Study of firms in an accelerator programme identifies key limitations to the ‘build-measure-learn’ scientific approach of early venture creation, says paper by Monique Boddington and Stelios Kavadias of Cambridge Judge.
Most new enterprises fail, so it’s well established that fledgling firms need to “pivot” their products, resources and capabilities to find a successful business model. Yet there remain divergent perspectives on the process by which founders and venture leadership teams reason on how to pivot.
On the one hand, there is deductive reasoning that focuses on hypothesis development and controlled experimentation, and on the other, inductive reasoning that synthesises complex information including past observations and more general knowledge.
A two-year study of 21 early-stage ventures helps cut through this friction: a close examination of these firms finds that deductive reasoning works best during a venture’s idea validation phase, yet at early stages of idea creation and later scaling stages inductive reasoning provides more benefit because it accounts for the higher ambiguity or complexity present in these stages.
“Our findings bring clarity to the tension between the two approaches of strategy formation, as they qualify the conditions that make each approach valid,” says the paper co-authored by Monique Boddington, Associate Faculty and Deputy Director of the Masters in Entrepreneurship programme at Cambridge Judge Business School, and Stelios Kavadias, Margaret Thatcher Professor of Enterprise Studies in Innovation & Growth at Cambridge Judge and Co-Director of the School’s Entrepreneurship Centre.
Importantly, the research identifies boundaries to the often-advocated “build-measure-learn” approach to enterprise creation, part of what’s known as the “scientific approach” to entrepreneurial decision-making that traces its origin to the Lean Startup movement. The study concluded that such a deductive reasoning model might limit venture development at certain crucial stages.
“There are times when venture founders need to draw inferences from ambiguous and complex information, and an overly restrictive experiment-based approach at certain stages could lead to erroneous inferences and therefore choices,” says co-author Monique Boddington. “Early-stage ambiguity or complexity emerging at later stages can render experiments ineffective.”
Adds co-author Stelios Kavadias: “The study brings into question the universal application of the scientific approach to early-stage venture formation. We conclude that alternative forms of reasoning can often work best in enabling successful choices for ventures.”
The study is based on 21 ventures that joined an accelerator housed at a European business school between June 2016 and April 2017, which were closely tracked over 24 months including via interviews with founders and other key stakeholders. A total of 129 pivots (defined as changes in the factors that may impact a venture’s performance) across the 21 ventures were identified over the two-year study period.
The research grouped pivots into four pivot dimensions: market, product or service, execution or operations, and organisation. For example, the market dimension of pivoting involved narrowing or widening of market, shift of market segmentation, and route to market, while the organisation dimension included changes to the founding team and adjustments to the partnership structure. One other theme also emerged: too much or too little pivoting can be hazardous to a venture’s long-term health: “In our sample, ventures that ceased to exist exhibited one thing in common: they did not pivot at all, or they pivoted very rarely. However, frequent pivoting also hampered venture progress.”