Article

Manufacturing and industrial companies: get paid for your records

By TrustPast

Updated November 7, 2026

This page is for the owner or operations lead of a plant, a machine shop, a fabricator or a contract manufacturer with 30 or more people and years behind it. You know the machines have a price, and so does the order book. What most plants have never priced is the paper trail behind every part they ever shipped.

That trail is the quote, the traveler, the inspection sheet, the nonconformance report and the engineering change that followed it. Each one records a decision made on the floor, usually under time pressure, by someone who knew the machine. A buyer pays to license records like these to train AI models, and it pays for how the work was done, not for who the customer was. The general case, written for trade businesses and just as true of a shop floor, is on what data buyers actually want from a trade business.

Here is what a manufacturer keeps, what micro1 says counts, what stays out, how the names come off, and what moves the price.

The records a manufacturing and industrial company keeps

These are our examples, not a buyer’s list. They are what a plant produces as it runs, year after year, whatever it makes.

The request for quote and the quote. A customer sends a print and a quantity. Someone works out material, setups, cycle times, outside processing and scrap, and sends a price. Most quotes never turn into orders. The ones that did show how the shop priced risk: the tight tolerance that added an hour of inspection, the heat treat sent out, the minimum lot that made a small order worth taking. The lost ones show where the price was wrong.

The traveler and the work order. Once the purchase order lands, the job gets a routing: saw, mill, deburr, inspect, plate, pack. The traveler goes with the parts and collects initials, dates, counts and notes at each step. Ten years of travelers are a record of how long things really took against how long the quote said they would. The job files and work orders page shows which parts of a work order carry the most.

First article and inspection. The first-piece check, the in-process checks, the final inspection report and the calibration log for the gauges used. Each says what was measured, against what, and whether it passed.

The nonconformance report. This is where the plant writes down a problem. A bore out of tolerance, a burr on a sealing face, a lot of castings with porosity. The NCR says what was found, how many, and the disposition: scrap, rework, use as is with the customer’s sign-off, or return to the vendor. Behind each disposition sits the argument that produced it.

The corrective action. When the same defect shows up twice, someone opens one. The root-cause worksheet, the fishbone, the five whys, the fix and the check that the fix held. This is how a problem was diagnosed, in the plant’s own words, with the false leads left in.

The engineering change. A customer revises the print, or the floor finds a better way to hold the part. The change order records what moved, why, which lots it applies to and who approved it. Read in order, the change orders show how a product settled into its final form.

Maintenance. Work requests on a press, the preventive maintenance calendar, the downtime log, the note that says a spindle bearing was replaced after the third chatter complaint.

The project history. A new product launch, a tooling build, a line move, a new machine install. Each has a plan, a schedule that slipped, meeting notes, supplier emails and a sign-off. Our project history page covers how those files show the gap between the plan and what happened.

Here is how one part number runs through all of it, as an invented example. The quote went out in 2017. The order came. The first article failed on a thread depth. The NCR sent the lot to rework, and a corrective action changed the tap and the inspection plan. In 2019 the customer revised the print, and the quote went up for the new feature. That is six or seven files across three systems. Together they are a full account of how a shop learned to make one part. Now count the part numbers the plant has run.

Then there are the records nobody lists. The shift handover notes. The quality inbox. The supervisor’s spreadsheet of jobs that went sideways. The thread where the floor and the front office argued about whether a late order could ship partial. Owners forget these because nobody paid for them. They are often where the reasoning lives.

See your estimate

What counts, in micro1’s own words

micro1 publishes the companies it looks for. On its data partnerships page: “We’re looking for established companies with 30+ full-time employees, mature business operations, and substantial historical operational data.” The same answer goes on: “This can include internal communications, SOPs, project histories, CRM records, support tickets, and other business documentation.”

