When you launch innovation activities or projects, you never have a perfectly clear understanding of the outcome and output. You never know in detail, if your activities are worth the effort, and thus it’s hard to justify the funding and resources allocation. How can you claim, that the investment will be fruitful?
You need to embrace this amount of uncertainty and risk, but there is a safe way of doing so. It provides transparency in your train of thoughts and activities, and gives comfort to you, your team, and stakeholders: Use hypothesis-driven experiments.
This blog post gives insight to the tool and template, and should be read as a follow-up to the blog post on speedboats.
The innovation activities we’re looking at in the modern workplace all have a need for fast results. All the initiatives are grounded in the “try, inspect, adapt”-thinking, in line with both the agile approach and the need for responsiveness. The outcome and output must be reached quickly (e.g. within 10 or 100 days), and have a tangible result. This also applies to activities, that are purely explorations.
The quest must be written in clear text, and presented with the end user, employee, or customer in mind. You can do that by finalizing these two sentences:
- Our quest
- “Our idea is that …”
- “Our quest is to explore / investigate / confirm / validate …”
Then, you break this overall description into these parts
- Our hypothesis
- “We have a hypothesis that …”
- “It will solve this problem …”
- “It will create possibility for …”
- “We execute this experiment: …”
The learning objective
The learning objective is inspired by the work of Eric Ries and his book “The Lean Start-up”. The learning must be explainable and presentable; either numbers, visitors, qualitative interviews, transactions, or similar things that are tangible:
- Our inspection
- “We expect this outcome: …”
- “We validate the learning via …”
- Our adaption
- “What did we learn …”
- “Use – or throw away?”
- “We propose the following next steps: …”
You can setup a hypothesis-log like this (click to expand):
You can also take a look at the toolkit among others here.
Thank you to Maz Spork and Søren Skov, for inspiration and co-creation on this approach!