What Manufacturers are Actually Struggling With

And Why the Answer Is Not More Spreadsheets

Based on real-world conversations with manufacturers across aerospace, defense, medical devices, electronics, and industrial production

We talk to a lot of manufacturers. Startups trying to get out of spreadsheets before a big customer finds out they are still in spreadsheets. Shops that have been doing it the same way for fifteen years because it worked well enough and nobody pushed hard enough to change it until recently. Operations running three facilities where the processes are different at each one and everyone has a reason why their way makes the most sense.

You hear the same things enough times, you start to notice.

• Nearly three-quarters were still running paper-based work instructions, travelers, or processes as their current operating reality.
• More than two-thirds had no MES in place at all or were running a homegrown workaround they knew was not sustainable.
• More than half described their ERP as the root of the problem: good software, simply the wrong tool for the shop floor.

None of this is unusual. LNS Research found that 42% of manufacturers cite workforce execution visibility as their top operational challenge. Once you have heard the same story from enough different people in enough different industries, it stops feeling like individual bad luck and starts feeling like something structural. Something in how manufacturing operations are typically set up that makes these problems almost inevitable.

What follows is an attempt to categorize them.

1. The ERP Stops at the Shop Floor Door

Why the system running your business cannot run your floor

Ask almost any plant manager what their ERP does well and you will get a consistent answer: financials, purchasing, scheduling at a high level. Ask what it does on the floor and the conversation gets quieter. That is not the ERP's fault. ERP systems were built to run businesses. They are very good at it. The problem is that running a business and running a factory floor are two different things, and for decades manufacturers have been asking one system to do both, filling the gaps with paper, spreadsheets, and people whose actual job has become moving information from one place to another.

One manufacturer running a major ERP platform described a system that handled everything up to the point a work order was released and nothing after. Work-in-process was not tracked. Work instructions were on paper. Drawings circulated as controlled documents, which created a constant problem around who was allowed to mark them up and which version was actually current. The manufacturing side of the operation had no digital backbone at all.

Another shop tried using their ERP's built-in manufacturing module. It worked in the demo. On the floor, the gaps between what the system expected and how production actually runs became apparent quickly. Time entries that could not be corrected after the fact. Material consumption that did not post to the right accounts. The workarounds required to bridge those gaps created more overhead than the system saved. After a few months they stepped back to paper travelers and manual entry. The digital transformation had resulted in both a system nobody used and the same paper process they had before.

A third manufacturer walked us through their process: sales order in, admin copies it to a spreadsheet, operators use the spreadsheet, admin copies results back into the ERP. The data moves through the same hands two, three, sometimes four times before it lands anywhere useful. Each pass is another chance to introduce an error. Nobody designed this process. It just accumulated.

Deloitte research found that 20% of frontline worker time is lost to activities like chasing updates, clarifying tasks, and duplicating effort. In operations built around manual re-entry and paper-to-system transcription, that is not a productivity problem. It is the process.

This is the space FactoryLogix was built for: everything that happens between a work order being released and a finished product being shipped. It connects to existing ERP systems, so neither has to be replaced, and it passes clean data back without anyone having to re-type it.

2. The Majority Are Still Running Paper

Why paper persists, and why the cost is so hard to see

Nearly three-quarters of the manufacturers in these conversations were running paper as their primary system for work instructions, quality records or both. Most of them knew it. What they had not done yet was put a real number on what it was costing, and that is partly by design. The cost of paper does not land in one place. It spreads thin across the operation: the manufacturing engineer who spends an hour updating documents before a revision can go to the floor, the operator building from a traveler printed before last week's change, the admin re-entering material consumption data on Friday that was already written down on paper Tuesday morning. None of those feels expensive on its own.

But nobody is adding them up. Because the cost is distributed, nobody carries it on a budget line and nobody is responsible for reducing it. It just gets absorbed.

One manufacturer had paper travelers, process control plans, inspection sheets, and material records that did not connect to each other in any meaningful way. Pulling a report or tracking down a unit's history meant physically locating documents and manually assembling the picture. That is not a workflow. It is archaeology.

