Supply chain and operations teams rarely suffer from a shortage of metrics. Most planners can recite their OTIF, inventory turns, forecast accuracy and cost per order from memory. What they tend to lack is a way of deciding which number is not good enough, and then organising a quarter around moving it. That is the gap Objectives and Key Results are designed to fill.
The distinction is easiest to hold onto as a pair of words. A KPI measures health: it tells you, continuously, whether the process is running as it should. An OKR describes change: a quarterly commitment to take one of those metrics from where it is to where it needs to be. Most published sets of OKRs for different functions lean towards sales and product teams, so the three examples below speak the language of the warehouse and the planning meeting.
KPI or OKR: which is which
A useful test is to ask whether the number would still matter if nothing about it changed this quarter. OTIF matters every month regardless, so it is a KPI. “Raise OTIF from 91 % to 96 % by the end of Q4” is a decision to change it, so it is a Key Result under an Objective about reliability. When the quarter ends, the metric goes back to being a KPI at its new level, and the team picks something else to move.
The SCOR model from ASCM is a sensible reference for choosing which metrics to standardise on, and it rests on the same principle: a good KPI is chosen because it reflects an outcome the organisation cares about, not because it is easy to count. The same applies when choosing which KPI deserves an OKR: pick the one costing you customers, cash or credibility, not the one with the tidiest data.
Example one: on-time in full
A team running at 91 % OTIF against a customer expectation of 95 % has a problem it can name. The Objective might read: “Customers can rely on our delivery promise.” The Key Results: raise OTIF from 91 % to 96 %; cut the share of late orders caused by picking errors from 40 % of all late orders to 15 %; and reduce the average time to confirm a delivery date from 48 hours to 8 hours. The second and third results are the levers the team believes will move the first.
Example two: forecast accuracy and a new S&OP cadence
Forecast accuracy at 68 % usually means safety stock is doing the job that planning should be doing. Here the Objective could be: “We plan from a forecast the business actually trusts.” Key Results: improve forecast accuracy at SKU level from 68 % to 80 %; hold a monthly S&OP review with sales, finance and operations in the room, up from an ad hoc call; and reduce the number of SKUs with manual forecast overrides from 220 to 90. The S&OP cadence is the process change; the accuracy number is the evidence that it is working.
Example three: inventory quality
A single turns target hides where an inventory problem sits. A sharper Objective is: “Our stock is the stock customers want.” Key Results: reduce slow-moving stock from 22 % to 12 % of inventory value; cut stock-outs on A items from 14 to 4 per month; and raise cycle-count accuracy on A items from 94 % to 99 %. Two of these pull in opposite directions, which is deliberate: a team that only cut slow movers could starve the fast ones.
The weekly check-in is the method
What separates an OKR from a spreadsheet target is a short weekly conversation, twenty minutes at most, where the owner of each Key Result says where the number is, whether it is on track or behind, and what is blocking it. A forecast accuracy figure drifting the wrong way in week three then gets a decision in week three, rather than a post-mortem in month four. AWS guidance on establishing KPIs, though written for engineering teams, describes the habit: agree the measure, agree who owns it, and review it on a schedule.
Where OKRs do not fit
Do not write an OKR for a process that is already at target. If OTIF has sat at 97 % for a year and customers are content, it needs a KPI with an alert threshold and nothing more; forcing a stretch goal onto it wastes attention and invites gaming. The illustrative figures above each start from a level the team has decided is not acceptable.
Safety and compliance are the other exclusion, an absolute one. Lost-time incidents, regulatory audit findings and food safety non-conformances are never stretch goals. They are floors that must be held every day, and framing them as something to improve by a percentage implies the current level is negotiable. Keep those on the KPI dashboard with a hard limit, and spend the OKR on the number that a good quarter of focused work could genuinely change.
Frequently Asked Questions FAQs
What is the main difference between a KPI and an OKR in supply chain management?
A KPI (Key Performance Indicator) measures the ongoing health of a process to ensure it is running as it should. An OKR (Objective and Key Results) describes a specific, quarterly commitment to change or improve a metric from its current state to a desired target.
How do I decide whether a metric should be tracked as a KPI or an OKR?
Ask yourself if the number would still matter if nothing changed this quarter. If it matters regardless of change, it’s a KPI. If you are actively setting a goal to improve that specific metric over the next quarter, it becomes a Key Result under an OKR. Once the quarter ends and the new level is reached, it reverts to being a KPI.
What is an example of an OKR for improving On-Time In-Full (OTIF) performance?
An Objective could be “Customers can rely on our delivery promise.” The Key Results to support this might include raising OTIF from 91% to 96%, cutting picking-error-related late orders from 40% to 15%, and reducing average delivery date confirmation time from 48 hours to 8 hours.
What is an example of an OKR for improving inventory quality?
An Objective could be “Our stock is the stock customers want.” Key Results might include reducing slow-moving stock from 22% to 12% of inventory value, cutting stock-outs on “A” items from 14 to 4 per month, and raising cycle-count accuracy on “A” items from 94% to 99%.
How should a team manage and track its OKRs throughout the quarter?
Teams should hold a short, weekly check-in (around 20 minutes) where the owner of each Key Result reports on the current numbers, states whether the goal is on track or behind, and identifies any blockers. This prevents surprises and allows for immediate course correction.
When should a team avoid using OKRs?
Do not write OKRs for processes that are already consistently hitting their targets; a simple KPI with an alert threshold is sufficient. Additionally, never use OKRs for safety, compliance, or regulatory requirements. These are absolute standards that must be maintained daily, not negotiable stretch goals.
