IN-DEPTH: Interview with LikeCube’s co-founder Eleanor Ford

Published: 31 Jan 2011

IN-DEPTH: Interview with LikeCube’s co-founder Eleanor Ford

Personalisation should be looked upon as a strategic journey with a marketing messaging that conveys the benefits for the user, all about relevance and time-saving.

A company like LikeCube says you know you will win when people will start referring to your service or brand as a "trusted advisor" or "trusted authority", or when they will bypass search engines to use your service. Reducing your reliance on SEO/SEM is one of the clear benefits of personalisation.

In order to know more about travel personalisation, EyeforTravel’s Ritesh Gupta spoke to LikeCube’s co-founder Eleanor Ford. Excerpts:

LikeCube uses machine learning and leverages semantics dynamically understanding what people might like at the moment they search, creating a unique personalised filter through which they can see what’s most relevant to them. What do you think have been major breakthroughs for machine learning considering that its still early stages for the same?

Eleanor Ford:

In recent years, very large data sets have been created (e.g hotel reviews) and have become more easily available through developer programs and APIs. The industry has also witnessed the mass availability of cloud computing offering high computing power at affordable cost. These two trends have enabled an increased level of experimentation in the field of machine learning. Certain techniques requiring very large datasets and high computing power are now delivering excellent results, both in terms of quality and relevance as well as in performance (results can be computed in near-real time), meaning that they can become viable solutions for commercial implementations. While it is still early stage and many more innovations are yet to come, the industry is now seeing a wider range of use cases being addressed successfully.

Travel personalisation is still in its infancy. How do you expect to shape up this year?

Eleanor Ford:

The current approach, profiling based on CRM and organising, analysing and segmenting the database, shows limitations in addressing the fact that people’s desires aren’t static. Machine learning can offer a real-time personalised discovery at the individual level. A bit like Amazon books recommendations, but for hotels, destinations and trips. There are clear drivers for it. The growing amount of content is making travel search more difficult, not simpler. Also changes in the industry like the Google-ITA deal or the more recent move from airlines with respect to the OTAs and GDS have strong implications in this space. More organisations are now concerned that without offering a truly personalised experience to their users, they will simply lose them to the organisations that simplify their search and discovery by using it. So the industry should see an accelerated adoption in 2011.

Last year you mentioned that the machine learning era has just started. Machine learning - algorithms that use data to infer behavioural patterns and predict things -, is highly dependent on the amount and type of data available. Considering that the industry is seeing an exponential growth of data that can be used in different ways to learn about people and places, and this involves identity, location, actions and things that are yet to be even thought of, what can the travel industry expect in the time to come?

Eleanor Ford:

The travel organisations can already benefit from using large amounts of travel specific data combined with social media data they already own or have access to. Innovations in machine learning and in using semantics also mean that less data is required to deliver personalised experiences that will drive loyalty while reducing reliance on search engines. So the solutions are available today. In the times to come, with the drive to further narrow-cast content, search, recommendations and personalised discovery will likely become more blended with innovations coming from the ability to effectively combine the personalisation potential of many types of data (eg searches, reviews, likes, ratings, bookings, check-ins).

Social media, from a personalisation potential point of view, can be of great use. How do you expect the utility of the same to gain traction?

Eleanor Ford:

Social media can be used in many ways to increase an organisation's ability to deliver a personalised experience. It provides simple access to some data about the user with little effort from them, which is great to address the challenge of dealing with anonymous users. It also can be used as a way to let users easily create data that will not only contains an interesting personalisation potential but also entice their friends and followers to create more data to be used in the same way. Learning how to interact with social media (eg via "likes" and "check-ins"), capturing this data and using it to best advantage is now a very relevant part of an organization's journey into personalisation.

The lack of data about users is a real issue for the travel industry. However, a few past experiences are great data points that can be used. How do you intend to address this lack of data going forward?

Eleanor Ford:

The cold start issue (lack of data, anonymous users) is very real for travel. But there are more ways to address it now. As mentioned, data from social media can be used to get some data about the user with little effort from them. Used wisely, efficient filters of content can be derived from it.. Also a few past experiences can be used to deliver some level of personalisation. For example, our cold start solution enables anonymous users to very simply express relative preferences (eg more interested in hotel service than location) as well as one or more past experiences and use this data to create a personalised filter and profile, matching them with results from users having similar preferences or with places similar to their previous experiences. Moving forward, we also expect cold start to be addressed by surfacing relevant content from reviews more effectively.

Considering that the website, email and mobile channels remain largely untapped in terms of personalisation, what do you recommend when it comes to capitalising on the same?

Eleanor Ford:

Organisations would benefit from taking a stepped approach and start with understanding their specific personalisation potential and goals. We can help with this, by assessing the data they currently own, thinking about new data they should gather (eg from social media) and making recommendations about where and in which channel personalisation would add most value today and in the future. In general, though, it often makes sense to start where you have the most user interaction, which in turn can generate more useful data for personalisation purpose, and then leverage it in other channels.. Our solution supports all these channels out of the box, making it simple to roll-out in any order.

What’s on your agenda for 2011?

Eleanor Ford:

With the expertise we've built over the last 4 years in the travel space, we’re certainly very confident for 2011 given these current market trends. And we have some exciting new solutions, using natural language processing (NLP), that we’re working on to deal with the overload of user generated content. These will appeal to both the travel sites as well as the users.