One number everyone trusts: rebuilding a single source of truth
Before the dashboards, before the artificial intelligence, before any of the sophisticated things organizations want from their data, there is a plainer requirement: one number everyone trusts. Amazon Robotics did not have it where it mattered most.
The situation
The Plan of Record drove demand visibility and field resource planning across eight organizations. It had grown inaccurate, lacked a single source of truth, and left teams chronically out of sync. Competing priorities had kept a fix from reaching the top of anyone’s list, which is the usual fate of shared data: everyone owns its consequences, so in practice no one does.
What I did
I wrote a white paper to make the case and win buy-in from the product leader, my manager, and the teams consuming the data, because a fix nobody has agreed to is just another competing version of the truth. I identified the right points of contact with trusted data, standardized the inputs, and wrote the user stories for the data team to build against.
Then came the choice I would make again. The data team’s backlog ran three sprint cycles deep, so rather than wait, I ran a weekly intake with the product and demand teams and shipped a manual interim fix within two weeks. It was a spreadsheet. It was ugly. And it meant no team was blocked or working from stale numbers during the four months the automated, real-time version took to build, with demand signals captured directly from their sources.
What happened
The rebuilt Plan of Record served more than 25 portfolio solutions, six engineering and product departments, four regions, and 19 countries, on an installed base of one million robotic units. Finance adopted it as the single source of truth, which cut approval and audit iterations, and the approach was adopted more broadly across the organization.
The interim spreadsheet is the part people push back on. It felt beneath the ambition of the fix. But perfection would have cost four months of continued misalignment, and trust in shared data is built by never letting people fall back on their private versions, even for a quarter. Ship the bridge, then build the highway.