Evolve your business, literally

One of the major advantages that a small business has over bigger companies is the ability to easily implement changes. This agility, however, is obviously only beneficial if you actually use it and are willing to adapt, even when things seem to be going well. One of the biggest impediments to this is uncertainty about what the "right" change is. But here's the thing: you don't need to know what the right change is as long as you can quickly determine if it worked. That's the basis for how evolution works, and it happens to work so well that the same strategy is used in applications to a huge variety of real world problems.

So how does evolution work? In the abstract, it's pretty simple. You begin with a starting point that already works to some extent. You then need a way to make changes (often at random) to the starting point. Finally, you need a way to tell if a change worked. With those three things in hand, you can imagine how the process works:

1. Make a change (any change) to your starting point
2. Evaluate the change; if it works keep it as your new "starting point", otherwise undo the change
3. Repeat

Now this probably sounds pretty obvious and simple (because it is), but the cool thing about it is that nowhere in that process do you need to know what the "right thing to do" is. That can be a pretty liberating realization as it lets you focus on the generally easier problem of evaluating a change that you've already made. Today I'll go into a little more detail about the key requirements, and later this week I'll talk about some specific ways to apply this type of process to software choices and other aspects of your business.

A starting point
This one's easy. You already have a company/website/etc that is successful to some extent, so that's your starting point. For many applications (including "real" evolution), finding a decent starting point is the hardest part of the whole process, but you've already done that so cheers.

A way to make changes
This one's easy too. In the unlikely event that you don't already have a million different ideas of changes to try out, spend half an hour brainstorming or browsing around online for some ideas and you'll be set for a while. The key point here is that you don't need to know ahead of time whether any of these ideas are going to work, just that they might possibly be able to work. For the most part, you also want these changes to be small enough that there's likely to be some impact, but not a massive one. If you own a restaurant, you might think about adding a few new menu items, but don't replace your entire menu.

A way to tell if a change worked
Well, the first two steps are easy, so this one must be hard, right? Not if you choose your problems correctly. The key tradeoff for this step is between relevance and speed. What you really care about is how a change impacts the overall value of your company, but that's a really hard thing to measure, even on a long time scale. Instead, try to figure out something that's easy to measure in a short amount of time. For websites, we focus on things like bounce rate (how many people leave your site immediately) or conversion rate (how many people sign up for your service, subscribe to your blog, etc). Those statistics aren't really the most important things to the owner of a website (maybe it's number of people who end up paying for a product or service, or how many people actively read a blog), but they're much faster to determine.

So stop thinking about what the right changes to make to your company are, and start thinking about how you can evaluate what happens when you make a change. Once you do, you've got a clear path to improving your business, even if you have no idea what specific steps you'll take.