Paper travelers vs. digital travelers: what to keep.

Every job shop has a traveler. It might be a manila folder, a clipboard, or a single sheet stapled to the drawing. It rides with the parts from saw to mill to weld to inspection, and it tells whoever picks it up what the job is, what to do next, and who signed off on the last step. Most of the advice you will read about travelers comes from people selling the software that replaces them. This post is not that. Paper does some things better than a screen, and a shop that throws it all out at once usually regrets it. The question is not paper or digital. The question is which parts of the traveler should move, in what order, and which parts should stay on paper for now.

What a paper traveler does well

Start with the honest case for paper, because it is a good one.

It is always available. A paper traveler needs no login, no battery, no Wi-Fi, and no training. A new hire can read it in the first hour. When the network drops, the job keeps moving. A recent experimental study of assembly work instructions put it plainly: the advantages of paper documentation are its easy availability, low technical requirements, and long use in industrial settings.^[1]

It is visual, and it lives with the work. The Lean Enterprise Institute defines visual management as placing tools, parts, production activities, and performance indicators in plain view, so the status of the system can be understood at a glance by everyone involved.^[2] A traveler clipped to a cart does exactly that. Toyota's kanban is the classic example: a card that travels with every item or box of items through production, comes off when the item is used, and goes back upstream as the order for more.^[3]

It points at one thing. Michael Ballé, writing for LEI, argues that posting more information is not visual management. To work as management, a visual has to be intuitive and point toward one specific thing.^[4] A traveler on top of a pallet says one thing: this job, this step, this is next.

It is cheap. A ream of paper and a printer. No per-seat license, no tablet that gets dropped in coolant, no IT ticket.

In my shops the traveler itself was never the problem. The problem was that what it carried lived nowhere else, so the office and the floor were never looking at the same numbers.

What paper cannot do

The same properties that make paper reliable at the workstation make it useless everywhere else.

Real-time status

A paper traveler knows where the job is only if you are standing next to it. The status lives wherever the paper lives. Toyota lists three principles of Just-in-Time, and the second is to not allow goods and information to be held up during production.^[5] Paper holds information up by design. It moves at the speed of the cart.

NIST describes the same problem at a larger scale. Information in most manufacturing operations flows in one direction, and gaps in the flow prevent the whole enterprise from using it.^[6] A traveler is the purest one-way flow there is. Information goes onto the sheet. It does not come back off until someone keys it in, if anyone ever does.

If you have read the real cost of chasing job status, you already know what that walking costs. The traveler is not the cause of the chasing. It is the reason the chasing is the only option.

Searchable history

Ask a paper shop a simple question: how many hours did we actually spend on the last five runs of this part number? The answer is in a filing cabinet, in five folders, in five sets of handwriting. The estimate for the next quote will still be a guess.

Paper also goes stale without telling you. Mark Hamel, an LEI faculty member, notes that visuals are often not maintained, and over time the workplace fills with unused or half-used visuals, which he calls visual pollution.^[7] A traveler from an old revision still looks like a traveler. Nothing about it says it is wrong.

Quality traceability

ISO 9000 defines traceability as the ability to trace the history, application, or location of an object, and notes that for a product this covers the origin of materials and parts, the processing history, and where the product went after delivery.^[8] Paper can hold all of that. The problem is retrieving it. When a customer calls about a cracked bracket from a lot shipped eight months ago, you need heat number, operator, machine, inspection result, and ship date in minutes, not in a day of pulling folders.

NIST's own history of measurement traceability makes a sharper point. Writing about calibration records, it observes that a paper trail tracing back to reference standards can be of dubious value in ensuring that actual measurements are valid, because auditors focused only on documentation and not on the whole measurement process.^[9] A signed box on a traveler proves someone signed the box. It does not prove the part was checked.

Transcription

Every number written on a traveler and later typed into a spreadsheet gets handled twice, and the second handling is where it breaks. The best measured data on this comes from healthcare, not from machine shops, so take it as a reference point rather than a shop-floor statistic. A 2019 study of clinic staff manually entering glucose results found that 3.7% of manual entries disagreed with the instrument's own value, and the authors note that most prior studies measured manual entry error rates in the low single-digit percentages.^[10] Apply a rate like that to quantities, hours, and lot numbers across a year of jobs and you get a lot of quiet errors.

