Last Responsible Moment

Probably the least understood (or most misunderstood) lean principle is "decide as late as possible". I have seen it used to justify all sorts of weird decision-making policies that generally involve never making decisions, because surely as late as possible means leaving it until the absolute last possible moment, or even later. I have seldom, if ever, seen it applied correctly. So let's take a look at this principle and see what it really means.

The other way to express this principle is "defer decisions until the last responsible moment". There are two points of confusion here. The first is what is the last responsible moment? The other is what exactly do we mean by deferring decisions? Let's look at the last responsible moment. What is the last responsible moment? Does it mean the absolute last minute? Do we leave all decisions until we are absolutely forced to make one because otherwise the whole endeavor will fall flat? No. That makes no sense at all. Leaving decisions until they are forced upon you is hardly being responsible. Does it mean making decisions early because that's the responsible thing to do? Again, no. Making decisions early isn't using the last responsible moment. The last responsible moment is a really hard thing to define, so let's not try. Let's re-word it instead. The intent of the last responsible moment is to make decisions with the maximum possible information.

At the beginning, we know the least about what we are doing, so making decisions early is making them with the least possible information. That's not the responsible thing to do. Neither is leaving them to the absolute last minute. If you do that, quite often, the decision is made by default - it's the only option remaining. It may not be the best option but it's now the only one open because you waited so long.

Making decisions with the maximum possible information means collecting enough information to make the decision, then making it. Don't make it too early. Don't leave it too late. Don't make it with too little information but equally, don't get stuck in analysis paralysis, collecting more and more information until it's too late.

Now what about deferring decisions. What do they mean by that? Do they mean do nothing? No. That's exactly what not to do. Deferring making a decision does not mean doing nothing. It means collecting information until you have enough information to make the decision well. When you are deferring a decision, you aren't doing nothing, you are learning.

When you defer a decision, you do so for a reason - you don't have enough information yet to make it. So you need to work out how to get that information. You may even need to work out what information you need to make the decision before you can go about collecting it.

When you defer a decision, you need to have a plan in place around what information you need (or what you are going to do to find out what information you need) and a plan for collecting it. It may be some research. It may be some experiments. You may need to build multiple prototypes to assess competing options. You may need to speak with customers and stakeholders. Whatever it is you need to do, you need to have a plan for doing it. You can't just sit around and wait. Deferring a decision is an active process not a passive one.

When you defer a decision, you need to consider until when you can defer the decision. How long do you have before you have to make the decision. At what point does your ability to make the decision get taken away because other options get closed off by default? If you're considering two options for a technology and one will take 3 months to develop and the other 6 months, if you are 8 months away from your release date you only have two months to make the decision. Any later and the 6 month option (which may be be the best option even though it's the longest - better performance, lower maintenance etc) is no longer viable. You have waited too long and only the 3 month option is available. You didn't make the decision, it was made for you.

At what point do your options close off on their own? At what point does your ability to make the decision vanish? That may be months or years away, it might be tomorrow. Some key decisions that impact the whole structure of your solution, like which development platform to choose, probably need to be made pretty quickly or they will delay the whole thing. Some decisions, like the colour of the buttons on screen, can probably wait for some considerable time.

For those key decisions, the ones you need to make quickly, you may need to collect information quickly. You need to collect as much information as you can to make the decision, before you need to make the decision. You probably won't have time to collect all the information you need so you will have to make do with just enough. Prioritise the information you need. Collect the most important information first. The stuff that will have the biggest impact on your decision making. Leave the less important stuff until later. If you can collect it in time, all well and good. If not, well you have the most important stuff sorted. On the other hand, if you collect all the information you need, long before the last possible date, just make the decision. You don't need to delay if extra delay brings no further benefits in learning.

Decide as late as possible is not an excuse for inaction. Deferring a decision is a process of active learning; information collection. The intent is to make the decision with the maximum possible information by actively learning right up until the point where you have to make the decision. Then, with as much information as you have been able to collect, and with all your options still open, make the decision.