Travel agents

Guest Post: The big data revolution needs joined-up thinking in travel

Posted by Lee Hayhurst on
Guest Post: The big data revolution needs joined-up thinking in travel

Big data is transformational for the travel business, Amr Hassan, chief executive of Match2Lists, discusses how the latest technology is helping the online travel agents stay competitive and bring on board more inventory

Every business category is striving to become data driven and travel is no different.

As an industry, it has been completely overhauled in recent years. The internet era has meant anyone can arrange an entire holiday or business trip from the comfort of their sofa.

However, with huge change comes the law of unexpected consequences and when everything is automated, what checks and balances are going on to ensure that you really are booking yourself into the Five Star Sunshine Hotel on the strip and not the Two Star Sunshine Hotel in a less salubrious part of town?

Data matching of Big data is the answer, and it’s being used behind the scenes to ensure customers’ travel plans run smoothly.

Transformational Travel

The first place many travellers turn to when booking a trip is the online travel agent.

The hotels they have on offer to their customers is supplied, behind the scenes, by the massive hotel inventory providers such as Expedia, Booking.com and Hotels.com.

These hotel data banks aggregate some 800,000 hotels worldwide.

For this partnership to work, customers booking a hotel on a travel agent’s site should of course expect to show up at their chosen hotel and find their reservation. There should be no confusion over which hotel they booked at a given location.

In an ideal world, every hotel around the world would be unique, without need to worry about two hotels on the same strip, village or town sharing the same name and address.

The hotel that the customer has booked on the travel agent’s site, should be the same as one booked through the hotel inventory’s provider.

But the world of data is never ideal

For any of the tens of thousands of online travel agents, their databases of hotels are updated at points in time. But hotels get sold, bought, rebranded or cease to exist at other points in time.

The suppliers of the hotel data, the bed banks, may carry different details for the same hotel (its name, address, postcode telephone numbers etc) which could be spelt or written differently on each of the competing bed banks’ systems.

The hotel beds banks, as aggregators of “bookable” hotels, and the travel agent, as front window shops to the end customer, both share the responsibility of ensuring that the customer gets to stay at the right hotel.

Otherwise, one of them will face compensation penalties.

Unintended consequences

The problem is, when everything is automated, how do you check that the room you just booked has actually been booked in the hotel you expected?

Hotel bed banks ( like Expedia, Miki, Hotels Pro, Booking.com) take care of their end of the data by ensuring every hotel is de-duplicated and given a unique ID, the bed bank’s property ID.

The Bed Bank then shares their data to their affiliated online travel agent. However, this just pushes the problem further down the stack to the independent travel agencies and affiliates.

Now the onus is on the agent to ensure they are matching their inventory to the correct bed bank property IDs of one or more bed banks. Easy for a handful of hotels, challenging when you have hundreds of thousands of hotels.

Making it a big data problem – and fixing it

It’s when an independent travel agent or affiliates has inventory consisting of only hotel names and addresses that it becomes tricky.

For starters, there is the challenge of linguistics. Spellings are sometimes slightly different – phonetic spellings, language differences or just plain typos.

Abbreviations and acronyms present challenges too.

For example, “St.” may stand for “Street” or for “Saint”, perhaps depending on whether it appears before a name (saint something) or after a name (something street).

The challenge comes with the quantity of data involved, the bed banks have hundreds of thousands of listings, as do their affiliated online travel agents, many of which overlap.

In the data there are many fields that could be compared to get a match – for example name, street name, postcode, telephone number.

It’s not just matching either, de-duping is required as booking rooms against a duplicate is just as big a problem as booking it against the wrong hotel.

It quickly becomes a problem worthy of the title Big Data.

Match2Lists’ answer is to develop the most intelligent matching algorithms (computer code that finds the linkage between differently spelt or written records) and combine that with the fastest analytic database, EXASOL, to deliver the easiest to use cloud-based data matching platform.

What previously took weeks or months with alternative solutions (whether in-house manual or alternative software solutions) can now happen in minutes.

The advanced matching algorithms allows bed banks to create bespoke solutions using Match2Lists to enable their affiliated online travel agents match their data with the highest accuracy and speed.

Giving the online travel agencies a competitive edge

The Match2Lists algorithms and the speed of EXASOL are now giving companies like Expedia and many other bed banks a key competitive edge for maintaining high quality hotel inventory data.

In the battle between the online travel agencies, it is a key tool to bring more hotels on-board and ensure the inventory of rooms on its site is up-to-date and accurate.

Big data is pervasive across many industries and it is bringing benefits that improve business operations, as we’ve seen with online travel agents.

With the right tools and insights, any business can derive genuine value from data.

Comments

This is a community-moderated forum.
All post are the individual views of the respective commenter and are not the expressed views of Travolution.
By posting your comments you agree to accept our Terms & Conditions.