Member-only story
The ROI of design collaboration
A quantitative approach to design ROI that puts collaboration at the center
“What’s the ROI on that?”
Trying to measure the ROI of design has become a joke. In a hamstrung system, what exactly are you measuring? It’s like first anchoring an air balloon to the ground and then measuring how high it can go. Collaboration is essential for design and UX to deliver value, but siloes are often the norm.
I started the no handoff method to work more collaboratively. This article adds a way to measure the ROI of this transition.
Measuring a new system can seem hard because there is no track record, but it’s actually a great opportunity for multivariate testing. Just as with quantitative measurements of UX, we must isolate an effect to measure it. To do that effectively and in the language of management we can use Variance Reporting.
What is variance reporting
A variance is simply the difference between an actual result and the standard you are comparing it to. A variance report is a collection of these differences. They are used across many industries and help managers quickly see where their attention is needed. Typically they are large documents produced at regular intervals (even daily) that track how multiple line items are deviating from the budget. It’s not just for budgets though. If you can measure it then it can be tracked in a variance report.
To use variance reporting to measure the ROI of collaborative design we need to choose a metric that can both be tracked and is a good proxy of the desired outcome. Labor hours are the primary cost in software development, and though we know that “all models are flawed”[Box, 1979] a reduction in labor hours is a reasonable proxy for ROI.
By producing a variance report on the average labor hours of a standard project compared to the labor hours of a No Handoff project, we can provide a quantitative measure of the ROI of collaborative design as compared to more siloed projects.
Setting up tracking
Variance reporting is actually very simple, the difficulty is in gathering the data you need consistently. Here are the steps.
