Making effective use of customer and product data is highly challenging
RESEARCH. Software engineering companies collect large amounts of data from their customers and their products. But it is difficult to ensure this data permeates and influences the way the organisation develops its products. Resources are devoted unnecessarily to working with functions that ultimately fail to generate added value for customers and the company. This is demonstrated by PhD student Aleksander Fabijan in his licentiate thesis at Malmö University.
As a PhD student, Aleksander Fabijan is involved in a major research partnership – Software Center – in which ten companies and five Swedish universities have joined forces in an effort to accelerate and improve the adoption of novel approaches to software engineering.
“As part of my research, I am investigating how software engineering companies make use of customer feedback and product data to understand exactly what creates added value for the customers and to understand the type of functionality that we ought to focus on in the future,” said Aleksander Fabijan.
The way they gather data
Aleksander Fabijan has studied five companies, all of which are involved in developing embedded systems, including cars and mobile networks, and where a high proportion of the functionality is in the form of software. Through interviews, observations, workshops and prototype development, he has worked closely with these companies to understand the way in which they currently gather data from customers and products – and the way they use the data to improve their products. Based on this knowledge, he has been able to create models and processes that support companies in their in-house improvement programmes.
Diffucult to use
The study shows that these leading software engineering companies have access to large amounts of data, not only from customer surveys, interviews and beta testing of products, but also from the products themselves, which generate information each time they are used. Aleksander Fabijan has shown that companies are finding it difficult to use this data optimally. If they had been able to do so, they could have improved their products more quickly than is the case at present. This applies in particular to qualitative data – data received directly from the customers – where it is difficult to generalise, acquire an overview and apply the data effectively.
“It’s all very ad hoc; very here and now. The purpose perhaps is to gather data for a certain local project at the company. But that’s where it remains.”
Four different types of functions
Aleksander Fabijan talks about four different types of functions in software products: functions that need to exist (such as a safety belt in a car); functions that are necessary because a competitor has something similar; functions he calls “wow functions”, i.e. bold, eye-catching functions that are perhaps not always widely used but which draw attention to the product and promote sales. But the most crucial type of function is the fourth type – the type of function that improves customer experience when they perform a specific task or activity on a regular basis and which is part of their work or day-to-day life.
Aleksander Fabijan explained: “It’s a function that it is believed the customer wants to use a great deal and which competitors do not have.”
His conclusion is that the difficulty of making effective use of accumulated data is that engineers and technicians spend time developing functions that do not generate this distinctive, image-enhancing know-how.
Aleksander Fabijan has produced a framework that will facilitate qualitative analysis. “Using the framework, data can be analysed and it will be easier to see how to prioritize in a development project and where to invest,” Aleksan