Recent Events for MainPageDiary (Blog) Previous Next

2008-09-13 The Corporate World Should Learn From Internet

The Corporate World Should Learn From Internet

A recent blog post of Alex Schroeder about corporate wikis and my comment on his post, remind me of a post I wanted sometimes ago to make about the difficulty for a corporation to learn from Internet. Internet, beside being a big mess, works often better to manage large volume of information and on-line/distributed collaboration than any other internal company designed software/process. I don't really want to understand why it is the case on Internet maybe due to the "trial-and-error" approach with the continuous feedback (well described by Linus Torvalds in this post about Kernel's evolution). I just want to point some areas where the corporate world could learn from Internet to avoid classical pitfalls.

Books about ideas and "innovation"Patents are only for the old machine

Start Small And Prototype Early (And Forget All Those Bloody Requirements...)

As nicely and visually explained in the post about corporate Wikis from Alex Schroeder, new software project in a large corporation often start by accumulating all the requirements from the "stakeholders"(in other words : anyone that could be affected by the introduction of the new project). But does it work in practice? From my small experience on the subject, I tend to say "not at all". First, it's usually impossible to find any (free or proprietary) software that meet all the requirements. In such situation instead of throwing away some requests from the "stakeholders", the software project is trying to extend the software to meet all requirements. That generate unrealistic requests like modifying existing software (the famous "small customization"), creating software from scratch, over extending a work flow to match those crazy requirements or worst making a two hundred pages requirements document sent to a software engineer already working on 10 "corporate" projects.

Successful Internet projects often take the opposite approach by starting something small and doing it well. The first release of was simple and clearly focusing on a simple multi-user bookmarking service. But it worked. Various Internet services started very small due to their small asset and money constraint. Starting small with a really focus objective limit the risks of creating a monster that no one want to use. If it fails, you start a new project. On Internet, failure is accepted and welcome. Especially that helps (only helps ;-) to avoid repeating the same failure and experimenting other ideas that could fail or work well.

Companies should look more into this approach of creating very small project, through away "consensus among all the stakeholder" and implement ideas by prototyping early. I know that's not an easy task in the large companies especially changing (read killing) the approval process of companies (often good Wikis don't care about approval as described in Why it’s not a wiki world (yet)) or the hiding failure as default policy and we don't code we just make meetings. Don't take me wrong, I'm not trying to say that everything is broken in companies. But they should start to understand they are not living in island and they can gain a lot from Internet practices inside the companies.

This post is somehow related to a previous stacking idea page about how to better operate a company.