Definition
Ratio of order variance to demand variance across the replenishment cycle, measuring whether the autonomous reorder system is amplifying or dampening demand signals as they propagate upstream.
The Measurement Gap
Autonomous replenishment systems place orders without human review, and the standard metrics used to evaluate them focus on fill rate and inventory turns. Fill rate measures whether shelves are stocked but cannot detect whether the ordering pattern itself is creating upstream volatility. Inventory turnover measures capital efficiency but is blind to the variance structure of the orders generating that turnover. The bullwhip effect, where small demand fluctuations amplify into large order swings as they move upstream, was well-documented in human-managed supply chains. Autonomous systems were expected to reduce it through faster, data-driven responses. But AI reorder systems that react to every demand signal without dampening can amplify variance faster than human planners ever did, because they respond at machine speed to every fluctuation. Sales per square foot and gross margin return on investment (GMROI) reflect outcomes but cannot trace those outcomes back to ordering variance as a root cause. Without measuring the variance ratio directly, an organization cannot determine whether its autonomous system is making supply chain dynamics better or worse.
Decision Supported:
Is the autonomous replenishment system making supply chain dynamics better or worse? Should ordering constraints or dampening rules be added?
Formula
Plain language: Compare how much the orders fluctuate to how much the demand fluctuates. If orders swing more than demand, the system is amplifying variance (bullwhip). If orders swing less, it is dampening.
Notation:
ARVI = Var(Orders_t) / Var(Demand_t)
Variables:
| Variable | Definition | Unit |
|---|---|---|
| ARVI | Autonomous Reorder Variance Index | Dimensionless ratio |
| Var(Orders_t) | Variance of autonomous order quantities over the measurement period, aggregated at the category-DC level per week | Units squared |
| Var(Demand_t) | Variance of actual demand (POS sell-through) over the same period, same aggregation | Units squared |
| t | Measurement period: rolling 13 weeks minimum for stable variance estimates | Weeks |
Proposed formula, pending validation pilots.