The Shop Floor Answer

Can the way you pay people cause your operation to fail? Most risk registers list machines, suppliers and people. Almost none list the bonus scheme.

In this episode, we look at remuneration policy as a source of organisational failure. Pay tells people what the business actually wants, whatever the poster on the wall says. If the scheme rewards one number, people will deliver that number, even when it conflicts with what the operation can safely deliver. We start with two examples from the operations management textbook: a salesperson rewarded only for closing orders, who promises delivery dates and specifications the plant cannot meet, and investment bankers rewarded for short-term return while risk goes unmeasured.

Then a real case: Wells Fargo in 2016. US regulators fined the bank $185 million after finding that thousands of employees opened accounts in secret to hit sales targets and earn bonuses. More than 2 million accounts may have been opened without customers’ permission, and 5,300 employees were fired. It is a bank, not a factory, but the mechanism is the same one that shows up on any production line: reward the count, and the outcome suffers.

We finish with three questions you can use to stress test any target or bonus in your own operation.

For plant managers, production supervisors, operations leaders and anyone who designs or inherits performance targets.

Ask yourself one question: who in your operation is genuinely irreplaceable? Most managers name the most senior person. That’s usually the wrong answer.

In this episode, we look at human failure as a source of operational risk, and specifically at the loss of a critical person through resignation, illness or absence. The key insight is that the critical person is often not the one with the biggest title. It’s whoever holds tacit knowledge: the undocumented understanding of how a particular machine actually behaves, which setting needs a small adjustment on a cold morning, or which sound means trouble. None of it is in a manual, and all of it walks out the door with them.

We cover why this belongs in your risk register rather than being treated as an HR issue, and how it connects to the other human source of failure: errors and violations by people who are present and working. Then we work through a simple test: which knowledge in your operation exists in exactly one head, and what would happen next Monday if that person wasn’t available?

The practical takeaway: list your three most critical machines or processes, write down who really knows how to run each one when something goes wrong, and count how many names appear only once. Every single-name entry is a risk you can reduce before it becomes a failure.

For plant managers, production supervisors, maintenance leads, and anyone responsible for keeping an operation running when key people are unavailable.

Tags/keywords: key person risk, tacit knowledge, knowledge transfer, operational risk, risk management, succession planning, manufacturing operations, human factors

Episode description:

Most people in manufacturing have heard of Six Sigma. Far fewer can say what the “six” actually refers to without reaching for a textbook. This episode explains it in plain language.

Sigma is simply standard deviation, the measure of how much a process varies. The idea behind Six Sigma is margin: the customer’s specification window should be six standard deviations wide on each side of your process centre. That means your process can drift and still produce good product. It isn’t luck, it’s built-in room to move.

Then we cover the distinction that trips most people up: defective units versus defects. One claim, part or order can carry several defects, and counting them differently changes the story. We work through a real insurance claims case step by step: 300 claims sampled, 51 defective, 74 defects in total, and 4 ways each claim could go wrong. By the end, that works out to roughly 62,000 defects per million opportunities (DPMO). We also explain why DPMO exists: it lets you compare a simple process to a complex one fairly, which a raw defect percentage never can.

We also take the criticism seriously. Critics say Six Sigma repackaged existing ideas, such as improvement cycles, statistical process control and data analysis, under martial arts branding. The fair counter-question is this: if the combination works as a problem-solving system, does the packaging matter?

The practical takeaway: pick one process in your operation and count the number of ways it can go wrong. That number, not the defect percentage, is the first step to measuring it fairly.

For quality engineers, production supervisors, Six Sigma belt candidates, and anyone who has nodded along in a Six Sigma meeting without being sure what the numbers meant.

Tags/keywords: Six Sigma, DPMO, defects per million opportunities, standard deviation, process capability, quality management, Lean Six Sigma, manufacturing quality

Episode description:

Count how many departments in your business speak to the same customer. If the answer is more than one, who actually owns that relationship? Often, nobody does.

In this episode, we break down the difference between function-based and process-based organisation, using the Business Process Re-engineering (BPR) case of a trading company that buys consumer goods, warehouses them, and sells to retailers. Five departments touched the flow: purchasing, marketing, sales, the warehouse, and finance. Three of them spoke to the customer. Two of them spoke to the supplier. Every handoff was a chance for an error or a miscommunication, and nobody owned either relationship from start to finish.

