← Back to Field Notes

October 15, 2025 · The Bloomfield Team

9 Questions to Ask When OTD Drops Below 90%

On-time delivery at 90% means one in ten jobs ships late. For a shop running 200 orders per month, that is 20 late shipments. Twenty phone calls from buyers. Twenty opportunities for a customer to start looking at other vendors. The number matters because customers track it, even when you do not.

When OTD slides below 90%, the instinct is to push harder on the floor. Work overtime. Expedite material. Move jobs in the queue. These responses treat symptoms. The nine questions below treat causes.

1. Are your quoted lead times based on data or habit?

Most shops quote lead times the way they always have. Four weeks for standard work. Six weeks for complex. Eight for assemblies. These numbers may have been accurate three years ago when capacity utilization was 72%. At 88% utilization, the same shop needs longer lead times or lower WIP to maintain delivery performance. Pull your actual lead time data from the ERP for the past 12 months. Compare it to what you are quoting. If the gap is more than 15%, your delivery problems start at the quoting desk.

2. Where does the first delay happen in the job sequence?

Late jobs rarely become late at the last operation. The delay starts upstream. Material arrives two days after the scheduled start. The first operation waits in queue behind three rush jobs. Programming takes longer than expected because the geometry is new. By the time the job reaches final machining, the schedule gap has compounded and there is no buffer left. Track where in the operation sequence jobs first fall behind plan. The pattern will point you to the bottleneck.

3. How many rush orders disrupted the schedule this month?

Rush orders are necessary. They are also destructive. Every rush job that jumps the queue pushes two or three standard jobs back. If rush orders represent more than 15% of your monthly volume, your scheduling system is effectively running in reactive mode full time. Measure it. If the number is high, the question is whether you are pricing rush work high enough to compensate for the disruption it causes to everything else.

4. What is your actual setup time vs. your planned setup time?

Setup time is the most underestimated variable in production scheduling. A plan that allocates 45 minutes for setup on a VMC when the actual average is 90 minutes creates a cascading delay across every job that follows. Collect actual setup times for your top 20 job types across two weeks. Compare them to your planning assumptions. The gap is often 40 to 80%.

5. Is your scheduling system seeing the same reality as the shop floor?

If the production schedule lives in the ERP but the floor supervisor adjusts job sequence based on what they see at 6 AM, you have two schedules. The one in the system and the one that actually runs. That disconnect means every upstream decision, material release, tooling prep, outside processing coordination, is based on a plan that the floor has already changed. Either the schedule needs to reflect reality faster, or the floor needs to follow the schedule. The current state wastes planning effort and makes OTD unpredictable.

6. How often does material arrive late?

A job cannot start without material. Track the percentage of jobs where material arrives after the scheduled start date. In many shops, this number is between 8 and 15%. Each late material delivery pushes the job start and compresses the remaining production window. If material delays are a consistent contributor to late deliveries, the fix involves vendor management, safety stock decisions, and longer procurement lead times built into the schedule.

7. What percentage of your rework is unplanned?

Planned rework, like a known secondary operation, is already in the schedule. Unplanned rework, a tolerance miss that requires re-machining, a surface finish that needs re-grinding, destroys schedule integrity because it consumes machine time and operator time that was allocated to the next job. Track unplanned rework hours as a percentage of total production hours. If it exceeds 5%, the quality problem is also a delivery problem.

8. Do you have visibility into WIP across all operations?

A job has five operations. It finished Op 2 yesterday. Where is it right now? In many shops, answering that question requires walking the floor or calling the supervisor. Without real-time WIP visibility, you cannot identify at-risk jobs before they become late jobs. A production dashboard that shows job status by operation, with color-coded alerts for jobs falling behind plan, changes the conversation from reactive firefighting to proactive intervention.

9. Is anyone accountable for OTD as a metric?

In most small shops, OTD is everyone's problem and nobody's metric. The production manager, the scheduler, the estimator, and the shop foreman all affect delivery performance, but no single person owns the number, reviews it weekly, and drives corrective action when it drops. Assign ownership. Review OTD weekly with root cause analysis for every late shipment. The act of measuring and discussing it consistently is often enough to drive a 5 to 10 point improvement within a quarter.

The Pattern Behind the Questions

Late deliveries are the output of a system. Quoting, scheduling, material procurement, production execution, quality control, and communication all feed into whether a part ships on time. Fixing OTD means diagnosing which inputs are producing the late outputs, and that requires visibility into each stage of the process that most disconnected systems do not provide.

The shops running above 95% OTD consistently are the ones that have connected their data across these stages and can see a job's risk profile days or weeks before the ship date, not hours. That visibility is now buildable for shops at any size, using the data already sitting in your ERP, your shop floor systems, and your team's daily observations. For a broader view of how production visibility works, see our guide to production visibility.