The truth about being ‘data-driven’: what airlines must understand

By Hélène Dubos | Personalization

Apr 29
Thuth about being 'data-driven'

From agriculture to aviation to automotive, companies around the world are increasingly using data to gather insights into how their businesses work, boost efficiency and develop new opportunities.

A data-driven business allows for a wide range of benefits. For instance, data allows companies to offer a more personalized service to customers, something that would have been much harder - if not impossible - just a couple of decades ago. Now, with the insights collected from more data, they can truly understand their customers and offer them unique services. And it's not just with personalization that data can help. From customer acquisition to marketing to operations, data can reinvigorate many different aspects of a business.

However, being truly data-driven requires a whole new approach to business. While hiring data scientists and using data management technology is an essential part of becoming data-driven, the greatest success is to be had when businesses transform the wider company and its processes. It’s about making changes from the top to the bottom of the company - about change management as much as technology.

For the aviation sector, there's a huge opportunity to become data-driven. But it can be valuable to look outside the industry to see how businesses in totally different sectors have changed their models to become truly data-driven. Let’s look at three companies that have not only invested in Big Data technology but really understood the need to change their processes and business models too.

Zara: using data to drive clothing design

As the world’s most successful ‘fast clothing’ brand, Zara demonstrates how Big Data can be a game changer for customer-facing businesses. Through a laser focus on collecting information about customer buying patterns, the company is able to respond to every shift in demand and taste, and then design new clothing items in response. Rather than predicting trends in fashion in advance or looking backward, the firm focuses instead on a quasi-real-time collection of data to figure out what consumers want right now.

Each day, Zara’s data HQ draws in enormous amounts of data from sales terminals, RFID tags on clothing and handheld devices carried by staff who record what customers like or dislike about everything from buttons to zips to the material. Rather than producing huge, time-consuming runs in factories in the Far East, they instead produce smaller batches in Europe to trial what is and isn’t popular thereby slashing production times and gathering information from customers and stores about what features are popular and which won’t sell.

This information from their thousands of stores is then fed back to analysts to crunch through. These analysts then provide insights to Zara’s fashion designers about what clothes are and aren’t popular - and this information guides them not only on the fashion items but drives their entire supply chain

Key strategy change: In the past, mass clothing design was based on a mix of imitating ‘high fashion’ and predicting what consumers wanted from surveys, observation and “gut feel”. Thanks to the new data-driven model, Zara has completely revolutionized its approach to fashion retailing.

Takeaway for airlines
Airlines have access to a huge amount of real-time data on customer preferences - from routes flown to meals ordered or the kinds of products bought from the duty-free magazine. Data analysis can help airlines respond much faster and responsively to shifts in demand - making sure they don’t stock up on goods or services that few passengers really want.

Assistance Publique-Hôpitaux de Paris: predicting resource requirements

The Assistance Publique-Hôpitaux de Paris (AP-HP) demonstrate how a healthcare organization changed its internal processes to become data-driven. In the past, resource management at hospitals was carried out by shift managers who would estimate how many nurses, doctors and other emergency staff would be needed on hospital wards at different times of day throughout the year. As a result, there was always a risk that a hospital would be under-staffed on a busy week, or over-staffed which would be costly.

The hospitals introduced a new Big Data management system which collated data about hospital admissions from the previous ten years, and then also crossed this data with other publicly available information (on flu outbreaks or weather extremes for example). This was then able to predict more effectively how many staff would need to be on hand during shifts. It also fed this into an easy-to-use dashboard for staff.

Key strategy change: The hospitals completely changed their resource management model, moving from the individual shift manager’s guesswork to giving them an easy-to-use tool which would help them plan better.

Takeaway for airlines
While airlines have long been able to predict passenger numbers, there is much value in drawing on their data to accurately predict a wider range of variables. For instance, even today one in four flights is delayed-airlines need to be using both internal and external data to predict disruption more effectively.

Danske Bank: using data to detect fraud

Fraud is a major challenge for any bank - and in the era of online shopping and phishing scams, the issue has become even more pressing. Nordic Danske Bank had been attempting to deal with this issue using human-written coding to alert any unusual behaviors. However, they were struggling with a low fraud detection rate and dealing with around 1200 false positives each day - which wasted time for staff and customers.

The bank introduced new ‘deep learning’ technology to analyze data more effectively and spot instances of fraud much more efficiently, reducing false positives by 60%, and boosting the true discovery of fraudulent behavior by around 50%.

Key strategy change: The bank changed its mindset about detecting fraud - moving from their own human-selected rules to choosing a system which used machine learning to discover fraud.

Takeaway for airlines
Airlines should review how they carry out many tasks which are currently manual, assessing how they can move humans away from time-consuming manual work to using data to make work more efficient.

Implications of becoming data-driven

Becoming a real data-driven organization is about much more than adding a layer of Business Intelligence software to your existing systems, hiring a data scientist and a team of analysts. When performed effectively it is about optimizing the entire system and re-evaluating how the company operates holistically, subsequently re-engineering systems and processes to serve the most valuable thing that every single organization in the world has - its customers.

In the next two blogs, we will look in more detail in two crucial areas: How airlines must transform their marketing and sales departments to become truly data-driven, and the transformation they need to apply to their commercial operations teams.