Agile practices for Startups
When working with accelerators, I have often been asked to present a workshop on agile processes. This post is a summary of they key points I highlight when talking to startups.
Startups are not smaller versions of organisations, rather a startup is an entity in search for a repeatable and scalable business model. The founder has a ‘big idea’, and it is essentially a faith-based institution. The business model of a startup is based on assumptions e.g.
- Who the customers are?
- Path to customers
- Price customers are willing to pay for product
What startups need to do is identify the riskiest assumptions and test these assumptions in the quickest way possible before running out of money. This process is often referred to as discovery.
Established organisations on the other contrary are designed to execute on a repeatable and scalable business model. In this case, the focus is on execution and delivery.
Given this context how should agile processes work in startups compared to organisations?
- Startups need to be really good at running experiments and validating assumptions. A startup needs to cycle through the Build, Measure and Learn loop as fast as possible and extract the maximum possible learning through each cycle. The backlog should be experiments that need to be run to validate hypotheses. The emphasis is not well-crafted stories rather on formulating clear hypotheses, running relevant experiments and validation. This may or may not include writing software, it maybe customer interviews, concierge MVPs, prototypes etc.
- One of the key values of agile is ‘Working software over comprehensive documentation’ . Working software is important in organisations, but in a startup learning is more important. Software craftsmanship and tested code are useful if you have customers willing to pay for your product. Quoting Kent Beck, one of the original signatories of the Agile Manifesto, for startups this value should rather be: ‘We value learning over working software’
- Another agile value is ‘Customer collaboration over Contract negotiation’. A startup does not have customers, so there are no customers to collaborate with. You have to go and find your early adopters using customer development.
- Agile processes typically focus on velocity as a key metric; this is because the focus is on delivery. This is not relevant in startups. The key unit of measurement is the velocity of validated learning. How quickly are you validating your assumptions and learning?
- One of the breakthroughs of agile thinking compared to waterfall development is an iterative and incremental approach to software development. This is still very relevant to startups. Testing assumptions in small steps with feedback loops reduces waste in startups.
I highlight the above points and prefer not to go into the details of any agile process like Scrum and Extreme Programming. This is more relevant when startups find product/market fit and focus shifts to execution.