Qubit’s Future of Travel 2017: Industry experts discuss findings of ‘Man v Machine’ report. Lee Hayhurst reports
Machine learning: Honing data ‘can sharpen results’
Machines will increasingly be used to help customers hone options to a manageable number, but humans are expected to remain vital in closing the deal.
Panellists agreed that for many holidaymakers, interacting with another human being offered reassurance and had a significant impact on conversion.
However, they felt technology should be operating in the background to ensure sales agents have the information on product and customer to give relevant advice and suggestions.
Brian Young, G Adventures’ UK managing director, said tagging hotels based on their amenities to match them to customer requirements had previously been a manual process.
“For me this is about migrating from people manually doing it to having the technology at the front end doing it to get better conversion,” he said.
“The more it picks the data up and hones it down, the more you get to a position where what you are offering exactly matches what the customer is looking for.
“That’s where machine learning comes in. You can take all that data and really crunch it so what you end up with is very honed but at the same time a small offering.”
Panellists said once a small number of highly relevant options have been suggested, human intervention should come into play to convert the customer.
Fred Olsen Cruise Lines head of marketing Justin Stanton said: “People will always want that interaction. It’s still the case that 80% of all retail transactions in the UK happen in store because people still want to look at the product or speak to a sales person. I think that will continue, albeit the proportion of business in store will fall.”
Recommendations: ‘Tech has yet to blow doors off’
Technologies designed to automate customer recommendations have proved to be underwhelming to date, according to the UK boss of flash sale retailer Secret Escapes.
Sebastian Fallert said he has yet to see any system that had “blown the doors off” by achieving significant conversion uplift.
“I have yet to bump into somebody from the travel industry who says ‘we have found the holy grail that does this for you’,” he said.
“Look at Booking.com and Expedia: they are not exactly super-keen on recommendations. For Secret Escapes, it doesn’t work particularly well.
“A human is actually pretty good at looking at a deal and saying ‘that’s a great deal, I will buy it’. Although we’d like to do it, turning that into an algorithm isn’t easy.”
Fallert said Secret Escapes’ business model – finding deals each week it hopes will be interesting to everyone on its email database of millions – is risky.
And he said not all the technology developed to try to make this task more efficient would prove beneficial.
“Machines don’t always make good recommendations,” he said.
“I know it’s all very hyped, but most of the things that people call machine intelligence are super‑dumb. How often do you go on Amazon and the recommendations are totally useless?
“You can see how the machine got there but it’s not relevant.”
Wayne Perks, managing director of Teletext Holidays, said technology can improve recommendations by basing them on what people are currently looking at online.
“Historic data can tell you what a customer might like,” he said.
“But in travel, data ages quickly and it’s only the most recent that’s going to inform your decision about what to recommend.”
Craig Ashford, director of sales and marketing at Travel Up, said recommendations work when retargeting prospects with deals for properties consumers have probably already looked at.
Automation; ‘Business travel is easier to crack’
The decision about whether to automate processes with technology will be determined by the nature of individual businesses and the products they sell.
Panellists said they expect firms retailing commoditised travel products like flight-only to strip out all human element before long.
Much of the corporate travel sector is likely to go this way, but the more complex the product, the more humans are expected to play a role in the future.
Sebastian Fallert, UK managing director of Secret Escapes, said: “The beauty of business travel is it’s repetitive. But B2C in leisure travel is tough. You have to infer what the customer wants literally from 10 to 15 clicks, probably less, before they bounce off your site. There’s very little context.”
G Adventures’ Young said: “The commoditised sector will move very quickly because it’s not hard and actually the more you use it the more you learn.
“But other travel is a very complex nut to crack, with lots of different moving parts.”
Geoff Cowley, managing director of Wyndham Vacation Rentals, said there are lots of pain points for frequent business travellers that could be solved by artificial intelligence.
“You can very easily see the application in business travel. It should see early adoption,” he said.
Chatbots: Travel Up reveals plans for automated support
Travel Up is giving serious consideration to employing chatbots as it expands to the US, Australia, Canada and New Zealand.
The Aldermaston-based group of brands operates a call centre in Pakistan and has already launched in the US and expanded into the consumer sector for hotel-only.
Compared with its core flight-only business, there are more customer touchpoints in hotel‑only and holidays, and Travel Up’s Craig Ashford said expansion would present challenges.
“Use of chatbots is hugely dependent on the product. For hotel-only and flight-only, it can work.
“But a customer looking for a twin‑centre in Florida wants the reassurance of talking to someone. They want someone who gets enthused when they chat, and I just don’t think that’s going to happen with machines.”
Tracking devices: Firms ‘must avoid appearing creepy’
Travel firms risk being seen as “creepy” by customers if they use technology that feels too invasive.
Fred Olsen Cruise Lines’ Justin Stanton said the company is likely to follow the likes of Carnival Cruise Line and MSC Cruises in deploying onboard technologies capable of tracking customers as they move about ships.
But he said this must be seen as a way to enhance consumers’ holidays rather than as a means to extract more revenue.
“Consumers don’t mind that trade-off of the operator knowing stuff about them because it’s making the experience better,” he said. “But if it just results in a series of spam messages, then it becomes intrusive. It’s getting that balance right.”
Stanton said the over-60 age group was particularly cynical about the use of technology.