We decided to ship a new secondary navigation pattern in one part of the product over honest dissent about long-term inconsistency. The reasoning, the data, and how the dissent shaped the rollout.
One of our design leaders brought a secondary navigation concept to our bi-weekly extended design leadership sync this week, aimed at a new experience we’re launching. The current state is tabs with sub-tabs with filters, and it’s getting unwieldy. His proposal moves the top tabs into a vertical left column, promotes the sub-tabs, and clears up what he called beachfront real estate in the hero section. Engineering sized the core change as a relatively light lift.
The concern in the room was legitimate, and another of our trusted senior designers voiced it best. If one area ships this and nothing else gets it for 6 to 12 months, we’ve introduced long-term inconsistency into the product, on top of brand work already in flight. He also pushed on whether the change was truly crucial for the launch or just better. And he was right to push. Some of our oldest pages carry legacy layouts that would fight any new container, so a platform-wide rollout is nobody’s idea of fast.
My call: ship it and test it in the one space, and don’t let it block the launch. Two reasons. First, adoption data from an adjacent product shows heavy churn in week 1 that eases as familiarity grows, so simpler early navigation is likely a real retention lever, not a cosmetic preference. Second, a pattern proven in one high-traffic space gives the rest of the design leads something concrete to govern and scale. The alternative is a platform-wide debate about a hypothetical, and I’ve watched those debates outlive the products they were about.
Then one of my design leadership partners did something that changed the shape of the whole conversation. He decomposed the proposal into 2 separable changes. Step 1: collapse the main nav, doable today, no structural change. Step 2: add the secondary panel, the bigger and more complex lift. What had been an all-or-nothing debate became a sequencing conversation, and sequencing conversations actually end.
What had been an all-or-nothing debate became a sequencing conversation, and sequencing conversations actually end.
I told our dissenting lead in the room that I appreciated every one of his gut checks, and I meant it. The dissent didn’t reverse the call, but it rewrote the plan. The next step is a level-of-effort pass with engineering across every area of the product, and the responsive impact he flagged is now the first thing we check. Deciding over dissent only works if the dissent leaves fingerprints on the execution.
The longer game here is governance. Ownership of the pattern sits with the teams shipping it now, and the design leads collectively own the question of global scale. Whatever we end up scaling, the goal stays the same: navigation should reflect the jobs to be done and the workflows, not the org chart. If a nav structure only makes sense when you know how the company is organized, it was designed for us, not for the people using it.
The teachable part
When a pattern debate stalls on platform-wide consistency, ship it in one high-traffic space, let adoption data carry the argument, and let dissent shape the rollout plan instead of blocking the decision.