Another manufacturer scaling up from low-volume engineering builds to full production had the same disconnect. Lot codes lived in the ERP. But when a material was applied to a specific part, the connection was written on paper and never digitized. Ask them which lot went into which assembly and they could tell you approximately. Not exactly. When a customer requires material traceability as a contract condition, approximately is not an answer.

One parts manufacturer described what closing out a job looks like in practice: admin staff goes back through the paper traveler and manually re-enters material consumption into the ERP. Incomplete stamps, entries that were skipped, data recorded without the work order context. Traceability evaporates regularly, and no single person is responsible because the problem is distributed across dozens of small steps, each one reasonable on its own, collectively a mess.

Deloitte research puts scrap and rework losses at 5% to 30% of total manufacturing cost depending on industry and process maturity. That range is wide because most of the variance lives on the floor, and without a way to capture what is actually happening out there, it stays unaccounted for.

The trouble with paper is not that it fails in a way anyone notices. It fails in the margins of every shift, absorbed into the routine until someone finally forces the question of what all that manual effort is actually worth. Most manufacturers have not forced that question yet. The ones who have tend not to go back.

3. Most Have No MES, or Built Their Own Out of Necessity

Why homegrown systems eventually meet their limit

More than two-thirds of the companies in these conversations either had no MES or were running something they built themselves. Spreadsheet macros. Shared folders with naming conventions only two people understand. Home-grown tracking tools that one engineer maintains in their spare time, usually the same engineer who built it, because nobody else knows how it works. These were not mistakes. They were reasonable solutions for a smaller operation, and they worked until the company outgrew them. The problem is that the moment they stop working tends to coincide with exactly the wrong time: a new contract, a customer audit, a second facility, a key employee leaving.

One manufacturer described their tracking system honestly. It was a homegrown tool built as a short-term fix, and everyone knew it. Making changes meant touching code and hoping nothing else broke. They were clear that it was never meant to be permanent. Scaling it to higher volumes and a larger team of technicians was not going to work. They just had not found the replacement yet. That last part, not having found the replacement yet, is where a lot of manufacturers sit for longer than they intend to.

A company early in its manufacturing journey put the core issue directly. The work instruction lives in their head. Fine at low volumes with an engineering team doing the builds. Not fine when you start hiring technicians and need them to produce the same result consistently on day one.

McKinsey estimates disconnected operations cost between $17,000 and $30,000 per frontline employee per year. Run that against a 50-person floor and you are looking at somewhere north of $850,000 annually, possibly closer to $1.5 million, in friction costs that never appear as a single line item anywhere.

What most homegrown systems cannot replicate is the concept that traceability, production tracking, and process control do not need to be separate activities layered on top of the work. In a purpose-built MES, they should just be byproducts of running the system. With a single scan of a unit, the operator sees the right revision of the right documents for their station, the system confirms their certifications and verifies materials, and the build record compiles in the background. The operator is not doing anything extra. The business is getting what it needs without anyone realizing it is occurring. That is a fundamentally different starting point than any spreadsheet or homegrown tool can offer.

4. Majority Do Not Know Where the Jobs Are

Roughly a third of the companies in these conversations had the same floor-level problem. They did not know, in real time, where their jobs were.

One capital equipment manufacturer held daily status meetings where planners reported what they knew, then spent the rest of the day walking the building to find what they did not. Finding things and pushing them through had become the job. Strategic thinking was not possible in that environment. They had hired planners specifically to chase jobs around the shop floor. That is a staffing solution to a systems problem.

Another manufacturer running continuous production described what happens when a job gets expedited: the tracking steps get skipped under pressure, the part disappears into the process, and the last recorded location is wrong. Meanwhile a final assembly sits incomplete because nobody can find the part it needs. The clock runs. Eventually the part gets scrapped and rebuilt because that is faster than finding it. A good part, scrapped, because nothing in the system could account for where it actually went.

The materials problem and the WIP problem are usually the same problem. Without a live view of what is happening on the floor, planning is reactive. Planners over-kit to protect themselves against uncertainty they cannot see. Buffer stock climbs. The warehouse carries more than it should while the line still stops waiting for the right component at the right time. It is an expensive way to manage uncertainty, and it does not actually solve the uncertainty.

ISA research estimates unplanned downtime destroys between 5% and 20% of annual productivity on average. For most facilities, that loss does not appear on a downtime report. It shows up scattered across cycle time numbers, missed ship dates, and overtime authorizations.

