(The following blog kicks off what we hope is an ongoing and lively discussion among service providers about the concerns of their business. At least once each month we will post another blog which may address ideas already raised, or suggest something else to consider. It should be fun and informative. Please comment.)
by Joost Pisters, Head of Product Development, LuxCloud
Recently, I have been captured by a book called “The Lean Startup” by Eric Ries. Ries defines a startup as “an organization dedicated to creating something new under conditions of extreme uncertainty”. Ries is promoting something called ‘validated learning’ which means that when you make improvements to your existing products or create new products you should first define a hypothesis and then seek a way to prove (or disprove) the hypothesis by measuring these ideas in the quickest way possible.
Very much like in online marketing, we should create A/B split tests for product features and test whether the improvements we want to make are actually getting the results we are seeking. This can be a small hypothesis like: “I believe that adding bigger buttons to our control panel will make it more attractive for SMB’s and will drive return visits” to the more fundamental: “We believe a hosting control panel should be more than a technical tool and is THE marketplace to engage your existing SMB customers with new and exciting products that ultimately drive more sales”.
I have a lot of these kinds of “beliefs” that I could argue about till the clouds come home but until I scientifically prove them they remain just that, beliefs. Even if I would create a new product that incorporates all these beliefs and that product becomes hugely popular, this still proves nothing. Maybe it was our great sales force or a blog post about the product that boosted sales. Maybe it would have been more successful if we would have left out this or that feature.
When it comes to improving the Parallels Control panels, I would love for Parallels to ask every partner to post one well-formed and testable hypothesis and then create a quick experiment with that reseller to prove or disprove it. Parallels Partner X believes that “adding a product suggestion engine to the upsell page will boosts sales”. End result could be that the experiment shows that the attach rate stays the same with or without this feature. Lesson learned, partner happy that Parallels took their idea so seriously, and we can move on to the next hypothesis. The ones that are reaching their goal get a place in the next release and everyone profits from the cloud knowledge. Let’s post these ideas and let’s convince Parallels to work on the Lean Control Panel!