Your demand planning team spent Q2 tightening the forecast model. Mean absolute percentage error dropped two points. By the time the improved model shipped, spot rates had already moved 25% year over year and the lane plan built around last quarter's numbers was three weeks out of date.
That gap is the real story right now. Freight markets moved from episodic disruption, a hurricane, a strike, a one-off shortage, into a structurally volatile default. A forecast model tuned for episodic swings chases a target that keeps moving before training finishes.
Decision speed is the time between an operational signal landing with incomplete confidence and a qualified person committing to an action inside a bounded window, with the decision logged. It measures execution, not forecast quality: how fast an organization acts when certainty never arrives, and how fast it corrects once better information does.
Written from 5+ manufacturing and distribution engagements, 2022–2026, where planning teams kept refining a forecast while the operational decision that mattered sat waiting on it.
Why forecast accuracy stopped being the supply chain advantage
For two decades, the lever most ops teams pulled was demand-sensing precision: tighter error rates, faster S&OP cycles, better statistical models. That approach worked when disruption was episodic. A port closed for two weeks. A strike delayed a lane for a month. The model re-stabilized once the event passed, and the accuracy investment paid back the way it was supposed to.
Uber Freight's head of client services, Bob Daymon, described what changed on the Supply Chain Management Review podcast in mid-July 2026: spot rates are up 25% year over year, and none of the operators he talks to expect near-term relief. That is the new baseline a forecast model has to hit inside, not a spike waiting to revert.
Daymon named four traits separating the operators pulling ahead in 2026 from the ones still tuning their forecast:
- Speed and agility in decision-making
- A continuous improvement mindset
- Operational flexibility and redundancy
- Strong strategic carrier relationships
Read that list again. Of the four traits Daymon named, forecast precision is not one of them.
The cross-functional silo problem no forecast model fixes
A better forecast also assumes something that stopped being true: that planning, sourcing, manufacturing, logistics, and fulfillment are optimizing toward the same decision. Coty's chief global supply chain officer, Graeme Carter, told the same podcast that the traditional model, each function running its own objectives and its own decision-making process, is no longer sufficient given current volatility.
Here's what that looks like in practice. A demand signal can arrive with full confidence and still take three days to become a purchase order, a freight booking, and a production schedule change, because each of those five functions makes its own call on its own timeline. Forecast accuracy improves the signal. It does nothing to the four handoffs that follow it.
This is what decision speed actually depends on: a named owner for the decision that spans functions, with an approval path that doesn't wait on five separate sign-offs in five separate systems.
Call that role a Decision Owner, distinct from a Process Owner. A Process Owner keeps their own function running well. A Decision Owner is accountable for the decision landing fast, across whichever functions it touches. Most mid-market operations have plenty of the first and almost none of the second.
Forecast accuracy vs. decision speed: the operating model comparison
The mindset shift shows up in five concrete operating decisions, not one abstract principle.
| Operating dimension | Forecast-accuracy mindset | Decision-speed mindset |
|---|---|---|
| Primary investment | Model precision, error-rate reduction | Decision cycle time, exception routing |
| Trigger to act | Forecast confidence crosses a set threshold | Action window opens when the signal arrives, regardless of confidence |
| Owner of the exception | Planning team, working in isolation | Named owner spanning planning, logistics, and fulfillment |
| Response to a stale number | Wait for the next scheduled forecast cycle | Act on the best available signal, revise after |
| Cost of being wrong | Re-run the model | Re-route the decision, same day |
The forecast-accuracy column optimizes the moment before the decision. The decision-speed column optimizes the moment of the decision itself.
What decision speed actually requires under incomplete data
None of this argues for ignoring forecasts. A demand signal is still better than no signal. The argument is about where the next dollar of investment goes. Taking a demand model's MAPE from 18% to 16% is real work for a marginal return, most mid-market planning functions are already past the point where another point of accuracy moves the P&L. Taking a decision cycle from 72 hours to 6 does move it, because the cost of a slow decision compounds every day the market keeps shifting under it.
Recognizing the gap is straightforward. If a stockout risk, a rate spike, or a supplier delay sits in an inbox or a weekly S&OP deck for more than a day before someone with authority acts on it, the constraint isn't the forecast. It's routing and approval.
The Decision Latency Diagnostic scores that exact gap, signal to action, across five stages. No forecast data required, just how your team currently routes and approves an exception.
That is also what decision infrastructure is built to close: a named owner for the exception, an approval path that doesn't route through five separate systems, and an audit trail that survives the question of who decided what. OpsGrid applies that model to Business Central specifically, surfacing the exception inside Teams and requiring a human sign-off before anything writes back to the ERP. The forecast underneath it matters less than how fast the decision moves once that forecast, however good it is this quarter, turns out to be wrong.
Frequently asked questions
Does prioritizing decision speed mean abandoning forecast accuracy?
No. A demand signal is still more useful than no signal, and forecasting doesn't go away. The shift is about where incremental investment goes once a planning function has already captured most of the achievable forecast accuracy. Uber Freight's Bob Daymon named four 2026 differentiators, decision-making speed and agility, continuous improvement, operational flexibility and redundancy, and strategic carrier relationships, and forecast precision isn't one of them.
What is decision speed in supply chain operations?
Decision speed is the time between an operational signal landing, with full or partial confidence, and a qualified person committing to an action inside a bounded window, with that decision recorded. It's an execution metric, not a forecasting metric: it measures how fast an organization acts when certainty never fully arrives.
Why did Coty's supply chain leadership call the traditional planning model insufficient?
Coty's chief global supply chain officer, Graeme Carter, said the traditional model, planning, sourcing, manufacturing, logistics, and fulfillment each running independent objectives and decision-making processes, is no longer sufficient given current market volatility. The problem isn't any one function's forecast. It's that a decision spanning five functions has no single owner and no shared approval path, so accuracy in one function doesn't speed up the other four.
How does an operations leader measure decision speed today?
Track the time between when a signal, a stockout risk, a rate spike, a supplier delay, first appears in a system of record and when a qualified person commits to a response. Most mid-market operations have never measured this directly because it spans systems and functions with no single owner. The Decision Latency Diagnostic scores this gap across five governance stages in about three minutes, without requiring an existing measurement system.