If a Startup is determined to ascertain which features to prioritize, how to convert more customers? And how to spread the word about its product? It must be willing to put its vision and business plan hypothesis to the test. Startup experiments championing different versions of a product are veritable means of conducting this test.
Upon discovering its error, a disciplined startup team can quickly adapt and change its methodology.
To truly draw on Validated Learning, a startup stands itself up to design and control its experiments, and to lead its customers on, instead of being led by them. Startup experiments are often designed as split-test experiments that get up different versions of its product forward to customers. The impact of the different versions of the product is revealed by changes in behaviour between the two groups of customers. The customer behaviours indicate the level of value ascribed to the two different versions of the product.
Through split-testing, a startup invariably learns that certain features considered important by its engineers have no impact on customer behaviour. The test helps to eliminate work that does not influence customer behaviour.
Startup metrics depart from the realm of vanity and become actionable when it can be linked to specific actions undertaken by the Startup team. In such a situation, the team is able to learn from its actions.
A cohort-based metrics highlight the periodic number of customers exhibiting behaviours that validate or disprove a startup’s value and growth hypothesis, its riskiest business plan assumptions.
