“A core component of Lean Startup methodology is the build-measure-learn feedback loop. The first step is figuring out the problem that needs to be solved and then developing a minimum viable product (MVP) to begin the process of learning as quickly as possible. Once the MVP is established, a startup can work on tuning the engine. This will involve measurement and learning and must include actionable metrics that can demonstrate cause and effect question.”
Eric Ries, The Lean Startup.com
I like the concept of developing a Minimum Viable Product (MVP) to gain initial momentum for a product and evolve it quickly to introduce what the market likes/needs (and hack what it doesn’t). In effect, it’s the rapid prototyping model that can be a fantastic approach to OSS product development, especially when the product is so new / innovative that customers haven’t seen its like before.
But today’s blog takes a slightly different twist on the concept – that of Minimum Viable Data (MVD).
As we all know, data is OSS‘s Kryptonite, but it’s also what makes OSS so valuable (potentially). Creating, maintaining. cleansing and manipulating data can take up enormous amounts of an OSS team’s time and attention, yet data quality can still quickly spiral out of control and render an OSS barely useable.
When time and resources are at a premium, I believe that lean data should be one of the tools in the “ruthless simplification” armoury. Less data equals less effort on creation, less effort on migration, less effort on maintenance and less effort on integration.
When you are starting from a position of having unreliable or incomplete source data, it is even more important to reduce your dependence on the data to avoid lengthy data rectification processes. When it’s your front-line staff creating source data (eg network audits, as-built record creation, etc), the data collection activities are generally slowing them down from achieving their key deliverables (eg network construction), so you want to minimise the time they spend on it (ie collect less!).
But what is MVD exactly? I define it as the minimal set of data that still allows an organisation to commence and continue using an OSS tool. When evaluating what data sets are essential to record and maintain, if in doubt, leave it out…. but ensure you have a build-measure-learn feedback loop to put it back in if it’s absence is hindering your organisation. Hypotheses have to be tested under an MVD regime.
Now I would like to point out an obvious quandary – just as there is a place for Big Data (especially when data collection and flows are automated) due to the insights that can be mined, there is also a place for Small Data (especially when speed / efficiency is of the essence!)Read the Passionate About OSS blog for more.