On the kind of data, the same page says: “We are looking for data that reflects how work is actually performed within an organization.” It names “standard operating procedures (SOPs), process documentation, internal knowledge bases, workflows, templates, operational frameworks, and other business process artifacts”. And it says “The most valuable datasets capture real operational expertise, decision-making processes, and repeatable workflows”.

micro1 doesn’t name manufacturers as a heading. It lists industry examples under “Operational data from every industry can contribute”, and the nearest to a plant is operations and logistics: “Operational SOPs, logistics processes, fulfillment workflows, inventory management, and internal operations”. Its customer support example includes “QA processes”. The referral page lists “All industries” among its company requirements.

Read that against a plant’s files, as our own reading. Work instructions are SOPs. The routing on a traveler is a repeatable workflow. An NCR disposition and a corrective action are decision-making processes, written down. The stockroom’s cycle counts are inventory management. micro1 decides what it licenses, but the words on its page describe what a manufacturer does every shift.

Two more lines matter for a plant. micro1 says “There’s no fixed minimum number of files, documents, or terabytes.” And “Companies with several years of records across multiple business systems are generally a stronger fit.” A shop with an ERP, a quality system, a maintenance system and a shared drive has several systems by definition. micro1 also says “Data should be primarily in English.” If the floor writes notes in two languages, our page on records not in English goes through what to raise.

What is never included

Customer names, the people on the payroll, payment details and anything personal stay out. That is the first thing a plant manager asks, and the answer is the same here as everywhere on this site. Our page on what is never included shows where each of those hides in a company’s files.

A manufacturer has two more lines to watch, as our own advice.

The first is the customer’s print. A contract manufacturer builds to someone else’s drawing, and the drawing, the specification and often the inspection plan belong to that customer. The supply agreement says what you may do with them. The traveler, the cycle times, the NCR and the way your shop solved the problem may be yours; the print usually isn’t. Our page on client-owned work and NDAs shows how to sort the files, and counsel has the final word.

The second is controlled technology. The Bureau of Industry and Security says it administers the Export Administration Regulations, “15 CFR parts 730-774, which govern the export, reexport, and transfer (in-country) of items “subject to the EAR.”” It says an item means “commodities, software, and technology”. The same page says the Department of State’s Directorate of Defense Trade Controls “has jurisdiction over defense articles and all defense services.” A plant that machines parts for defense or aerospace programmes may hold technical data under either regime. Keep any such files out of the inventory until counsel has looked at them. Ask counsel; we can’t tell you which of your files fall under either rule.

Then the plain ones. Employee files, injury and medical records, payroll and bank details stay where they are. micro1 says that, before any participation, the two sides confirm “U.S. and allied buyers only”, and that the data “is licensed exclusively to labs and enterprises based in the U.S. and allied countries”. That is worth knowing, and it is no substitute for counsel on controlled files.

How de-identification works

micro1 says you don’t do the scrubbing. Asked whether a company must anonymize its data before transferring it, its page answers: “No, micro1 has processes for de-identifying and transforming sensitive information.”

Its published confirmations, made before any participation, include “De-identified before sharing”, explained as “Personal and identifying information is removed or transformed before your data is shared with third parties”. And “Data segregation”: “Your data is kept separate from other companies’ data, with access restricted to authorized personnel”.

micro1 goes further on what happens after the data is shared. It says “personally identifiable information may be removed and certain datasets may undergo synthetic rewrites that preserve workflow structure while reducing direct ties to original records.” It says “companies are given the opportunity to review representative samples before prepared data is used.” And it says customer data is “processed through isolated pipelines that prevent it from being commingled with other engagements.”

Here is what that looks like on a traveler, as our own illustration. The customer’s name in the header, the buyer’s purchase order number and the operators’ initials are the identifying parts. The routing, the setup and run times, the counts at each step, the inspection result and the note that says the fixture was shimmed are the work. The first set is what the process exists to take off. The second is what a buyer is paying for. Our page on how de-identification works goes through the steps and what to ask about each.

What decides the price

micro1 publishes three tiers on its data partnerships page. $100K-500K for “Qualified Enterprise Data Partnerships” (indicative, not an offer). $500K-3M for “Large-scale operational datasets or ongoing participation across multiple teams and business functions” (indicative, not an offer). $3M+ for “Highly unique, proprietary operational data” (indicative, not an offer). Above the tiers it says “Compensation is based on the quality, uniqueness, and value of the data provided”.

