I was in a hallway conversation where a designer peer who has been around the block asked for my take on a problem. It was about changes the surface representation of a given digital experience – let’s call it a “skin” though we were talking about a speech interaction.
If the way that you apply that skin takes a couple of steps for procedural reasons, such as identification, authentication, purchasing, and changing default settings. The way this system is architected, when you got to the end, you got a confirmation saying, “hey, I applied that skin like you said, and that’s how it is now.”
The question was, “What do you do if the user says, ‘oh hell no change that back!'”
So I asked: Are we sure the user wanted to do the thing?
Well, this is the fourth step of a flow so it would be fairly difficult to get all of the way there with false accepts at every step so, yes, let’s assume so.
The next thing I asked was: And the assumption here is that because the thing is both a purchase and a pretty specific choice – you don’t buy an Andrew Dice Clay comedy album or a Scientolgy text by accident, so is this more like that? Or more like ordering in a restaurant, “actually, could I have the salad instead?”
Definitely Andrew Dice Clay. You picked this on purpose and did work to get it.
So what is the concern?
Well, some people who we talked through the design were like, “How do I cancel out if I suddenly realize I did the wrong thing?”
My final question I kept to myself. We talked through ways that you could identify and capture ranges of intents and do daily or hourly log queries, some general tech capabilities we might be able to apply here, and that was that. Back to work.
My final question was, “Why did we build a Chinese finger trap?”
Ignoring the ways you might implement a system to use progressive disclosure, familiar words, legalese tick boxes and numerous steps to ensure that a user would not even end up in the situation where this problem were possible, my primary concern was that the people who had driven the product from the beginning had built it as a trap.
Get the revenue! Get the impressions! Get the clicks! Get the engagement! Roach motel it!
(Obviously these folks would not say “roach motel it” – Except for people who straightforwardly adopt the most coercive tactics of Hooked and other manipulation textbooks, most people in my experience solving this problem this way are merely doing what their boss said. If they were a role playing game character they’d be “Unprincipled” or “Neutral”.)
The real problem they were trying to solve was once they saw the implementation-level experience, they saw the user would see it was a trap, and back out. The problem was it wasn’t deceptive enough.
But it was too late to build it differently, so now we could only bolt things on at the end and say they represented safety and choice.
Like seat belts in 1968.
Build a velvet rope and an exit door to the roach motel if people might decide they don’t want to stay.