“Whilst doing rather than planning has been a hugely successful tactic for entrepreneurs and their investors, before I go any further I want to note that as with everything you can take it too far. To get the best chance of achieving huge success, and avoid getting stuck at a local maxima, a certain amount of thinking should be done before building starts. There’s a balance to be struck and whilst best practice is definitely to maintain a bias towards action in our experience an increasing number of founders are starting to build product before they’ve done enough thinking, sometimes encouraged by investors who want to play with product before they invest. Many of these founders end up failing when with a little more customer research they might have built a slightly different product which would have resonated much better and allowed them to iterate to success.”
Nic Brisbourne here on his blog.
As usual, Nic’s blog articulates some great ideas, none of which have anything to do with OSS, but have everything to do with the future way of work, including on OSS.
One of the biggest challenges with the OSS of the past is that the build-measure-learn loops have had a cycle time that could be measured in years. That’s simply too slow for the modern communications industry or any of the industries that make use of communications technologies.
For most OSS, a vendor might develop a concept, take months to build it and then install it. The first question is whether they’re able to stick around long enough before measuring anything meaningful (that part usually becomes the responsibility of the customer when they’re using the commissioned OSS tools). And if the builder is not measuring, then are they also removed from the learning stage of the loop?
By contrast, the small-grid OSS has a much tighter involvement in the build-measure-learn loop and also builds tools that allow customers (the xSP) to quickly build-measure-learn from the tools that offer a service to their customers (businesses / individuals) who can also build-measure-learn.
That means the small-grid OSS model needs to think ahead to deliver three levels of build-measure-learn loops from the outset. Can you think of any OSS that currently facilitate this?
PS. I tend to agree with Nic’s opening quote too – it makes sense to put a bit more time into an initial measure and learn cycle before the first iteration of build.