Meters sites one at a time, monthly, and keeps the ones that moved the wrong way in the same table as the ones that improved.
A group-level energy figure will always look tidier than the sites underneath it. It is also where the inconvenient sites go to hide. A single average across a portfolio can fall year on year while a third of the buildings in it are getting worse, and nobody holding the average would know.
So each site is metered separately and read monthly. It takes longer, it produces an uglier table, and it is the only version of the table that tells an estimates director where to spend money.
Consumption is one half of it. The other half is what that consumption is being compared against, which is where most building reporting quietly falls apart. A site that took on a third more floor area did not become more efficient because its intensity fell, and an intensity figure with no denominator behind it says nothing at all.
A certificate bought after the fact is not the same claim as a contract that brought new capacity onto the grid. Both are legitimate; they say different things, and a report that blurs them is making an additionality argument it cannot support. We write down which one you hold, and date it.
Two of eleven sites moved the wrong way last year. They stayed in the table. They were the two worth a conversation.
Sometimes a portfolio does not improve. Reporting a flat year as flat, with the reason attached, costs less credibility than a reorganised baseline that makes the line bend. Clients rarely believe this until the second year.
