Big Data = Business Intelligence?

Big Data – it’s a big buzzword in the tech scene. This week I was asked by Slush, the annual startup and innovation event in Singapore, to give a keynote on the value of Big Data and how data insights can be invaluable when developing digital services – and ultimately spell the difference between success and failure.

Did you know that there has been over 4 million hashtags for Big Data on Twitter this year alone? Or that 90% of all data in existence has been created the last 4 years? Big data is definitely a topic in the wind. But what is it really? And how does it relate to business intelligence and creating value for customers?

In my daily job as Chief Marketing Officer for the global telecom operator Telenor Group, part of my responsibility include driving the development of new digital services, digitize the core business of the company as well as ensuring the best customer experiences through data insights and cutting edge technology.

Big Data is what it sounds like: Big sets of data that can be analyzed to show patterns and trends for example for human behavior. Three main characteristics of big data are Volume (more and more created), Velocity (real time), and Variety (sms, unstructured, xml, video, text etc.). These three V’s gives challenges such as capturing the data, structure and storage it, visualizing it, and at the end use it for business intelligence to improve or find new customer offerings. On top of this, data privacy needs to be taken very seriously, also increasing complexity.

And then the big question: How can you use big data to add value through an existing product or a new service? You must be relevant in your customer interactions, have an engaging product, and be trusted.

The good and the bad

In the market today, there are both good and bad examples of how big data and business intelligence goes hand in hand. In my opinion, a good example of how a customer offering is improved through big data and BI is how Google has applied data on traffic patterns and queues to optimize travel. To do this, Google couples machine-learning algorithms with rich and reliable data from multiple sources. Smartphones are a key source of this data, providing speed, drivers’ location and itineraries from phone-based GPS to Google Maps. Thanks to this, Google can determine the real-time traffic situation on a particular section of the route or notify of cars that, due to an accident, are moving 30 km slower than what is normal.

On the not so good side of using big data you have many online retailers. How many of you have been looking at a specific product, say a pair of shoes, and then suddenly your Facebook feed and other marketing channels are full of shoe commercials? And what happens when you have actually bought the shoes? The shoe commercials are still there for a long time. Another slightly annoying example is when my wife or kids use my computer for online shopping –the underlying algorithms do not understand that when they’re finished, it’s really no point targeting me with ads for female products or Justin Bieber promotions.

The business of combining big data and business intelligence has come far in some areas, but there is still a large improvement potential.

Mobility analytics

Coming from a large telco, many envy our access to data about our customers; which we do have a lot of. That could be from social media, web/content data, social data, demographic data, device profile data and more. But as you all know, having data is only the first step. Using it for analysis and decision making is a whole other thing that we are continuously working to improve.

In Norway, we have a project working with mobility analytics – in this context meaning analysis of movement patterns based on which base stations your cellular phone connects to. This data creates a range of new opportunities.

An outtake from a dashboard showing the movement over a day in the Norwegian capital of Oslo thanks to mobility analytics.

These data can also show attributes such as which nationalities are present at any given time. This information can be used to tailor travel experiences by adapting to tourist flows, or smarter city planning by getting information on where to locate hospitals, retailers, schools, parking lots and more.

Mobility analytics can also help us in predicting disease outbreaks and avoiding pandemics. In this area, many systems today depend on reporting from hospitals and doctors for tracking the spreading of disease – often with significant lag times.  With the help of keywords used in online search related to the disease and associating it with location, health experts can track the spread of a disease virtually in real time, allowing them to allocate limited resources more efficiently, and target communications down to the neighbourhood level.

As an example, in 2013 we analysed anonymised call data records from more than 30 million Telenor Pakistan subscribers during a dengue outbreak. The research team found that mobile phone-based mobility estimates accurately predicted the geographic spread and timing of epidemics in both recently epidemic and emerging locations. The study combined the data with dengue climate-suitability maps and estimates of seasonal dengue virus importation to generate fine-scale risk maps. The resulting model contributed to the design of more effective national response mechanisms in the country and other at-risk nations. This has also been researched on malaria in Myanmar with promising results.

The same concept of using data insights to reduce risk and improve the customer offering can be applied to many other industries, including health and car insurance (big data can be used to reduce the “uncertainty cost” added to premiums by assessing the disease prevalence and cost of potential treatment in a certain area or driving patterns.

Personalized & contextual marketing

In terms of personalisation, data from telecom and AdTech, search, payment, location and online behavior in general can be used to go away from segmented marketing and target each and every individual with a unique offer. We are experimenting with this now, and have promising results so far with one of our markets showing a more than 300% improvement on conversion rate by applying contextual marketing in the campaigns.

The way we do this is by using our data to create customer profiles that can help us predict which offer to give an individual customer. This is accomplished across different devices by using an identity service and data science together with reporting and analytics on attribution and media delivery. Based on the outcome the individual could be given a personalized offer in a relevant AdTech (social media, online classifieds etc.) or telco (sms, websites, applications) channel. In our case, the offers are either to join as a new customer, or give an up or cross-selling offer to existing customers. As we progress with our experiments the goal is to use more and more advanced machine learning and artificial intelligence to improve and optimize our personalisation of ads.

I can’t talk about any of this without mentioning Data Governance (data control and consumer privacy). We have a very high focus on this area, yes; regulations demand high privacy standards, but that is not our driving force. We believe strongly that our customers must be able to trust us as a company if they are to stay loyal and join us on our common journey going forward.

The examples I have mentioned provide a view of different customer needs and which new services that can be created based on what in many ways can be seen as the same big data set.

The competitive advantage is in the data

So how does that relate to the likes of entrepreneurs and investors? You need to understand and utilise the information you have available. The data, and more importantly the analysis will make your company more intelligent through getting a better understanding of those you are trying to serve. Start in the small, and work your way towards more advanced analysis and machine learning algorithms as you get more data through different sources. Start with a use case, then build analytical capabilities, deploy technology and solutions, not the other way around.

For Telenor, within digital marketing we can see which kinds of offers that are relevant for whom and through that improve our conversions. Within insurance we can offer a new and better product with lower risk and improve our customers’ life. Within mobility analytics we can do everything from assisting governments in health and transport planning, to tailor tourist offerings to people from a specific nation. I could go on with many more examples, and all of these are because we have started to understand the data we have available.

I would go as far as saying that companies not in control of their data and data analysis in the future will have a serious challenge to beat its competitors. On the other hand, if you are able to capture, structure and use all the data you have for business intelligence purposes, you will have a major competitive advantage.

Be relevant, engaging – and trusted

Let’s return to the abundance of data available to us. While you have been reading this, the following has happened:

  • >43 000 000 GB of internet traffic data created
  • >63 500 000 YouTube videos viewed
  • Almost 7 000 000 Tweets written
  • >2 300 000 000 emails sent
  • >28 000 smartphones sold (meaning we will have an even larger amount of data created the next 15 minutes)

Be relevant, be engaging and become trusted. If you can accomplish that, I believe that big data and business intelligence mean big opportunities.

Kirkeng on stage at Slush