How Managers Can Build a Culture of Experimentation

You need reliable data. Over 80% of the time and costs of data scientists and IT groups is spent cleaning up data due to, for example, inconsistent inputs and outdated views of buyer behaviour and legacy assumptions.

You should build tests from data you are confident about.

The most difficult challenges in testing are internal processes. While managers are good at asking questions, they fail to specify what would constitute a feasible answer to those questions. Managers must always interpret data with a specific end in mind.

For example, pricing has many dimensions, such as base price, discounts, rebates tied to volume, special offers, etc. The evaluation criteria for price testing should be established because testing means evaluating alternatives.

Companies tend to allocate testing time and money to large initiatives while ignoring small ideas that can have a bigger impact. Seek progress, not perfection. Invest in processes that encourage employees to submit small ideas.

• Use your website purchase interactions. For example, when airlines ask if a trip is for business or personal, they gain insight into price sensitivity for upgrades.
• Occasionally rotate the questions you ask to gather insights from a different perspective.
• Engage users and non-users.