Half in
Why design can’t use the long grass.
Any design decision that can be tested will eventually stop being owned by design. Contrast ratios for accessibility get a check in the build. Screens get compared against the last release, pixel by pixel. The component library gets rules that fail a pull request when someone breaks them. Each of these was once a designer’s judgement, argued over in a review. Now it passes or fails, and whatever passes or fails belongs to engineering.
These decisions only move one way, and the reason for that is the kind of question being asked. “Does this work” ends in either a pass or a fail. The answer to “is this what people want” emerges slowly (if at all), noisily and after the feature hits the reality of customers. Definition of done is written in engineering’s terms because engineering’s questions have answers of that kind, but it means design then inherits a stopping rule built for a different kind of question. What’s left in design’s house once these testable parts have moved out are often those that are hardest to close and finish. The extreme end of this are questions of taste, which is one claim nobody can objectively disprove.
Nothing to show was about the rare organisation that is able to hold an idea open until it’s right. This post is about everywhere else, where design sits half inside engineering’s process holding the work that can’t be tested. I don’t think this situation can be fixed from inside design, but it can at least be described, which is more than usually happens.
Engineering has its own unresolved questions, and somewhere to put them: spikes, hardening sprints, investigation tickets, a tech-debt allowance (often branded as ‘maintenance’). Nobody outside engineering ever asks what the ROI of the refactor was. That isn’t dishonesty though, it’s just that 40 engineers and a few hundred tickets dilute the uncertainty to background noise, and so it goes quietly into the long grass. Four designers carrying whatever isn’t easily testable are a third of the function, and their uncertainty shows up on every board. The long grass simply isn’t available at design’s typical headcount.
So design ends up half in. It’s in the ceremonies but never estimated. It has tickets on the board but they don’t close cleanly. It works ahead of the sprint but in a way nobody can see. That produces one complaint from two directions. Engineering says design is never done when they need it. Management says it doesn’t see what design contributes. Both mean the same thing: we don’t know what you’re doing. Dual-track agile was meant to solve this by giving the long grass a serious name, but the discovery track (if it’s working properly) often looks like it’s not making progress or is giving stakeholders ugly truths, so it’s the first thing cut.
Being half in is rational if you look at it week by week. In any one sprint it’s the least-commitment option and nobody would argue with it. The cost arrives at the end of the half, when someone asks what design did. You’ll all have your own version of this. The designer who worked flat out all quarter keeping up with the team, and at review has a handful of their own small wins to show. The design leader asked by someone on the executive team what the design team actually contributes. It’s the same motif: the heavy lifting was done inside features that engineering shipped, and so it gets credited to the feature. What design can call its own happened outside the process, which makes it small by definition. Nobody asks what engineering contributes, because its output is the product. Design’s contribution is inside that output, and from outside it looks like part of it.
I’ve seen the consequences of this stance happen more than once. The person read as phoning it in is sometimes doing the most important work on the team; the organisation priced the deficit it could see and never saw the surplus1At end of year review these people will be told that they should have made the further effort to make their work more visible, adding injury to insult.. The organisation’s view of someone can reverse completely with nothing changed except what’s made visible to them: same person, same hours. And I’ve said the same sentence to more than one designer: make that idea you’re carrying in your head the thing you’re known for internally. I don’t offer it as a method, it’s what I found myself saying a few times.
The judgement is knowing when you have to work inside engineering’s process and when you can afford to step outside it and spend past 75%. Some people make that call well, although I haven’t found a way to teach it, but if you’re a design leader it’s your job to pull these people under your protection and it’s your job to make their work more visible while they carry the cost of untestable questions.
At end of year review these people will be told that they should have made the further effort to make their work more visible, adding injury to insult. ↩︎