Giving permission to experiment comes first, but it isn't the only thing.

Giving permission comes first, and I wrote about that before. But permission is the way in, not the whole list, and you learn that by watching things die that had the permission granted, the budget approved and everyone nodding in the meeting.

What was missing was never in the speech. It was in how the company works on the inside.

Tim Brown, in Change by Design, lists the conditions that allow an experimental attitude inside an organization. The first one is that room to experiment reaches everyone, not just the designers, not just the engineers, and certainly not just management. There are five more, and Eric Ries, in The Lean Startup, adds another three from the resources side.

Nine conditions in total. I'll go to the four I've seen the most.

The problem is almost never money, it's priority

Ries argues that a team experimenting inside a company needs scarce but secure resources. Scarce because abundance doesn't force you to choose, secure because nobody explores calmly when every step depends on a new approval.

From what I've seen, budget is rarely the problem. When what you ask for is reasonable, it usually gets approved. What competes isn't money, it's priority, the place your experiment takes in a list that also holds operations, the year's commitments and what was already promised to a client.

And it's understandable, even if it stings. An experiment doesn't bring in revenue today. If it works out, and depending on what it is, it can open the door to more experiments and more spending, which over time improve a process or promise some return. But that return isn't today's, and the client is.

And there's a pattern there that's hard to accept. Incremental work fits without discussion, because its return comes fast and can almost always be measured. The further from that what you propose sits, the more its priority moves, and it has to wait for room. Sometimes the room never comes.

Whoever benefits isn't always whoever carries the cost

This is the one that has taught me the most.

Whoever experiments should have something at stake in the outcome, Ries argues. That's true, and it's also incomplete, because the outcome doesn't land the same way on everyone. I've seen teams push the implementation of something they won't even use, just because the benefit comes back to them some other way. And I've seen the opposite, an experiment that makes life easier for one area at the cost of adding work to another, and one of the two shows up to the meetings willing while the other always has something more urgent.

When the benefit and the cost land on different people, adoption is decided right there, not in the quality of the idea. It's worth looking at how it splits before starting, because afterwards it's too late to negotiate. And if it comes out very lopsided, finding a way to involve both sides, even if it means splitting the work so it hurts everyone a little.

Ideas never get to gather followers

Brown proposes something that sounds odd: an idea should have followers before it has a budget. Few, but loud.

Odd because in most companies that's impossible by design. Ideas don't circulate, at best they get presented. They live in someone's head or in a private document until the day of the meeting, and that day everything is decided in front of whoever approves. There were no earlier weeks for two or three people to get excited about it, defend it in a hallway and find its flaws. They arrive alone, and a lone idea is easy to dismiss.

Brown has another observation that points the same way. Suggestion boxes fail, he says, not because people don't come up with things, but because there's no mechanism to do anything with what gets suggested. The box fills up and that's where it ends.

It isn't anyone's bad faith, it's how the processes are built, but it explains why so many ideas die in the first presentation.

Having a strategy doesn't save you the explaining

Another of Brown's conditions is that an overarching purpose exists, so people know where they are headed and don't need constant supervision.

Here reality is more uncomfortable. The strategy can exist, the objectives can exist, and still what reaches senior management gets justified from scratch almost every time. Low impact work moves along without much explaining, and that helps more than it seems. But when something carries weight, you have to defend it anyway, and it's worth knowing that before sitting down to present, so you don't arrive assuming you already have the backing.

One I haven't solved

There's a condition of Brown's that says ideas shouldn't be favored based on who creates them.

It isn't easy, and I'm not going to claim I've seen it solved. What I am clear about is where the way out points, and it isn't asking people to be impartial. It's having an agreed way to evaluate and test ideas, so the question stops being who it came from and becomes what it proved.

So

Permission is the condition for entry. After that you need a priority that doesn't move every quarter, a split of benefit and cost that someone has looked at, some room for an idea to gather people before staking everything on one meeting, and a direction that doesn't force you to justify the world every time.

No company has all four.

Which one are you missing?

This is the second in a series. In the next one I get into the human factor and resistance to change, which is where a good part of what did have permission falls apart. And later into something left pending here, how an experiment gets commissioned, because a condition becomes concrete the day someone assigns it and defines what would count as success. If you want to follow the thread, follow me or subscribe to the newsletter.

Ideas taken from Tim Brown's "Change by Design" and Eric Ries's "The Lean Startup", reworked from my own experience.