Companies generally exist to provide goods to customers and make a profit. However, there are a lot of different variables in how businesses can run their operations to be profitable. For new companies, the data generated by customers is just as valuable as any product they sell to them, if not more so. To make the most of each new customer opportunity in these changing markets, it’s important to make each transaction smarter.
Why change is necessary
Traditional businesses sell a product or service to meet a customer need, and make a profit from each one that is sold. New businesses set up to disrupt those sectors have changed how the customer solves their problem, creating data at every step that can be used to improve the user experience. The likes of Uber continuously take and use data to improve the service, from tracking where customers regularly hail cabs from to managing the supply and demand for drivers.
For these organisations, each action that a customer makes is a recommendation. For example, entertainment services like Spotify and Netflix use a customer’s playlist and interactions to generate recommendations for viewing or listening choices. These recommendations provide customers with more options, while the service provider keeps the customer happy and paying for the service.
At the same time, Internet of Things companies are springing up to connect everyday items as part of a service for customers. Thermostats and scales are boring, everyday items that people normally take for granted, but the data they collect can help companies improve their customer relationships. Rather than being invisible within the home, transactional data can help improve their day-to-day lives.
For traditional companies looking at these digital transformation projects being rolled out by competitors, this data-led model represents a wholesale shift in approach. To make this change, IT has to think in new ways about how data is used across the business. This starts with applying analytics to each transaction so that operations can be personalised. As each customer interacts with the company, data is created. Look at how that data can be used as the transaction is created in real time. Using this data is harder if you only look at historical trends and after-the-fact analysis.
Another important area to understand is why existing businesses shift their approaches over to new and uncertain models in the first place. In some cases, companies don’t have a choice. Existing businesses may see new market entrants taking big market share from them and have to respond. Even if this immediate competition is not taking place, it’s important to make critical changes before outside pressure comes to bear.
What is different around each transaction?
For CIOs thinking about how to support new ways of working within the business, it’s worth looking at how transactional and analytical approaches can work together. Normally, business intelligence and analytics projects are concerned with large scale historical trends that can be found through analysis of volumes of data. This is where Hadoop has been deployed to make sense of all this data that has been created over time. While this will still continue, companies are looking to do more with their operational data as it is being created.
Content from our partners
If you look at the history of databases, they can be divided into two camps – transactional or analytical. Transactional powers the application and analytical is used for business intelligence. There has always been a feedback loop between them. In the past, that loop used to take days if not weeks before results from analysis could be fed back into the online system. Today, that analysis happens in real-time to provide more relevant results for the customer in their recommendations and lead to more effective personalisation.
For example, this can include more accurate product recommendations that lead to a better user experience or further purchases. Alternatively, it can be used by the company to improve its internal processes by focusing on issues like preventing fraudulent transactions as they are attempted. Keep in mind that each transaction can be used immediately to affect how the company relates to its customers, both in market segments and as individuals.
Where will this data go?
All the data that customers create has to be stored and analysed. Rather than saving it to be analysed in large batches, data can be looked at in real time to get the most value. New tools like Apache Spark can conduct analytics faster, while the combination of open source platforms like Spark and Cassandra help scale up both operational transactions and real-time analytics together.
As more companies start to develop services based on connecting devices to the Internet, this use of data will also increase. The goal for many new Internet of Things services is to improve quality of life – for example, connecting up a set of scales to the Internet can be used to track weight over time and improve health through greater visibility and recommendations on activity levels. Similarly, using Internet-connected devices within the home can improve energy efficiency based on greater insight into what people are really using power for, rather than "best guess" estimates.
This use of data will come with its own considerations around data privacy and security. For example, is the data used solely for personal recommendations, or are there other use cases for that information? Insurance companies could find that data from Internet of Things devices on living styles and health would be very useful to their own calculations and risk analysis, even if the data was only available to them in aggregate and anonymised forms.
For customers, knowing how their data is being used is essential for them to buy into any long-term relationship with the companies they buy things from. if the vendor uses customer data without their consent, it will risk that relationship and trust.
At the same time, building up time-series data that covers the entire customer base and their actions will be a commitment for companies. While the cost of storage has dropped considerably, there is still a price to pay to manage the customer relationship around the data. It is only through having this time-series data that companies can find those new opportunities to improve customer service or deliver more value.
Personalisation and recommendation services require real-time analytics in order to be more effective. For companies that are looking at how to make use of data in new ways, working real time can open up new avenues for greater profitability and providing customers with more of what they want. However, making use of this data involves a change to how organisations think about their business models and how IT teams at these companies manage that use of data over time.