FactoryLogix addresses the materials side through a pull-based replenishment model that most ERP systems cannot replicate because they do not have access to the detailed real-time floor data required to make it work. The system knows actual consumption rates at each station, run rates of each asset, when jobs will complete, and how long it takes to move material from the warehouse to the line. It uses all of that to generate replenishment signals that keep lines running without overstocking.

There is also a less obvious problem that shows up in operations without live WIP tracking: duplicate assemblies. A job is already being built somewhere on the floor, but because nobody can see it is in process, it gets re-kitted at another station and started again. Double the materials consumed, double the labor, and often a partially completed unit that has to be scrapped when the duplication is eventually discovered. It happens more than most manufacturers expect, and it is almost entirely a visibility problem. When customers have addressed WIP tracking through FactoryLogix, they have reported a 13% reduction in unplanned downtime and 100% elimination of duplicate assemblies.

5. Material Traceability Is Now a Customer Requirement

What customers and regulators are asking for

Nearly a third of the companies in these conversations were dealing with traceability not as an internal goal but as an external requirement. Customers asking for it. Regulators expecting it. Contracts specifying it.

A manufacturer serving defense and aerospace supply chains described what newer customers want: a report that traces a component's certificate of conformance all the way through the build to the finished unit's serial number and out to shipping. They described it simply as needing to know the complete lineage of every part, where each component came from, where it went, and what happened to it along the way.

Another manufacturer described the stakes plainly. When the product cannot be retrieved, the data is the only record that exists. No physical unit to examine. No way to reconstruct what happened without a complete build history. The record either exists or it does not.

One manufacturer moving from R&D into production had lot codes in the ERP but no digital link between those codes and specific parts. The connection was handwritten on the traveler and never captured anywhere searchable. They acknowledged there was no way to digitally connect a material lot to the specific parts it went into. Fine internally. An urgent problem the moment a customer asks for it.

The National Association of Manufacturers estimates non-compliance fines and delayed audit costs run $50,000 or more per employee annually in regulated industries. Incomplete traceability is among the most common sources of those findings.

With FactoryLogix, there is no separate traceability process. The complete build record, components consumed, lot codes, sub-assembly serial numbers, tools validated, operators at each step, test results throughout, is compiled in the background as work happens. Operators are simply doing their jobs. The device history record, the compliance reports, all of it is a natural output of production rather than a separate documentation exercise. And the lookup works in both directions: start with a material lot number and the system shows every unit that material touched; start with a serial number and the system shows everything that went into that unit. Customers have reported a 59% reduction in audit preparation time, and reporting that previously consumed significant hours each week now runs automatically.

6. Quality Is Still Being Managed After the Fact

Why reactive quality is a design problem, not a performance problem

A consistent share of companies described quality processes that were reactive by design. Inspect at intervals. Something fails. A human decides what to do. By the time that decision is made, more units have moved on.

One manufacturer with an automated assembly line described their standard process: inspect at defined intervals, and if something fails, manually identify and pull the affected units, then re-inspect. Every step of that, the decision about scope, the tracking, the rework routing, was done by hand. That was not an edge case. That was Tuesday.

Another manufacturer described what happens when a defect threshold is crossed and there is no automated response: someone gets a notification, acknowledges it, and moves on. The defect gets logged. Whether it gets investigated, whether root cause gets documented, whether anything changes as a result depends entirely on whether the right person remembers to follow up. Often, they do not. The same defect reappears the following week.

A high-volume production manufacturer described a quality process where every defect was discovered downstream of where it happened, never at the source. By the time a defect was identified, the affected units had often already moved through additional operations. The cost compounds at every step downstream of where it started.

The reason reactive quality persists is structural, not cultural. When inspection is a separate activity from production it will always lag production. You cannot inspect your way to quality. The only way to fundamentally change the outcome is to make quality checks part of the work itself, so that the two happen simultaneously rather than sequentially.

The American Society for Quality reports that 33% of quality-related problems in manufacturing are due to human error not because people are careless, but because ambiguous instructions, inconsistent processes, and decisions made without the right information at the right moment make errors almost inevitable.

