The decision and the reasoning

An operator whose four hardest functions answer to one regulator

A licensed cannabis operator engaged Common Ground in September 2023 under a consulting agreement, a contracted engagement from the first day. The ask had two parts. First, research across all of the operator's stores: how each one actually runs, and where the operation as a whole loses sight of itself. Second, a master control center: one place where the people running the business can see the supply chain, the warehouse, the inventory and the compliance position for every store together, and act on what they see. We delivered both.

A licensed operator is not an ordinary distributor with an unusual product. Every unit it holds has to be accounted for from the moment it arrives to the moment it leaves, in the seed-to-sale tracking system the regulator reads as well as on the shelf. A count that does not match the tracking record is not a bookkeeping error to fix at month end. It is a compliance event. Product ages, and product that ages out or fails a test cannot simply be marked down and sold; it has to be held, remediated or destroyed under rules, with a record of each step. Supply can only come from licensed sources, so a missed reorder cannot be covered by a phone call to whoever has stock.

That is why the four functions in the brief belong together. In most businesses, supply chain, warehouse, inventory and compliance are four departments with four reports. In this one, the regulator reads all four as a single story, and a mistake in any one of them shows up in the others. What the operator needed was not four better reports. It needed to see the handoffs.

Why the research came before the build

There were two easy ways to get this wrong. The first was to start building software on day one, from a feature list, because an operator asking for a control center usually has a picture of it already. The second was to wire a screen into every system the operator runs before anyone knew which of those signals mattered. Both skip the same step: finding out which decisions actually matter, who makes them, and what they look at when they do.

So the order was fixed before any code. Research across every store first, because a dashboard built before you understand the decisions it serves ends up showing the data that was easy to reach rather than the data the decision needs, and because a gap one store has learned to live with stands out when it sits beside all the others. Then the build, scoped from what the research showed rather than from the original wish list: which decisions the screen serves, which function owns each one, and which of the operator's systems it has to read.