You have to love how when some people speak, everyone stops and listens. A shareholder letter from Jeff Bezos is making the rounds and I love how each of us reads the letter and internalizes it differently. We all find ways to relate our own world to different bits of the shared wisdom.
Myself, there were two ideas that have been spinning in the back of my head lately that were put together in a really great way by Jeff Bezos.
When user testing isn't the whole story
This is one that I've been working so hard to process. We're all focused on measuring and data and so we work hard to include user testing in the design process. It's 100% correct to measure and work to understand what actually happens out there in the real world.
Everything can't be a simple a/b test. I have seen and fought getting too carried away with doing what the testing says. It's so easy to let go of responsibility and respond "the user testing says ...". I think that user testing is great when you're looking for small tweaks and changes or general feedback on if something appears useful and interesting. If I want to improve the click through rate, rearranging the content, the design of call to action buttons, or even the information architecture all might help. NONE of these though are breakthrough ideas that provide something new and excited for users. They're not truly things that are part of the identity of a product.
I often find that in Juju users will nearly always provide feedback that's the shortest path from their task to having it work the way they want. It makes sense from where they're sitting. A user might request some fast path to making a quick change after a deployment in a one off way. Juju is built on a consistent model. Anything that's done that's not part of that model, the state, is lost for future decisions and understanding. We often find we need to take the user feedback and get into what they want to do in this "one off" and see where we have an opportunity to improve the model so that the user is able to leverage Juju but the model is still consistently the focus on the product. I feel like the key word I want to use here is intent. You can filter ideas, test results, and then put some true intent behind them to build something of value.
Getting others to agree by allowing them to disagree
This is a lesson I first learned myself in code reviews. I'd see some code and completely hate it. However, if I stood back, I'd realize what I hated was that it's not "how I would have done it". I would have to watch myself and not be negative just because someone else thought differently.
I like how Bezos's letter makes disagreement a reason to get together. "but will you gamble with me?" is a really great way of putting that. I find that normally this is a lesson for myself. To make myself willing to gamble on someone else's work or ideas. I obviously am already sold when it's my work.
Sometimes, in order to get folks on board, they just need permission to not be held accountable if it fails. In code review, your goal isn't to be perfect all the time. In decision making, some will be good and some will not work out. However, you need permission to let the team move forward on something and learn from the outcome. None of us will be right all the time.
What do you find interesting or personally meaningful in the letter? Do any of the ideas really speak to something you've been noodling on in the back of your mind. Let me know. @mitechie