FactoryLogix approaches this by making defect capture part of the work rather than a step that follows it. Operators see a visual of the actual product and tap on the part that has a problem. A filtered list of defect codes appears, showing only what is relevant at that specific point in the process. Defects that cannot physically occur at this stage simply do not appear, which means operators are not scrolling through irrelevant options or guessing. When a threshold is crossed the system reacts without waiting for a supervisor to notice: a minor issue triggers a rework route, a more serious one opens a structured corrective action workflow. Quality engineering sets the thresholds ahead of time. Everything else runs automatically. Customers have reported a 50% reduction in rework time and an 80% reduction in defective parts per million.

7. The Tribal Knowledge Problem Gets Worse Before It Gets Better

What happens when the people are the process

This one shows up in almost every conversation, just in different forms.

A key technician who is the only person who knows how a particular assembly goes together. A process that works because the right three people happen to still be employed. A facility that has been running the same way for years, and nobody has written any of it down because nobody needed to. Until now they do, and the ask of writing it all down feels impossibly large, so it keeps getting deferred.

One manufacturer that had grown significantly described their current state: assembly procedures written in Word documents, photos taken on the floor, methods gathered by interviewing technicians, tool information tracked in a separate spreadsheet. A scrappy system that worked for where they were. It would not survive the growth they were planning, or a serious customer audit, and they knew that too.

Another manufacturer whose production team was expanding well beyond its founding engineers described the inflection point clearly. As long as the engineers who designed the product were the ones building it, informal knowledge transfer worked well enough. The moment they started bringing in technicians who had not been part of the development process, something documented and accessible was needed. That is the exact moment the informal system stops being an asset and starts being a liability. It does not happen gradually. It tends to happen at once, usually right when you can least afford it.

There is also a separation that most manufacturers miss between process revision and product revision. A product might stay at the same engineering revision for two years while the way it is built changes multiple times, as lines get automated, methods improve, or new tooling is introduced. In most informal systems those two things get conflated, which means process improvements either get tangled up in engineering change processes or simply never get formally captured at all. The work gets done better. Nobody writes it down. Six months later a new operator builds it the old way and nobody knows why the results are different.

Deloitte research found manufacturers lose an average of 32 working days per year per employee from toggling between systems to find necessary information. When process knowledge lives in people's heads or scattered across documents and drives, that number climbs further.

And one manufacturer trying to open a second facility found out the hard way: the process they thought they had documented well enough to replicate was not really documented at all. It was just the way certain people did things. When those people were not there, neither was the process. Replication is the test that most informal systems fail.

What This Adds Up To

Across a wide range of real-world manufacturer conversations, spanning industries and operational complexity, the same challenges kept surfacing:

• The majority were running paper-based processes as their current operating reality.
• More than two-thirds had no MES in place or were running something they knew was not sustainable.
• More than half said their ERP was the root of the problem, not for lack of capability in its own domain, but because it was never built for the shop floor.
• Roughly a third had no reliable way to know where their jobs were on the floor at any given moment.
• Nearly a third were responding to traceability requirements from customers or regulators.
• A significant share had no documented processes. The knowledge existed only in the people who had been there long enough to accumulate it.
• Many were running quality processes that caught problems after the damage was already done.

No two conversations were identical. Different industries, different products, different scale. But the underlying problem was consistent: a disconnect between what is happening on the production floor and what is being recorded anywhere, filled by paper, spreadsheets, manual effort, and people who have been there long enough to know how everything works.

The ERP handles the business transactions. Engineering systems manage product data. But the space where the actual work happens stays largely unmanaged for most manufacturers. YouGov research puts the global annual loss from inefficient tools and fragmented execution at $162 billion. That number is abstract until you map it to your own floor: the planner walking the building to find a job, the operator building from last month's instructions, the admin re-entering data that was already captured on paper three hours ago.

The manufacturers who address this do not just learn where their jobs are. They build the data foundation that makes improvement possible for the first time.

The question is not whether you can afford to address it. It is whether you can afford to keep not addressing it.

Statistics attributed to third-party sources: LNS Research; Deloitte; McKinsey; ISA (International Society of Automation); National Association of Manufacturers; YouGov; MESA.org and Tech-Clarity. FactoryLogix outcome statistics reflect documented customer results.