Driving API Consumption with Demand-Driven Schema Design

GraphQL’s flexible type system allows us to model our data and how we access it in almost any way we wish. However, without a cohesive design strategy, a graph will most likely devolve into an API that is tailored too much for one type of graph stakeholder and have limited utility.

At Apollo, we’ve seen many companies run different GraphQL experiments and we’ve developed best practices around what we’ve seen work. We advocate for a demand-driven approach to schema design, where graph consumers begin the schema design process by proposing what they need and having other graph stakeholders help shape the final proposal.

In this talk, we discuss practical implementation suggestions and why applying the principles of a demand-driven approach will help you realize the vision you have for the graph, whether you are just starting a POC or you’ve been in production for years.

1. The importance of developing and adopting a unified vision for the graph

2. A demand-driven approach to schema design and how it can help you realize your vision for the graph

3. How to identify graph stakeholders and address their roles and needs

4. Practical advice for an iterative implementation of this approach for both new and mature graphs