← Back to Agenda

Truth: no user wants another technically right, but effectively wrong solution from your data team. It’s not good for you, your team, or your customers—but it happens frequently because solutions are not designed with intention around adoption, usability, and utility. In this session, I want to share a human-centered approach to creating useful decision support applications that allows the value of your ML and analytics work to be realized. This approach is often referred to as having “a data product mindset” or “product orientation,” and I’m going to explain why thinking about data products as a method and not an output will get your team producing better ML and analytics solutions that actually get adopted.    During this non-technical session, we will use my simple, human-centered definition of “data products” that is practical, useful, and decoupled from any output-based definitions heard in spaces such as data mesh, analytics products, and data-as-a-product. I will also share the messy truth of innovation work you must get comfortable with, why UX design and product management skills are essential to working in this style, and three design activities your team must embrace to routinely produce useful, high-value data products. Additionally, we’ll cover:

• Why 1 leader tossed out 18 months of data science work to adopt this approach
• Why you can’t achieve business value with data without first solving the adoption problem
• Why change management is the wrong way to think about introducing ML/AI and decision support apps
• Why customer and stakeholder requirements are friendly lies you can’t trust
• What my performance with The Who has to do with data products
• The producty definition of data products that myself, and most members of The Data Product Leadership Community, gravitate toward – and why others aren’t helping them
• Core practices your team can (and must) adopt now if they want to routinely deliver data products that get used and produce value
• What the future looks like when you start adopting a producty approach
• What types of data products and usage scenarios need UI/UX the most and why GenAI poses even greater challenges around adoption
• Where to find data product management and design resources
• Why data products must be measurable, and how to measure their impact (yes, they can be measured!)    15 additional ways CDOs can increase adoption of data products (hint: they don’t involve GenAI, Python, or any other tech)

Speaker

Brian T. O’Neill

Founder and Principal, Designing for Analytics