Fog Creek

From Idea to Product with Creek Weeks

Fog Creek Software has been going for over 15 years. During that time we’ve created many products. Some have worked out well (like FogBugz, Stack Overflow, and Trello), others, not so well. But one thing they all have in common though is how they were created: through Creek Weeks.

Creek Weeks are one way we embed innovation into our culture. Innovation can be a difficult subject to pin down – there’s a lot of hand-wavy nonsense that’s written about it, trying to make it seem like this mysterious and magical thing. So I want to eschew that and give you the detail about this specific, actionable activity. Creek Weeks help us focus our efforts on new, potential product ideas and kick off a process that turns those creative ideas into usable and saleable products.

betaTestingBanner

What’s a Creek Week?

A Creek Week is a week away from your day-to-day responsibilities to work on an idea. At Fog Creek, anyone can take a Creek Week at any time, with just one requirement: you have to convince at least one other person to join you to work on that idea too.

Ideas are managed on a shared Trello board, where people contribute ideas, discuss problems, potential solutions and seek out other people to work on them. An important thing to get right up-front is to specify the types of ideas that you as a company are looking for. This may be in terms of financial goals, the number of customers, area of expertise etc. There are lots of potentially good, viable ideas. If though what your company needs strategically is to diversify, or to reduce capital expenditure, then small-fry ideas or capital-intensive products aren’t going to work. So, it’s better to be clear about this from the outset and keep things focused.

Another important point is that managers and fellow team members need to do whatever’s necessary to facilitate that person’s time away from the team. If you allow “we’re too busy right now, maybe later” to cut it, then you’ll never do it. After all, in practice when are you ever not busy? Innovation is an important thing, it needs to be a priority and time must be given to it.


“At Fog Creek, anyone can take a Creek Week at any time”


What’s Involved in a Creek Week?

The goal of the Creek Week is to find out why your idea won’t work. That sounds pessimistic I know, but as a small company we want to fail fast so that we can focus on the ideas we do think will succeed.

We start by trying to identify the most difficult or important aspect of the idea. As a developer-led company, the temptation is often to jump straight into the technical details, but in truth, that’s rarely the most challenging part. More often it’s a market research issue, like how many other people have this problem? Will they use the solution and will they pay for it? Of course, you can’t solve all problems within a week, especially not the most difficult ones. So we don’t necessarily try to solve them, but instead, we assess whether they’re solvable. Either way, a typical week involves a mixture of thinking, coding, research, and a lot of talking.

Before and after, fellow Creekers contribute advice, questions, and feedback. However, during the week it’s heads-down time – the team may seek out people for input, but otherwise it’s about exploring the idea for yourself.


“The goal of the Creek Week is to find out why your idea won’t work”


Beyond the Creek Week

By the end of the Creek Week, you have a good idea whether you’re on to something or not. If you think you are, then during the week you will have identified a number of other key problems, and resolving these should be the focus as the project develops.

The ideas that we decide to keep on developing continue with the initial members. There’s often a lot of interest in these projects, many people will have opinions, they may want to be involved and there are all sorts of practical questions and considerations. However, it’s all too easy for a developing idea to get derailed or distorted. With many more people involved it’s difficult to reach consensus and keep the project moving forward, especially in a flat organization, like Fog Creek.

What’s more, there’s deliberately no heavy management over the innovation projects. It’s supposed to be lightweight and feel like a startup. The only people that can cancel the projects are the team members themselves, who are trusted to decide whether an idea continues to be viable, or by the founders if they think the project has become lost and there’s no alternative way forward. In practice, we rarely jump to canceling a project straight away – to have gotten this far the idea itself is solid, so the approach might just need to change. Instead, we might pivot and try to attack the problem in a new way.

As the project progresses, we have numerous check-ins and points for reevaluation of the project. These are primarily fortnightly meetings with the founders where they check on progress, ask about issues and provide advice. As well as quarterly reviews, in-line with wider strategic company planning.


“It’s all too easy for a developing idea to get derailed or distorted”


As the product develops, and the idea and approach solidifies, we add additional resource as needed. Glitch, for example, quickly moved from 2 to 3 people. They built the core product before additional resource was added in the final months before launch, ending up with a team of 8 (4 full-time, 4 part-time) contributors.

How many projects we have on the go at any one time is dictated by the company’s wider goals, but we’ve pretty much launched at least one new thing each year since the company began back in 2000. Creek Weeks don’t guarantee the success of the product, but they are one way to make sure you’re giving time to innovation. Of course, to be truly successful with innovation it needs to be a continuous effort, rather than a periodic one like with a Creek Week. So they’re just one tool of many. But maybe they’ll work for you too?

In a follow-up post, we’ll walk-through how Creek Weeks work in practice by reviewing two recent projects – HyperWeb and FogHorn.