The factors are named in its questions: “Factors considered include dataset size and volume, workflow complexity, data quality, domain expertise represented, and the uniqueness of the operational knowledge contained within the dataset.” And “Each partnership is evaluated individually”. micro1’s own estimate page asks three things: how many employees, how many years of data (“Typically the number of years you’ve been in business”), and where the company is headquartered. Under its result it says: “Indicative estimate only. Final payout is determined based on data sources, volume, quality, access terms, and due diligence.”

For a plant, as our own reading of those words:

Years. A shop that kept travelers and NCRs since 2009 has more historical depth than one that went paperless in 2022 and shredded the rest.
A complete trail. A quote that can be followed to its order, its traveler, its NCR and its change order is worth more than four piles that don’t connect. Workflow complexity is easiest to show when the links are there.
Domain expertise. A corrective action written by a quality engineer with twenty years on the floor carries judgment a clean production report doesn’t.
Systems. The ERP, the quality system, the maintenance log and the email are four sources, not one.

What lowers it is the reverse. Gaps from an ERP migration where the old data was never exported. Travelers thrown away after the retention period. A quality system that only kept the final disposition and none of the reasoning. If the plant changed systems, our page on whether old systems count says how to keep the old records before access ends.

On timing, micro1’s steps say: “We pay in full within 30 days of acceptance, with payments typically ranging from $100K to over $3M” (indicative, not an offer). On 9 October 2026, micro1 announced a commitment to spend $1 billion over the next 12 months acquiring and licensing enterprise operational data. More on each factor is on what decides the price of your records.

Which programme fits

This page is about micro1’s programme. Its line is 30 or more people, now or at peak, or a company of any size outside the US. micro1’s referral page lists the requirements for its enterprise partnership as “30+ employees”, “US and Western markets”, “English” and “All industries”. Count everyone: the floor, the office, the second shift. A plant that ran 80 people before a downturn and runs 25 now has the records of an 80-person company. Our estimate asks your headcount and names the programme that fits it.

If the plant has closed, the records haven’t. micro1’s referral page lists a Defunct Company Data Partnership: “We are acquiring data from companies that are no longer operating.” Its requirements are “30+ employees”, “10+ Countries”, “English” and “All industries”, and the company compensation it lists is $100K-$3M+ (indicative, not an offer) “for approved data packages”. The question for a closed plant is who can sign and where the servers went. Both are covered on the page for closed companies.

A publisher or education company keeps the same kind of production trail, a manuscript or a course moving through review and revision instead of a part moving through a routing. A hardware engineering company shares a plant’s change orders, test reports and controlled-data questions, one step earlier in the life of a product. And a SaaS and subscription company keeps its own version of the NCR and the corrective action: the support ticket and the incident review.

TrustPast

Questions and answers

Does micro1 take manufacturers?

micro1’s referral page lists “All industries” among its company requirements, next to 30 or more employees and English. Its data partnerships page doesn’t name manufacturing as a heading, but it names operational SOPs, inventory management and QA processes, which a plant keeps every day.

Do we have to hand over our customers’ drawings?

No. A print the customer owns stays the customer’s, and only the contract says what you may license. That is our advice, not micro1’s: have counsel read the supply agreements before contracting.

Our travelers are on paper. Do they count?

micro1 says there’s “no fixed minimum number of files, documents, or terabytes” and that it evaluates historical depth, breadth, quality and relevance. Scanned travelers are still records of how the work went; raise the paper on the call, and keep it until you have.

Do we need to scrub operator names first?

No. micro1 says it “has processes for de-identifying and transforming sensitive information”, and that personal and identifying information is removed or transformed before the data is shared with third parties.

How much does micro1 pay?

micro1’s data partnerships page lists three tiers, $100K-500K, $500K-3M and $3M+ (indicative, not an offer). micro1 prices each company, and our estimate is a range, not an offer.

We closed the plant last year. Is it too late?

Not under micro1’s published programmes. Its referral page lists a Defunct Company Data Partnership for “companies that are no longer operating”, with the same 30 or more employees line.