We walk through how the company redesigned around two end-to-end processes instead: one owning everything to do with suppliers, and one owning everything to do with customers, speaking with a single voice.

Then we give the necessary criticism. BPR started with a bold idea: don’t automate the old work, obliterate it. But in practice it often became a respectable cover story for downsizing, with most projects cutting more than 20% of staff. It also carries a quieter risk: if you redesign radically and cut deeply at the same time, you can remove the experienced people who knew how to handle the unexpected.

The practical takeaway: pick one customer complaint from last month and trace how many people and departments it passed through. If you can’t name a single person who owned it from start to finish, you’ve found your first process to redesign.

For plant managers, supply chain leads, operations managers, and anyone who has watched a customer request get lost between departments.

Tags/keywords: business process re-engineering, BPR, process-based organisation, end-to-end process, customer service, supply chain management, operations management, organisational design

A customer has a problem, complains, and your team fobs them off. Is that worse than if they'd had the exact same problem and never said a word? The research says yes — and the gap is bigger than most people expect.

In this episode, we break down four scenarios and rank them by customer loyalty: no problem at all ranks highest, unsurprisingly. But second isn't "no complaint" — it's a customer whose complaint was properly and fully resolved. Below that sits the customer who had a problem and stayed silent. And at the very bottom is the customer who complained and was dismissed.

The implication most teams miss: a service failure isn't automatically a loss. Handled well, it can produce more loyalty than if nothing had gone wrong at all. Handled badly, it does more damage than the original failure ever could. Dissatisfaction, in other words, often has less to do with the failure itself and more to do with how the organization responded to it.

We also walk through a practical four-step recovery framework — discover what happened and who it affects, act to contain it and inform the right people, learn the actual root cause rather than the convenient one, and plan the fix into the process so it doesn't repeat.

The practical takeaway: the next complaint that lands on your desk isn't a problem to make disappear quietly — it's the single best chance you'll get to turn a dissatisfied customer into your most loyal one.

For plant managers, quality leads, customer service teams, and anyone responsible for handling service failures or customer complaints.

Tags/keywords: service recovery, customer complaint handling, customer loyalty, quality management, service failure, customer satisfaction, operations management, complaint resolution

Of ten possible failures in your process, which one do you fix first? Most people instinctively pick the one with the worst consequences. FMEA often points somewhere else entirely — and the reason is the one factor everyone forgets to weigh.

In this episode, we break down Failure Mode and Effect Analysis: scoring every potential failure on three things — how likely it is to happen, how severe it would be if it did, and how likely it is to be caught before it reaches the customer. Multiply the three scores together and you get a Risk Priority Number, and that number — not severity alone — is what should actually drive your priority list.

We work through a real transport case with three failure modes behind the same complaint: damaged goods. One mode scored highest on severity, another on likelihood — but the one that won top priority wasn't either of those. It was the mode that was hardest to detect. A failure that announces itself gets caught and fixed. A failure that slips through quietly is the one that actually costs you.

The practical takeaway: next time you're ranking risks or defects for action, don't stop at "how bad is it" — ask "how likely are we to even notice it's happening" before it reaches a customer.

For quality engineers, plant managers, production supervisors, and anyone running root cause or risk prioritization in a manufacturing environment.

Tags/keywords: FMEA, failure mode and effect analysis, risk priority number, quality management, root cause analysis, risk management, Six Sigma, manufacturing quality control

Lean thinking says inventory is bad — it hides problems instead of solving them. Risk management thinking says thin inventory is dangerous — it removes the cushion that protects you from disruption. Both are right. This episode doesn't pick a side.

We start with the river-and-rocks analogy: inventory is the water level, problems are the rocks sitting on the riverbed. Keep the water high and the rocks stay hidden — production keeps flowing, but nothing gets fixed. Deliberately lower the water and the rocks are exposed, forcing you to deal with the actual defects, breakdowns, and delays you'd otherwise never see. From there we cover the seven wastes, with overproduction singled out as the root cause behind several of the others.

