Articles
Becoming more opinionated about your product
An opinionated product makes the key choices for people and makes them well. Three before and after examples of moving from universal to opinionated.
Becoming more opinionated about your product means deciding how it should work for the people using it, instead of leaving every choice open. I spent years building something universal, where anyone could make anything, and I am now choosing a clear, specific shape, and becoming comfortable with that. Below are three real examples from my own product, each shown as before and after.
The short answer
An opinionated product makes the important choices for you and makes them well. A universal product gives you every possible option and leaves the choosing to you. Both can be good. But when people are busy and stuck, the opinionated version usually gets them to the result faster, because it starts them on the path that works.
On a walk this week I noticed it in myself: "I'm getting much more opinionated with like how I want to do things, and I'm becoming more okay with that."
Universal versus opinionated
For a long time my instinct was to build the most general thing possible. A system that knows nothing about any one person, that treats every use the same, and that anyone can bend to any job. There is real value in that. It keeps the code clean and it means nothing has to be rebuilt when needs change.
The cost is that a general tool asks the person to do all the deciding. They open it and face a blank page. The better the underlying system, the more options there are, and the harder it becomes to know where to start.
An opinionated product does the opposite at the surface. The engine underneath can stay general. What changes is what the person sees first and what happens by default.
Example one: what you see first
Before: the product opened as a studio. You could build any tool or workflow you could imagine, which meant you had to know what to build.
After: "I want the application to be about the pursuit." The app opens on what you are pursuing, and it shows you one thing above everything else: "this is your absolute biggest needle mover." In my words, "So this is much tighter of a vision. It's more opinionated of how the app should be."
The general engine is still there. It is simply no longer the front door.
Example two: how every decision starts
Before: any process was allowed, so the order of thinking was left to chance.
After: every decision starts the same way I start them when I work with people in person. "Here's my philosophy mode. This is what I think everything should start with. Like, what's worth the pursuit?" Only once that is clear do we look for the single biggest move, and then act on it.
This is a strong opinion. Someone could argue for a different starting point. But a clear first question saves far more time than it costs, because people stop spending energy on the wrong things.
I learned this order on a tennis court long before I put it in software. When a player comes to me, I do not hand them twenty drills and let them pick. I find the one weakness that is costing them the most points, and we work on that first. In my experience, players improve faster with one clear focus than with a long menu of everything they could practise, because their attention is not split across ten things at once. The product now works the same way.
Example three: how much the AI takes the lead
Before: the AI waited to be asked. It did what you told it and nothing more.
After: "I want the AI to have a little bit more agency." It looks through what you save, such as screenshots and notes, notices what could be useful, and acts on it. That is an opinion about how the product should behave. It is still configurable, and a person can set their own rules. "But everybody's getting this automatically to start, which is great."
Each of these three changes is small on its own. Together they change what it feels like to open the product. Instead of asking, what could I build here, a person sees, here is what you are pursuing, here is the biggest move, and here is what has already been started for you. That is the whole difference between universal and opinionated, felt in the first ten seconds.
Why opinions make building faster
There is a practical reason this matters to me. A tighter vision is easier to build. As I noticed, "it is making it easier for me to have it developed by the AI." When the target is clear, each new feature has an obvious place and an obvious job.
A universal design pushes the other way. In a fully general system, "you probably wouldn't implement that specific feature for all users to have." So useful features never get built, because they do not fit everyone. An opinionated product can include the feature most people need, while the general engine underneath still lets anyone change it.
A well-known example is Apple when Steve Jobs returned. The company had many overlapping products, and he cut the line down to a small number of clear ones. Fewer choices, each made well, made the products easier to understand and easier to build.
Being opinionated is not the same as being rigid. The opinions live in the defaults, the order of steps and what appears first. The engine underneath still treats everything as data that can change. If an opinion turns out wrong, changing it is a matter of changing a default, not rebuilding the product.
How to become more opinionated without boxing yourself in
- Write the one outcome your product exists to deliver, in a sentence a stranger would understand.
- Decide what a person should see first, and remove everything that competes with it.
- Pick the default path that works for most people, and make it the starting point.
- Keep the engine underneath general, so the defaults are settings rather than walls.
- Let people change the defaults for themselves, but do not ask them to choose before they start.
- Watch where people still get stuck, and make a clearer choice there too.
The short version
- A universal product offers every option; an opinionated one makes the key choices well.
- Busy people usually get further with a clear default than with a blank page.
- Keep the engine general and put the opinions in the defaults.
- A tighter vision is faster to build, for people and for AI.
- Let people change the defaults, but never make them choose before they begin.