managed and targeted marketing so it’s not just the preserve of the big boys.

“We have seen year-on-year increases of 30% in search volume. And it’s not only volume, it’s the complexity and choice.

“Agencies want to sell more content, more ancillary services and they want to trip down those billions of results sets to something that’s more relevant.

“The whole industry has been structured around checking availability, checking schedules, pricing options, doing it multiple times over and over again.

“Large companies have had enormous stores for decades and have been able to crunch the numbers but now for smaller companies it’s accessible.

“It used to be about averages; now it’s every single event you can do something with it. We want to solve those complex problems. It’s not your core business to do that, it’s ours.

“Typically we have more data than any single participant in our network and because we have the search queries we can spot the trends that no one else can spot.”

Increasingly the value of the GDSs will be measured by how they can exploit their position at the centre of the travel industry to do three things:


  • Buffer airline technology against the rising demands of the direct and indirect customer base;

  • Enable travel agents to become more effective and efficient by targeting their products to more closely match their customers, and;

  • Allow the end consumer through which ever channel they choose, to quickly and accurately search and shop on whatever device whenever they want to.

Warlop said Travelport has three main propositions emerging for its industry partners: personalisation, benchmarking and recommendations.

“They all require a lot of data to be processed to establish who the customer is, to determine to which segment this customer belongs.

“We are in a position to marry that up for partners who specialise in segmenting customer for behavioural targeting.

“Travelport is sat on a goldmine of data. We could sell it to financial analysts but we are more interested in applying it to our own transactions.”

Remember, the fundamental concepts behind this aren’t revolutionary.

Ever since Dunnhumby helped Tesco overtake it’s closest rival Sainsbury’s in the 1990s firms have strived to emulate it’s approach to targeted marketing.

The hospitality industry was quick off the block to use insights derived from unstructured Big Data to improve their offering and Warlop said OTAs like Vayama, Expedia and Kayak have started too.

“The momentum is really building up,” he said, “already we are exposing more data, collecting more raw data and making it more available.

“It’s a win win; the consumer will be getting less noise, the agency will be able to offer better service and the airline will be able to sell more.”

A new search platform being developed by Travelport will be based on the principals of Big Data.

While exploiting the data the GDS already has this will rely on agents making sure their customer data is up to date and inputted in the right way.

An open platform and external developer network will also be able to siphon off the data it needs to come up with new products to advance to use of this data.

“It’s about control, brand diversification and options to upsell and cross-sell. These are fundamentally different design principals we are working on,” Warlop said.

“People spend more and more time planning their trips and we need to cater for that consumer behaviour. Why are metasearch sites so often used by consumers?

“Only so much is down to brand and SEO. The other thing a lot of meta sites have developed is an awesome consumer experience, far superior to many OTAs who still look like 2001.”