Hotels.com CTO, Thierry Bedos, discusses what worked for him when undergoing big data analytics projects
It’s one thing to understand the principles of big data analytics, but it’s another to ensure projects are successful.
As an early adopter of advanced and predictive data analytics, Hotels.com has gained in-depth knowledge on providing consumers with a personalised user experience and custom-made recommendations.
Drawing on data from clickstreams, reviews, personal preferences, and hotel profiles, we are able to build algorithms behind the booking procedure that understand the customer journey beyond just one session.
Here are five key things that helped make big data analytics projects at Hotels.com successful, which could help any business.
More than big data
Big data is closely linked to analytics, but think beyond just big data. At Hotels.com, one of the most important uses of data is to deliver a cross-device user experience for multiple screens and independent from where they are.
One part is analysing the user data in order to determine which devices are being used. The other part is logging data in order to be able to recognise the destinations each specific user has been searching for on the company’s booking platform, and to ensure that customers are recognised on any device so they can pick up where they left off in their search for a hotel.
This allows users to begin looking for accommodation, for example, on a tablet while travelling by train, before confirming the booking on a desktop at home without the need to start the booking journey again.
This is only one example of how Hotels.com uses its data platform to improve the booking experience and make it as personal as possible.
Choose the right platform
We all know that the right technology platform makes or breaks an IT project; the same can be applied to big data and analytics. So how do we make the right choice?
A technology decision should be based on a thorough assessment of business needs and deciding what could benefit from data analytics in the future.
CIOs also shouldn’t focus purely on investment costs when choosing the technology. The evaluation should also include performance, reliability, usability, data security and, most importantly, scalability.
After weighing up all the pros and cons, Hotels.com chose Datastax Enterprise as its online data platform, which is based on the open-source Apache Cassandra NoSQL database. This allows us to benefit from features such as built-in management services, extra security features, and external Hadoop integration that complements the features of a pure open-source solution free of charge.
We also benefit from Datastax’s support, maintenance and update services. That leaves us free to focus on data analytics and supporting the Hotels.com business.
Get the bosses on board
A big data and analytics project requires both investment and cross-company collaboration; silo thinking could be an obstacle to long-term success.
For Hotels.com, one of our main goals was to break down the barriers between the online world of our booking platform and the offline world with our data warehouse solution at its heart. This required a massive change in entrenched processes.
We found that the best way to get backing from the bosses was to prove that we could achieve a quick return on investment (ROI). To demonstrate this we collected more than 150 business use cases that would be possible with the proposed platform.
We then selected a subset of 10, which were suited to illustrating a proof of concept within a narrow time frame. Seeing our proposed platform delivering quick wins helped to convince the board and served as a stepping stone to more challenging use cases.
Data privacy comes first
Customer trust is a precious commodity and respect for data privacy is one important key to success. Therefore CIOs should ensure that their big data analytics strategy is carefully balanced with a commitment to protect customer data security.
The use of anonymisation is vital to protect every user’s privacy, especially when analysing large quantities of aggregated data.
Make data analytics business as usual
Regarding cost, time and organisational restraints, project leaders should always keep sight of long-term development to build a reliable, efficient and future proof platform.
Scalability is an important component in long-term planning to ensure that the technological platform is able to keep pace with the ever increasing flood of structured and unstructured data.
CIOs should start to integrate the analytics processes into everyday business once the platform has been established and first use cases have yielded results. Only when this has become business as usual should IT and data teams address new projects to advance the company’s business even further.
Original content was posted here: http://www.cio.com.au/article/591129/5-key-things-make-big-data-analytics-work-any-business/