Build the right thing first.

A backlog tells you what someone asked for. It rarely tells you why. Six months in, nobody can say which user a feature was for, what problem it solved or whether it worked, so nothing gets taken out and the product keeps growing sideways.

We work the other way round. Our delivery framework, HASTE, starts with two phases before anything is designed. In Hypothesise we write down what we believe about the problem and who has it. In Assess we test those beliefs for desirability, feasibility and viability. The output is a recommendation to proceed, pivot or stop, and we mean all three.

The evidence does not end up in a report on a shelf. It lives in Wayfinder, our own discovery and planning platform. Users, their jobs to be done, their pain points, research findings and hypotheses are all records, and every feature on the storyboard links back to them. A feature with no user or job linked to it is not ready for delivery, and Wayfinder says so.

The approach is older than the tool. At the Fair Work Commission, stakeholder interviews and research with different audience groups shaped the new website, and the average time employees spent on key tasks improved by 79%. At DITRDCA, workshops with 46 stakeholders across 14 business divisions narrowed a merged intranet down to the five services staff use most.

Every build cycle ends the same way: we check what shipped against the hypotheses it was meant to test. Then you decide, with the evidence in front of you, whether to keep going, change direction or stop. Stopping early is a good result when the evidence says so.