The Key to Data Product Success: Management Principles Are as Important as Analytical Assets

Editor's note:Julie Iskow isWorkiva's Chief Operating Officer, and Jon DiGiambattista isALM Media's President of Information Services. The views expressed in this article are solely those of the authors.

As financial leaders, CFOs know that leveraging data can improve efficiency, productivity, and turn insights about the future into action. They also understand that data investments can drive innovation and create greater value for customers. However, they are equally aware that identifying opportunities for innovation is one thing, but turning them into tangible returns is another.

Simply having assets, tools, and a team of data scientists is not enough to guarantee success. Countless enterprises have invested heavily in data capabilities without achieving the expected returns. The challenge lies in the ease with which people can be distracted by the dazzling features of data exploration, losing sight of the fundamental goal—creating value for customers and end users.

Our experience in practice shows that the challenges CFOs and their colleagues face when developing data products and the supporting infrastructure can be overcome if leaders consider the following factors.

Creating Business Value

Data science innovation is transformative, but it is not the only most important component when developing data products. You may have a powerful analytics platform and still miss the mark. Creating meaningful data products means avoiding common pitfalls that hinder integration, adoption, and successful user experiences.

Successful companies know the following key points:

• When defining the product-market problem to solve, adhere to core product management principles rather than being led by data science principles.
• Look for problems where solutions can bring value, and do not mine data just for the sake of mining. Validate value through proof of concept (POC), minimum viable product (MVP), and pilots, and iterate repeatedly with users, customers, and partners to confirm you are achieving the desired outcomes.

At Medidata, the data analytics platform we have both worked on, we found that progress was made once we began to view the expected outcomes as goals to solve real and significant business problems for customers. This became our guiding principle. To this end, we worked with customers to define, design, and refine products and features. We invited feedback early and often, conducted real-world pilots, and iterated continuously while incorporating customer feedback throughout the process. This is the foundation of success.

Actual Adoption and Use

Everything starts with the customer. Being customer-centric means asking the right questions to understand the potential impact your product may have on the customer's entire ecosystem.

For example, do your customers have the capability and willingness to integrate data solutions into their operations? What are their needs regarding account privacy, security, performance, and customer support when scaling?

Understanding the answers to these questions will help enterprises adopt data products and ensure the products meet customer standards. Here are some practical suggestions:

• Consider ease of use, the effort required for adoption, and the change management needed for usage.
• Understand whether customers prefer self-service discovery (curated data, tools, and insights) or full-service delivery (recommendations, suggestions, and conclusions).
• Meet customers where they are currently and help them transition to the desired future state. This may include training, providing services, or engaging third-party partners to drive change management.

Through collaboration, we have realized that focusing on the customer experience is crucial. Solutions need to be tailored, simplified, and intuitive, allowing users to quickly and easily adopt and gain value—this is a key goal for us. This is also often why some companies adopt a heavy service model in the early stages of analytics product development; they want to accelerate customer adoption and ensure continuous feedback.

Collaboration and Execution Teams

Building the right culture will help ensure you fully leverage the expertise at hand. An open and successful mindset among everyone on the team will lead to higher productivity. To this end, it is important to—

• Acknowledge that building, delivering, and productizing innovation, as well as scaling data products, requires far more than data scientists (though they are crucial). It also requires data security/privacy, delivery, data systems engineering, enterprise and data architecture, and product management (and engineering if integrated with the core platform).
• Prevent silos by understanding that data scientists are ultimately part of a larger value chain that delivers products to the market, and they should be integrated with product, engineering, and the broader technology organization.
• Prioritize fostering a culture of respect, trust, transparency, and collaboration. At Medidata, when we aligned the data science team with the broader organization, sharing outcomes and goals, we immediately reaped benefits. Bringing functional teams together allowed us to fully leverage each team's strengths, learn from one another, and achieve greater, more innovative results. This model led to higher-quality, more relevant capabilities, faster time to market, and happier, more engaged employees.

Building for Scale

The pitfalls of building for scale are numerous, but they are all avoidable. If enterprises take the time to build a framework aligned with their expected outcomes, they are more likely to achieve sustainable growth.

• Ensure automated operations are in place to make data 'fit for purpose' to achieve the expected outcomes; establish frameworks for connecting new data sources and for continuous data cleaning and curation.
• Align enterprise platform architecture and infrastructure to support the innovation lifecycle from research to production, providing a secure, controlled, integrated environment that supports experimentation and large-scale delivery.
• When work moves from ideation to productization, implement an effective software development lifecycle (SDLC) so that engineering and product follow a set of quality control standards in code, data, and implementation.
• Implement rigorous accountability systems designed with cultural values of respect, trust, transparency, and collaboration, and test them frequently.

At Medidata, we developed a process from innovation to product that leverages and becomes part of the core software development lifecycle process. We ensure that new products are compatible with the company's overall architecture and ecosystem; they are well-governed, controlled, maintainable, and secure (but still easily accessible); and they empower high-velocity exploration, discovery, and innovation.

Key Takeaways

These key factors stem from our years of experience in practice, where we have seen what works and what does not. Enterprises that learn from the successes and failures of others will be better positioned to seize the data opportunities at hand.

Authors Iskow and DiGiambattista previously worked together at Medidata Solutions, serving as Chief Technology Officer and Head of Data and Analytics, respectively.