The Product Analyst as a new role for agile teams

We meet a lot of managers and executives who have transformed their software and hardware engineering into an agile manner. The degree of adoption and success is surely varying with peoples’ passion and budgets. The next wave of evolution is Artificial Intelligence (AI), particularly Data Science. With citizen blockchain already doing the paperwork for the next initiative – I believe AI to be an extension to engineering paving the way for intelligent software, as Tim O’Reilly stated. Consequently, I suggest herein a new role for agile teams: The Product Analyst, which is a data scientist embedded into multifunctional software teams.

Why a new role?

Riding the new wave of adoption, companies have started hiring and building their own data science teams. We meet a lot of talented people from different branches of science in those teams. They eagerly dive into the data of their company, likely equipped with priorly defined big data architectures. Smart and eager to model the whole company. With euphoria backed support of manifold different departments handing in requests.

Nevertheless, a repetitive moment is uniting a lot of those AI/Big Data/Smart Data/Large Scale initiatives: After initial requests, departments suddenly realize, that knowledge transfer for their business processes is time consuming and results do not excel expectations.  Thus, Data Science is certainly no magic blackbox; Additionally, data scientists try to advertise internally to propagate more experiments and collaborations. If not given, those get demotivated.

Learning from agile development

Adding the product analyst to an agile development team leads to more insights and better products.

The product analyst embedded into an agile development team. More insights. Better products.

Learning from agile development, we already had understood the necessity of a product owner being in close contact to business requirements and stakeholders. Being the product owner is a full-time position. Consequently, in most projects, the product owner needs support for doing complex analysis across customer interactions and past data. Therefore, I suggest adding the role of Product Analyst to the orchestra of agile teams. This role can avoid siloed organizations according to Conway’s Law, away from analytics departments into embedded data science.

Qualifying the product analyst

The Product Analyst should be responsible for all product related analysis, utilizing customer and product data. Furthermore he delivers insights into existing customer base and forthcoming features. The Product Analyst is part of both hardware and software engineering teams. The essential level of education is formulated by the product complexity and the amount of data present. While an AI driven smart home application requires a data scientist, fully qualified in natural sciences, a rock solid university degree and a passion for number-crunching should be sufficient for a KPI driven business application. No doubt, the requirements for the hardware and software engineers will vary accordingly.

Embedding product analysts in agile development teams

The embedded Product Analyst is superior to analytical departments, since the role can be laid out in agreement with product requirements. According to the complexity of the product, the Product Analyst coaches the team in using quantitative methods ranging up to artificial intelligence and deep learning. He remains part of the engineering team. For which the product analyst should work in close collaboration with the product owner to find the best and pragmatic algorithms to fulfill the business requirements. However, the whole team will implement the algorithms and finally ship the AI models. They build it, they ship it.

How to get started?

So how do you get started with product engineering and AI? During consulting, we develop a data strategy with the client and identify the most promising fields for monetization of intelligent software. During implementation phase, we either train the agile teams to integrate data science into their methods or help adopting with our own data scientists.