Personas and Test Data

I was reviewing the excellent work Unicon’s Gary Thompson has done in developing personas related to uPortal (really they seem applicable across Higher-Education), which jogged my to write about some thinking: leveraging well-developed UE personas to build a suite of synthetic test-data.

User Experience is often not fully integrated within other aspects of the software development lifecycle, due to user experience’s ambiguous relationship with the “hard-types”. One aspect that stands out however is the similarity between UE personas and good sets of synthetic test data. Both seek to identify, quantify, and provide a means for verifying functionality against common scenarios and edge cases. Also, both are a lot of work to do well.

In work within the context of the uPortal project — I’ve begun to wonder whether involving the UE personas in the generation of seeding our test data might have some nice synergies.


  • Focus developers on UE personas during testing
  • Provide iteritive development & refinement of personas and test data
  • Provide common language for uniting UE tasks & developer edge-cases
  • More tightly integrate functional and UAT testing scenarios

In the case of uPortal, very simple test data has always shipped with the project, generic accounts for “student” “admin” “faculty” and the like. How much richer would our demonstration and design capabilities be if we actually had an accounts for Owen Oldschool which developers could test against that defaulted to larger fonts, and had a UAT script that tested the contextual help?

In open-source in particular, where there’s been a historical disconnect between the design/UE community, it seems like trying to integrate out tools, processes, and language could show great benefits.