Then we take the opposite argument seriously: in recent years, supply chains stripped of inventory buffers have proven far more fragile to disruption, because the cushion that used to absorb a late shipment or a supplier failure simply isn't there anymore. Efficiency and resilience pull in different directions, and neither the lean argument nor the risk argument is complete on its own.

The practical takeaway: the real question isn't "how much inventory should we hold" in the abstract — it's which specific problems you want exposed and fixed, and which specific risks you genuinely can't afford to be caught by. Those two answers, for most operations, aren't the same number.

For plant managers, supply chain planners, production engineers, and anyone weighing lean inventory reduction against supply chain risk.

Tags/keywords: lean manufacturing, inventory management, waste elimination, supply chain risk, JIT production, operations strategy, supply chain resilience, seven wastes

Every plant is under pressure to cut cost. It's usually the first thing leadership asks for. The sandcone theory argues that's exactly backwards — and that cutting cost first is often why the improvement doesn't hold.

In this episode, we break down the idea that operational improvement has a natural order: quality first, as the foundation everything else depends on, then dependability, then speed, then flexibility — and only then cost. Crucially, it's cumulative, not sequential. Moving to the next layer doesn't mean abandoning the one before it; each new layer sits on top of the last, like sand poured into a cone.

We work through why this order makes sense operationally — an operation with poor quality wastes capacity on rework and scrap before speed or cost improvements can even matter, and why chasing cost reduction on an unstable base tends to produce short-lived gains that quietly reverse.

We also give the honest counter-argument: business pressure rarely allows the luxury of sequencing improvements one at a time, and leadership often wants cost addressed immediately regardless of what theory recommends. This episode doesn't pretend that tension away.

The practical takeaway: before starting your next improvement initiative, ask which layer of the sandcone your operation is actually missing — because fixing cost on top of a quality problem rarely produces a result that lasts.

For plant managers, production engineers, operations leaders, and anyone deciding where to focus limited improvement resources.

Tags/keywords: sandcone theory, operations strategy, continuous improvement, quality management, cost reduction, manufacturing priorities, operational excellence, lean manufacturing

A supervisor reports 83% on-time delivery in a review meeting. Is that a win or a failure? The honest answer: it depends entirely on which target you're comparing it against — and most people never ask which one is being used.

In this episode, we break down the four types of performance target — historical (compared to last year), strategic (compared to what the business decided it needs), competitor-based (compared to rivals), and absolute (compared to a theoretical ideal). Then we run the same 83% number against all four: excellent against a 60% baseline from last year, poor against a 95% strategic goal, good against competitors sitting at 80%, and a clear failure against a 100% absolute standard. Same number, four completely different verdicts.

We also cover how this gets used — deliberately or not — in management reporting, and why the choice of comparison point can make a struggling operation look fine, or a genuinely improving one look like it's failing.

The practical takeaway: the next time you see a performance number in a meeting, the first question worth asking isn't whether it's good — it's which target it's being measured against, and who chose that target.

For production supervisors, plant managers, quality leads, and anyone who reports or reviews performance metrics in a manufacturing or operations setting.

Tags/keywords: performance targets, KPI benchmarking, operations metrics, delivery performance, performance measurement, manufacturing KPIs, operations management, plant performance reporting

Every plant manager has heard the same advice: find the best practice, copy it, implement it. It sounds like the safest possible strategy. It may also be the reason your operation never gets ahead of anyone.

In this episode, we break down what benchmarking is genuinely good for — stimulating fresh thinking and importing ideas from operations completely outside your own industry. Then we go straight to the uncomfortable part: best practice isn't "best" in the sense that it can't be bettered. It's only the best that someone happened to find and publish. Copy it exactly, and the ceiling you've just built for yourself is someone else's old result.

We also cover why a practice that works brilliantly in one operation can fail outright in another — differences in staff skill level, differences in strategic priority, differences in what the operation is actually optimizing for — and why benchmarking without understanding why a practice works is really just copying, not improving.

The practical takeaway: before adopting any "best practice" from a case study, a competitor, or a consultant, ask what specific conditions made it work there — and whether those conditions exist in your operation at all.

For plant managers, production engineers, quality leads, and anyone evaluating whether an outside "best practice" is actually right for their own operation.

Tags/keywords: benchmarking, best practice, continuous improvement, operations strategy, process improvement, lean manufacturing, manufacturing benchmarking, operational excellence