Data Driven Development

In today’s world, companies of all sizes and in all industry segments are collecting an ever-increasing amount of data. In the last couple of years there has been a growing interest in how data can help redefine product development processes to the level of Data Driven Development (DDD). In short this means that instead of basing product development on expert opinions, development is directed by a feedback loop with the customer.

"In short this means that instead of basing product development on expert opinions, development is directed by a feedback loop with the customer."

Thus you need to ensure that data collection and analysis are part of the process, so that system and user responses can be measured. Your drivers are higher customer satisfaction and better understanding of the market and the merits of the product.

The main enablers for doing DDD are the following:

Iterative development, to develop and deploy frequently, so techniques like continuous development will increase your ability to experiment.

Data collection. The ability to collect relevant data in many different shapes and fashions:

• Qualitative data can be retrieved from users expressing their views about a product through e.g. user surveys, ratings or social media – or even video recordings of consumers using the product. Understandably, this type of data requires careful consideration in the analysis step, but is often vital.
• Quantitative data are often collected through automated systems, like code probes that track user actions, system response and diagnostics or third-party market data.

Data storage and validation. You need somewhere to put your data, and you want to secure the quality and validity of them, including proper handling of privacy and security.

Data analysis. Most crucially you want to derive insights and knowledge from your data. This is not only about statistical methods, but can also involve understanding user feedback and psychological and behavioral factors. Ideally you would like your development team to be able to do its own analysis, but expert support may be needed to get started.

With this in place, an organization has the tools to seek answers to more or less any question or validate any assumption it might have – and to take action accordingly. DDD can be taken one step further by reshaping the creative process into an experimental approach. This requires an ability to break down product ideas into iterations and experiments and to ask the right questions.

Your experimentation could be about: – New product ideas and how to hone these into something that fulfills real needs on the market. – Evolving existing products with new features by ensuring that each new thing you add makes sense to your customers and creates real end-user value. – Pruning (or removing) products by examining the actual use of the current feature set.

It’s not a coincidence that many of the big players in the tech industry are using DDD extensively. They have found that it is a key to ensure user acquisition, drive increased customer satisfaction, and establish a market and customer driven innovation system.

"ensure user acquisition, drive increased customer satisfaction, and establish a market and customer driven innovation system"

DDD will be one of the major changes in the years to come. As with e.g. Agile and Continuous Delivery, it also requires a transformation of culture. All parts of your organization need to agree to the methodology and work much closer together than they are (probably) used to, e.g. the product planners with the engineer, all having customer data and experimentation as a main vehicle for innovation.

Addalot did interviews with some 20 development managers in different domains to get a better understanding of where they are heading. In the area of DDD, the status in today’s software organizations can be summarized as:

• They collect (huge amounts of) data – but do only limited validation and analysis
• They make quick updates when needed – but do not build a continuous, cyclic (innovation) flow (including Continuous Delivery)
• They ask what the consumers think – but don’t experiment actively

This article is taken from the Addalot Report "Trends in Software Development"