What the evidence says about digital instructions

There is a growing body of controlled studies on paper versus digital work instructions, and the results lean toward digital, with caveats. In a 2026 experiment, five experienced assembly workers built an unfamiliar mechanism first from paper documentation and then from step-by-step digital instructions. Average assembly time fell from 1,084.8 seconds to 599.4 seconds, total errors fell from 56 to 9, and requests for help dropped from 29 to 1.^[1] The authors are careful to say the sample was small, the paper condition always came first so a learning effect cannot be ruled out, and the results should be read as an indication rather than a general conclusion.^[1] Digital instructions help most when the work is unfamiliar. For a 20-year operator running the same part for the 300th time, the sheet on the bench is fine.

A migration order that keeps the floor running

The MEP National Network's guide to Industry 4.0 for small and mid-size manufacturers gives the right frame. Start small. Starting small can mean collecting and interpreting data that is already available, which often proves its value immediately and removes much of the employee skepticism. Implementing in stages makes sense for smaller shops, and the savings from early steps fund the next ones.^[11] NIST MEP also lists the usual hurdles: choosing a platform, the up-front investment, and lack of coordination among departments.^[12] A phased order addresses all three.

Here is the order I use. It is built around one rule: digitize the data people chase, and leave on paper the data people only read at the workstation.

A few notes on each step.

Step 1 is the whole game. If you move only one thing, move status. A steel products maker I worked with had over $700,000 tied up in late orders. Six months later it was under $50,000, with no new system. Everyone looked at the same numbers every day. You cannot do that when the status lives on a sheet riding a cart. (For the bigger argument about ERP and the floor, see You don't need a new ERP.)

Step 2 feeds everything after it. Hours and counts per operation are the raw material for estimating, capacity, and the rework math. Capture them at the point of work. The moment you let them ride on paper to the office for later entry, you have bought a transcription step and all the errors that come with it.

Step 3 is where paper fails worst. Quality records on paper are fine until you need one. Move the records tied to customer traceability requirements first. If you run a mix of commodity work and aerospace or medical work, move the regulated jobs and leave the rest for a later phase.

Step 4 is about the master copy, not the medium. The failure mode with paper work instructions is not that they are paper. It is that nobody knows which of the three copies in circulation is current. Once the system holds the revision, a printed copy with a print date is safe to use. An undated copy is not.

Step 5 is the part most vendors skip. Keep the physical traveler. Print it from the system. It is the visual that tells the forklift driver which pallet goes to paint without anyone opening an app. The screen is for the scheduler. The paper is for the cart.

Keep the floor running while you do it

OMEP, the Oregon MEP center, recommends a parallel, phased, or piloted implementation to upgrade systems without stopping the production line, and warns that change should not be imposed. It should come out of a discovery process that involves shop floor employees along with everyone else.^[13] In practice that means:

• Pilot one cell or one product family, not the whole shop.
• Run paper and digital side by side for status only, for two to four weeks, then stop the paper status column. Do not run two systems of record for anything longer than that.
• Pick the operator who complains loudest about lost travelers. Put them on the pilot. They know where paper fails.
• Define what counts as done before you start: status updated within ten minutes of an operation finishing, for example. Measure that, not adoption percentages.

The Illinois MEP center makes the point that digital tools fit high-mix, low-volume shops in particular.^[14] A shop that runs the same part all year does not need searchable history. A shop that quotes 40 new parts a month lives on it.

Two mistakes to avoid

Buying the whole stack to fix step 1. A Pennsylvania manufacturer working with its MEP center described the typical small-shop setup before an ERP project: one system for CRM, one for manufacturing, one for finance, and none of them talking to each other.^[15] Adding a fourth to track travelers does not fix that. If you are going to add software, add it in the order above and make it carry the job number from your quoting and invoicing system from day one. What is an MES, and does a small shop need one covers where the line between tracking and a full execution system sits.

Forcing operators onto screens for data they only read. A screen at every machine is a cost and a maintenance burden. If the operator reads the instruction, prints it, and tapes it to the fixture, that is not a failure of adoption. That is the right tool. Put screens where data is entered, and let paper handle the reading.

This is the shape of the small-shop system HixonOps builds: work orders and inventory first, modules added one at a time, with travelers that print from the system so the cart and the screen agree. Details are at /#shop. Shops with an existing ERP that need floor status tied back into it are usually looking at the MES instead.

Where to get a second opinion

You do not have to figure this out alone or from a vendor demo. The MEP National Network has nearly 1,400 manufacturing advisors at more than 450 service locations across the U.S. and Puerto Rico,^[16] and MEP centers will help assess how digital tools fit an operation before anything is bought.^[14]

Paper is not the enemy. Scattered data is. The traveler on the cart is doing its job. The traveler in the filing cabinet is not doing anyone's. Move the parts people chase, keep the parts people read, and make sure the screen and the sheet carry the same job number. That is the whole plan.