← Back to Field Notes

Knowledge

February 17, 2026 · The Bloomfield Team

The Difference Between Documentation and Usable Knowledge

Every manufacturing operation we walk into has documentation. SOPs in three-ring binders. Setup sheets in filing cabinets. Process instructions laminated and zip-tied to machine guards. Work instructions in the quality management system. Inspection protocols. Material specifications. Customer requirements files.

Every one of those operations also has the same complaint: "We have the documentation, but nobody uses it. When something goes wrong, people ask the guy next to them."

The documentation exists. The knowledge does not flow. That gap is where most operational inefficiency in manufacturing lives, and more documentation cannot close it.

What Documentation Actually Is

Documentation is information written down and stored. An SOP for setting up a Haas VF-2 is documentation. A work instruction for incoming material inspection is documentation. A customer packaging requirement saved on the shared drive is documentation.

Documentation answers the question: "Does the information exist somewhere in this building?" In most operations, yes. The shop floor is not suffering from a lack of recorded information. Across the ERP, file server, quality system, email archives, and physical files, thousands of pages exist.

Documentation as typically created and stored in manufacturing has three properties that make it functionally useless at the moment of need.

Organized by document type, not by situation. Setup sheets filed with setup sheets. Quality procedures with quality procedures. Customer requirements by customer. When an operator needs to set up a job, they need information from all three simultaneously. Gathering it means visiting three locations, physical or digital, and assembling the pieces.

Static. An SOP from 2021 reflects how the shop ran in 2021. Tooling changed. Machines updated. Customer requirements evolved. Experienced operators know which parts of the SOP are current and which need mental revision. A new operator follows it exactly, which may or may not produce a good result.

Disconnected from the work. Documentation lives in a system. Work happens on the floor. The distance between them, measured in minutes of searching, a walk across the building, a phone call to the office, is where mistakes happen. When retrieving information costs more than guessing, people guess.

What Usable Knowledge Looks Like

Usable knowledge is information that reaches the right person, at the right moment, in a format they can act on without interrupting their work. The difference between a map filed in a drawer and directions spoken while you are driving.

Four characteristics define it in manufacturing.

Contextual. When an operator is about to set up Job #4417, the knowledge they need is specific to that job: setup notes from the last successful run, tooling used, cycle time achieved, quality issues that occurred, customer-specific requirements. Usable knowledge delivers all of this in one place, organized around the job.

Current. It reflects what happened last week, not three years ago. If the tooling approach changed on the last run, the system reflects that. If a customer added a new requirement after the last order, it is flagged.

Searchable by symptom. When something goes wrong, the operator knows the symptom: chatter, surface finish degradation, dimensional drift, tool breakage. They may not know the root cause. Usable knowledge lets them search by experience and returns history of similar problems on similar parts with causes and solutions. Manuals organized by category cannot do this.

Cumulative. Every job generates knowledge. What worked. What failed. What the customer accepted. What they rejected. Usable knowledge captures this automatically from job travelers, inspection reports, quality deviations, and operator notes. The knowledge base grows with the operation.

A Concrete Example

A shop runs 14 configurations of hydraulic valve bodies for an industrial equipment manufacturer. Nine years of production. Hundreds of job records.

Under documentation: setup sheets in a binder by part number. Quality requirements in a separate folder. Last known tooling list in the programmer's files. Special notes about material behavior, fixturing, customer preferences in the senior machinist's memory.

New order for configuration 7B. If the senior machinist is available, setup takes 90 minutes. If he is on vacation or retired, setup takes three and a half hours because the operator follows the setup sheet literally, does not know about the fixturing modification preventing chatter on the thin cross-section, and runs first parts at conservative parameters the senior machinist would have overridden.

Under usable knowledge: Job #4417 for configuration 7B arrives on the floor. The system presents the complete setup from the last successful run, including the fixturing modification with a note explaining why it was adopted in 2022. The tooling list with current tool numbers and life remaining. Cycle time from the last run with a note that the programmer optimized roughing in 2024, cutting 12 minutes. Customer inspection requirements including a dimensional check added after a 2023 rejection. A material note that heat lot variations from the current supplier occasionally require a feed rate reduction on the deep bore.

The operator, regardless of experience level, runs the job correctly. Setup takes 90 minutes whether the senior machinist is there or not.

Where the Knowledge Already Lives

Most knowledge a manufacturing operation needs is already being generated. In the ERP as job history, routing data, cost records. In the quality system as inspection reports, nonconformance records, corrective actions. In emails as customer communications, supplier quotes, internal discussions. On the floor as setup sheets, operator notes, institutional memory.

The systems storing it were not designed to deliver it contextually. The ERP manages transactions. The quality system ensures compliance. Email handles communication. None answer the question: "What does the person doing this job right now need to know?"

A custom knowledge system connects these sources and organizes information around the work being done. Estimators see pricing history and comparable jobs when opening an RFQ. Programmers see tooling selections and cycle times from similar parts. Operators beginning setup see the complete context of every previous run.

The Return on Usable Knowledge

Setup time reduction: 15% to 30%. Operators with full context from previous runs avoid trial-and-error on unfamiliar jobs. For a shop running two setups per machine per day across 15 machines, a 20% reduction in average setup time is 6 machine-hours daily. At a $125/hour loaded rate, roughly $190,000 per year.

Scrap reduction: 20% to 40%. Most scrap in custom manufacturing comes from the first few parts of a run when setup is being dialed in. When previous run knowledge is available, the learning curve compresses. For a shop with $200,000 in annual scrap cost, a 30% reduction saves $60,000.

Knowledge retention: compounding value. Every retirement that would have cost $2.4 million in lost knowledge becomes a manageable transition when knowledge lives in a system. The new person still develops skill. But they access the accumulated experience of everyone who came before, changing ramp-up from 24 months to 6 to 12.

The operation treating knowledge as a living system rather than a filing obligation builds an advantage that compounds with every job run, every problem solved, and every person trained.