This is the sixth article in a series from Accenture Applied Intelligence on Data Science Transformation. It focuses on how to ensure that data science can deliver the best value for an organization.The prior article in this series is Much More Than Toys – Why Organizations Need to Industrialize the Data Science Playground.
Many companies still struggle to integrate data analytics into everyday business decisions. Why? Often, their analytics teams lack the capabilities needed to deliver insights at the speed of business. Analytics Pod teams – which include experts across various skillsets - provide a viable solution to this challenge.
Combining interdisciplinary expertise with an agile approach, Pod teams can transform analytics delivery and ensure far greater responsiveness to business needs.
How Analytics Works – the Promise and the Reality
What difference does an Analytics Pod team make? Imagine the following scenario. A sales director in a snack food company contacts the data science team with an urgent request: “We’re losing sales in the convenience store channel. What can we do about it?”
The data science team puts together a Pod team consisting of a business analyst, data engineer, data scientist, visualization developer and platform architect. A governance team prioritizes the request for a two-week sprint, and within two days, the Pod team is ready to roll.
With the right blend of resources, the team delivers a cross-channel analysis of sales for the category, looking at competitive performance, marketing tactics, promotional activity, assortment and pricing. It proposes an assortment and pricing strategy that will drive sales in the convenience channel without cannibalizing other channels, all supported by meaningful insights and visualizations. It even integrates the outputs into a sales-automation tool, allowing sales representatives to provide targeted recommendations that drive significant volume growth.
Unfortunately, this is not how most organizations mobilize their analytics teams today.
The more common scenario is that the analytics team has just two members: a data scientist and a business analyst. Since the data they need is not readily available and they lack the resources to gather it, this team undertakes a basic analysis of sales and recommends adjustments to the portfolio. Unconvinced by a potentially incomplete analysis, the sales director decides not to implement their recommendations. Instead, to hit volume targets, the sales team increases promotional spend, driving volume back to the convenience store channel from another, more profitable channel (e.g. large format grocery). So, although volume remains constant for the year, profits decline.
Why was this approach unsuccessful? Although the team had business acumen and the right data science skills, it lacked the interdisciplinary expertise needed to deliver relevant solutions on time. Simultaneously, an absence of governance meant the team was unable to prioritize this request over existing projects. Finally, the team’s delivery approach was too rigid to accommodate changing requirements.
To overcome the obstacles that can prevent data science teams from delivering real business value, companies should take the following steps:
· Enforce interdisciplinary ways of working – Having a data scientist team with a business analyst is a good start. However, it takes more to move the needle. Delivering solutions effectively and on time requires additional skills, including data engineering, data visualization, solution architecture and project methodology.
· Align to the business – The solutions that Pod teams develop become more relevant as organizations assign dedicated teams to business functions. This will enhance and specialize their business acumen and problem-solving skills for specific use cases.
· Get agile – To implement agile delivery, Pod teams should deliver projects in short sprints, spread over two to three week periods. Sprints should focus on business capabilities or value rather than technical milestones. At the beginning of each sprint, teams should re-prioritize work, reserving capacity for high-priority immediate needs.
· Establish governance – The above requires governance, including a process for defining what is most important and most urgent, and what requires the most resources.
Successful organizations have learned to treat Pod teams as a venture capital fund. They recognize that one or two successes can pay for two or three failures. By taking these steps, data science teams will be empowered to deliver more responsive and relevant analytics solutions.
About the authors:
Robert Berkey is a managing director at Accenture Applied Intelligence, where he leads the Strategy & Transformation offering globally.
Dr. Amy Gershkoff is a data consultant; she was previously Chief Data Officer for companies including WPP, Data Alliance, Zynga, and Ancestry.com.
Brandon Joffs is a managing director at Accenture Applied Intelligence.