Stairway to Heaven

These days, successful software development requires involving the clients in the design. However, this means that the management of customer feedback already has to be in order. Stairway to Heaven is a model which was created at the Universities of Malmö and Chalmers for industrial co-operation. Through the N4S programme, a similar, sufficiently long-term model will be created for the use of companies in Finland as well.

In Sweden, companies have carried out long-term co-operation with universities in program design. Software Center is a joint research organisation of Swedish research institutes and companies. Involved in it are the Chalmers University of Technology, the University of Gothenburg and Malmö University, as well as Ericsson, Volvo Cars, Volvo, Saab Group, Axis Communications, Jeppesen, Grundfos and Sony Mobile AB. The starting point of the operations of the centre is a continuous exchange of information between the companies and universities. The companies and universities decide together which projects to start out on.

The University of Oulu has carried out comparable co-operation with Nokia since as far back as 2004. The researchers of the University have participated in many European ITEA research projects with Nokia. ITEA is a cluster programme which supports program-intensive research and development projects.

Lucy Ellen Lwakatare on Stairway to Heaven model

“The co-operation carried out in Sweden and Oulu is basically about the same thing. With research co-operation carried out with companies, the aim is to make the whole of product development more agile. At the same time, research data that has been validated in practice is produced, and this can be utilised by other companies as well in their own development work. Companies derive a financial benefit from being able to react more quickly to changes occurring in the market and the new demands presented by customers relating to the products of the companies,” says researcher Pasi Kuvaja from the University of Oulu.

Photo: Kimmo Lehtonen

According to Kuvaja, for example in on-going cooperation with Nokia the target is to reduce Nokia’s response time to the new requirements of online customers from months to a days. So, the increase in productivity can be seen on this level and not only in how many lines of code are done in some period of time.

Five levels of software development

The Stairway to Heaven model describes moving from traditional software development to continuous adoption into use. The research results confirm that organisations do not adopt agile methods without careful presentation and familiarisation. An agile working culture is also not born is large product development teams, but instead they have to be made smaller. The starting point is that the team is responsible for the whole service or product. The development work also has to be focused on the features of the program, not its components. In the Stairway to Heaven model there are five stages through which a company must go through to get to its objective, i.e. quick software development that reacts to customer feedback.

Stairway to Heaven model

“Although the STW model was invented in Sweden, we’re further down the path in Finland than in Sweden,” claims researcher Pasi Kuvaja.

According to Kuvaja, on the basis of research results Swedish companies are at the integration stage whereas in Finland they have reached continuous deployment. In the N4S programme, web technology companies in particular are already trying out the benefits of stage 5.

In a way, the N4S programme has continued on from where the four-year Cloud Software programme (2010-2014) ended. According to Kuvaja, the companies that took part in the Cloud Software programme in Finland moved to stage 2 (agility) and began to take on the third stage, i.e. continuous integration. In the N4S project, there has been another step further still.

“Now, in the N4S programme those who are not yet at stage 3 have moved to it. Those that were already at that stage have moved to stage 4. There is now a terrific phase of transformation under way and companies such as F-Secure and Bittium have already attained the last stage. N4S provides a lot of support to these companies.”

The goal of the N4S programme is to get the fifth stage to all the software companies of the N4S programme.

“It’s important that software updates are made up to date. When web-based applications are updated all the time, it’s possible to test the customers’ reactions to the new software features at once.”

Data collection fundamental for understanding customers

The greatest challenges in companies are to do with a rigidity in the working culture. The researchers of the University of Oulu have found four problems that should be dealt with. The first is the lack of a direct connection with end users and the deficient understanding of the demands of customers that arises from that. The second problem is that the choice of the features of the software packages is based on the opinions of the workers and these have not been verified through the users. Thirdly, testing is seen as a way of identifying faults, but the gulfs that exist between product offers and customer needs are not noticed. Fourthly, companies have no systematic way to collect, analyse and incorporate the data obtained from customers into the product development process.

Pasi Kuvaja emphasises the importance of data collection and understanding the feedback received from it. In many organisations, data collection is extensive but the information is only accrued by a part of the organisation. Data collection should change so that all the employees of the organisation would have a uniform view of the feedback received and of how it should be reacted to. Operational data is also a gauge of the level of software development.

“In this way one can realise applications that are not crammed with thousands of features of which only a few percent are actually used. So, unnecessary work is not done and it’s possible to invest in the features that interest users instead.”

As an example, Kuvaja mentions a company with a huge number of usage possibilities in its software, of which, however, only a few are used in practice.

“The upkeep of software like this is a massive drain on resources,” sighs Kuvaja.

“It’s important that software updates are made up to date. When web-based applications are updated all the time, it’s possible to test the customers’ reactions to the new software features at once.”

The Internet of things will change data collection

Data is no longer collected by sitting in front of a terminal. In the Internet of things, data is collected directly into databases from the use of objects.

“A drill, for example, is connected to the Internet, meaning that feedback information is received through being collected by machines. In this way, the software can be fixed quickly.”

In addition to external data, data on your own development work must be collected to speed up your own work. One area is automatic testing, which generates internal big data for companies which can be used to develop their own work. In this way, decision-making is improved and it can be dispersed.

When a request for a particular feature comes from a customer, it usually takes a long time before that feature can be delivered. There is a desire to reduce this time and success in that will be achieved through internal data collection. In a way, data and its analysis determine the direction and decision-making.

Successful data collection is based on two principles. Firstly, the features that the customers really want are made quickly for systems. Secondly, features are created in application areas that can be broadened and grown.

“It’s not possible to reach this kind of situation if one is not ready to change rapidly.”

In the opinion of Kuvaja, the Stairway to Heaven model provides one possible way to go forward.

“STW has already been validated in practice, so the guidelines it provides can be relied upon.”

Kuvaja believes that it’s important to establish how far the doers and deciders are from each other in companies. For that reason, an operational development model for decision-making (Development of Operations) has been designed at the University of Oulu.

“The aim is to make organisations more agile. In addition to working practices, the degree of transparency must change. The data collected must be made visible to everyone and it must be possible to support decision-making with the right data. Data collection must be tested in order to know that unnecessary data is not collected. This affects strategic decision-making.”

Ari Turunen
August 17, 2015