RedHub AI
How the roll-up gate works
How the Estimate-to-Actual Feedback gate works
This tool grades every closed job on whether what it cost ever reached the people who price the next one. It reads 8 closed jobs and rolls them into one verdict — the worst, never the average — then a gate that can only make it worse.
The 8 closed jobs
Two triggers, and the second is unusual. Any job whose cost codes do not map to its estimate forces LOOP OPEN, because that comparison is not merely undone — it is not possible, and it is invisible to anyone measuring drift since there is nothing to measure. The second reads the WHOLE register rather than any row: if no job anywhere ever changed the estimating basis, the loop has never once closed. One basis update anywhere releases it, and a recorded decision to change nothing counts.
Why: Five of seven in-scope jobs read FED BACK and 71% of the practice fed back, which is a respectable number. Both triggers still fire. Municipal Annex's cost codes do not map to its estimate, so that comparison was never possible. And across all seven jobs, not one ever changed the estimating basis — a condition on the register rather than on any row, and the one a share can never see.
This grades the loop, never the drift. It computes no variance and never says what anything should have cost, because in construction that figure is usually unavailable rather than bad — every project is a one-off and nobody went back to look. So the practice reads on the share that fed back rather than its worst job, since one sloppy closeout does not break a loop — and the gate catches the two things a share cannot: a comparison that was never possible, and a register where nothing has ever changed the basis. Grades a register of closed jobs you fill in, never a person. It reads no accounting system, no estimate and no job cost, computes no variance, productivity factor or unit cost, and makes no claim that your numbers are good — only about whether anyone ever found out.