The hardest screen we have ever built is an empty field with a blinking cursor. There is no list to pick from, no levels, no tiles with pictures of subjects. There is only a question, what do you want to be able to do, and room to answer it.
People pause in front of that field. We have seen it at our own desks, with our own product. So this post is about what happens to your text in the next two minutes, and why each part is there.
Your words stay your words
If you type "I want to move into development", that sentence is kept as it is. We do not match it against a catalogue and quietly turn it into "Programming, beginner".
This matters more than it seems. "Move into development" and "understand the code my team writes" would end up on the same catalogue tile, but they are different goals. The first is reached with a first job. The second is reached the first time a pull request stops being frightening. The routes to them go different ways within a fortnight.
So the goal stays on the screen in the words you used, and you can rewrite it later. It is the one thing in the product that you word yourself.
You can also write more than one goal. Two is common, usually a profession and the language that profession uses. Both are on the same screen and neither blocks the other. The product asks about one goal at a time, and an evening is still one step long.
Two questions, not a form
Then EvoluteMe asks you two questions. It is not a registration form and not a survey with a progress bar across the top. Just two questions.
Which two depends on the goal. For a new profession: what you have already done that is related to it, and how much time an ordinary evening really gives you. For an exam: what the exam covers and how much time is left before it. For a child: the year at school and the most recent thing that went wrong.
We chose two deliberately, and we got there by removing questions. Every extra question looked reasonable on its own, and every one cost the same thing: one more moment when a person who came to learn is filling in a form about themselves. Anything we can find out later from what you can do, we find out later.
There is one more rule. Your two answers do not have to be accurate. "Something with computers, ages ago" is a perfectly good answer, because these answers do not set your starting point.
What comes back is a guess
A few seconds later you see a route: named skills in order, with the first few marked as ones you probably already know.
That route is a hypothesis, and we try to make the screen say so. At this point EvoluteMe knows your goal and two sentences about you. From that it has built the shortest sensible path from a person who knows nothing to a person who can do what you named. It does not know you yet.
It would be more comfortable to call it a plan, but it would not be honest. A plan is something you follow. This is a guess that needs correcting, and the correction is the next step, not an optional extra.
We still show the route before the check, because you can argue with a route you can see. People read it and say: interfaces before collections, really? Or: I do not need the first four of these. Both reactions are useful to us. The check is about to confirm or reject the second one.
Then a short check
The last thing in the first two minutes is a check. It moves forward along the route until it reaches something you cannot confirm yet, and your first lesson starts from there. It takes only a few questions. What you get from it is a place on the route, not a result.
Why help is switched off during the check, and what happens afterwards to a skill it confirmed, is explained in a separate post.
What we took out of these two minutes
A level test with a result at the end. We built one. It gave the name of a level. Nobody on the team could say what to do with that name on a Tuesday evening, and that is the only question this screen has to answer.
A tour of the interface. Four screens with arrows, explaining a product the person had not used yet. People clicked through it quickly and nobody remembered it.
A choice of learning style: visual, practical, theoretical, the usual list. It sounded considerate, but it mostly measured what people wanted to believe about themselves. Instead we watch which explanation actually helps you move forward. That costs you nothing.
Where this still goes wrong
Two minutes is the good case. The bad case is a goal too big for one route. People really do type "I want to be better at my job", and the route they get is wide, shallow and not much use. We ask a question to narrow the goal, and it does not always work.
The other problem is the opposite. Someone types a goal so narrow that it is really one lesson, and gets a route of three steps with nothing after it. That is a good outcome for the person, but a strange one for a product built around a long route. We have not decided how the screen should behave in that case.
We can see both problems only because we keep the goal in the person's own words. A catalogue tile would have hidden them, and we would have built a route for something nobody asked for.
