The Friction Audit. How to Find What You've Stopped Seeing.

Taiichi Ohno, the architect of the Toyota Production System, had a training exercise for new managers. He'd draw a chalk circle on the factory floor, tell the manager to stand in it, and watch. Just watch. Don't move. Don't fix anything. Don't talk to anyone. Stand there until you can see the waste.

Some managers stood for hours before they could see what was in front of them.

The waste wasn't hidden. It was normalised. The team had been working around it so long that it looked like "how things work." It took a fresh pair of eyes standing still long enough to see the difference between the work and the workarounds.

That's the friction audit. Not finding what's broken. Finding what you've adapted to.

The Cost of Not Looking

Inefficiency costs companies 20 to 30% of revenue annually. That's not a small business problem. That's an industry wide finding across companies of every size. Operating expenses wasted on process inefficiency sit in the same range: 20 to 30%. And globally, process friction accounts for over $3 trillion in lost value per year.

Managers spend 40% of their time resolving internal issues that shouldn't exist. Not client issues. Not strategic challenges. Internal friction. Problems that only exist because the processes around them were never designed properly, or were designed for a business half the current size.

Productivity growth across advanced economies has slowed by nearly 40% over the last decade. Not because people are lazier. Because the operational infrastructure they work inside hasn't kept up with the complexity of the work.

And here's the number that should make anyone cautious about jumping straight to solutions: 70% of all transformation efforts fail to meet their objectives. Nearly 50% of RPA programmes fail. Failed projects cost organisations an estimated $2.3 trillion per year globally. The problem isn't that businesses don't try to improve. It's that they try to improve the wrong things.

Why Most Improvement Efforts Fail

Eliyahu Goldratt, the creator of the Theory of Constraints, said something that should be tattooed on every operations manager's wall: spending time optimising non constraints will not provide significant benefits.

Most businesses optimise whatever's loudest. The process that generated the most complaints last month. The tool that crashed on Tuesday. The team that missed a deadline. But if none of those are the actual bottleneck, fixing them produces zero system level improvement. You've spent money, time, and political capital improving something that didn't matter.

The other failure mode is even more common: automating a broken process.

Then there's the monolithic automation trap. Organisations that build massive end to end automated workflows create systems that are brittle, hard to debug, and impossible to modify without breaking everything downstream. Modular, targeted automations dramatically outperform.

The Eight Wastes (Translated for Knowledge Work)

Toyota's framework identified eight categories of waste. They were designed for manufacturing, but they translate directly into every office, agency, and service business.

Waste type	Factory version	Knowledge work version
Transport	Moving materials unnecessarily	Forwarding emails, moving data between systems manually
Inventory	Excess stock sitting idle	Backlogs, unread messages, stalled projects
Motion	Unnecessary physical movement	Switching between 12 browser tabs to complete one task
Waiting	Idle time between process steps	Approvals sitting in inboxes for days
Overproduction	Making more than needed	Reports nobody reads, meetings that could be emails
Overprocessing	Adding unnecessary features	Formatting a spreadsheet that gets pasted into another spreadsheet
Defects	Products that need rework	Errors from manual data entry, miscommunication
Skills	Underusing workers' abilities	Senior staff doing admin because "it's faster if I just do it"

That last one is the quiet killer. Every time a senior person does work that a system or a junior team member could handle, you're paying premium rates for commodity tasks. And you're building a dependency that makes it harder to ever stop.

Three Layers, Not One

Most audit guides stop at layer one: find the slow, manual, or redundant steps. That's necessary but not sufficient. The reason improvement efforts fail isn't because organisations can't find friction. It's because they remove friction without understanding the adaptations built around it.

Layer 1: Map the friction

What's slow? What's manual? What requires data to move between systems by hand? What steps exist because "that's how we've always done it"? This is standard territory. Value stream mapping, process documentation, time tracking. Most consultants stop here.

Layer 2: Map the adaptations

This is where the real insight lives. For every friction point, there are workarounds. Sarah keeps a personal spreadsheet because the CRM doesn't track what she needs. Mark sends a Slack message every morning because the project tool doesn't notify the right people. The team has a "pre meeting meeting" because the actual meeting is too large to have real discussions.

These adaptations are load bearing. Remove the friction without understanding the adaptation, and things get worse. People built these workarounds for reasons. Those reasons need to be addressed in whatever replaces them.

Layer 3: Map the dependencies

Which adaptations would break if you removed the friction they're built around? What needs to replace them? This is the layer that prevents the "automating a broken process" failure. If you automate the CRM without replicating what Sarah's spreadsheet does, you haven't improved anything. You've just broken Sarah's workflow.

Why You Haven't Done This Already

Knowing you should audit your operations and actually doing it are separated by a wall of perfectly rational psychological resistance.

Status quo bias. Your brain's default mode actively resists deviation from established patterns. The current way of doing things feels safe because it's known, even when it's objectively terrible.

Loss aversion. Auditing means confronting the possibility that what you've been doing is wrong. That's a loss of competence, status, and certainty. Most people avoid it unconsciously.

The boiling frog. Leaders adjust expectations downward so gradually they don't notice. What was unacceptable two years ago is now "just how things are." The team adapted. Expectations dropped. Nobody remembers the original standard.

The 70% failure rate as self fulfilling prophecy. Your team has probably lived through failed improvement initiatives before. They have rational reasons to be cynical. The worst thing you can do is launch another "transformation" that dies quietly in quarter two.

Fear of competence threat. Long tenured staff may feel that changing processes threatens expertise they spent years developing. If the workaround they built gets automated, what's their role?

How to Actually Do It

Goldratt's five focusing steps are the simplest framework that works:

1. Identify the constraint

What is the single biggest bottleneck in your operation right now? Not the five things that annoy you. The one thing that, if it were faster, would make everything downstream faster too. In a service business, it's usually one of: intake and onboarding, approval routing, or the handoff between sales and delivery.

2. Exploit the constraint

Before spending money, squeeze more throughput out of the bottleneck with what you have. Can approvals be batched? Can the bottleneck person delegate any of their decisions? Can you eliminate any steps in the constrained process?

3. Subordinate everything else to the constraint

Nothing upstream should produce faster than the bottleneck can process. Nothing downstream should wait for anything other than the bottleneck's output. Align the whole operation around the single point that's holding it back.

4. Elevate the constraint

Now invest. Automate the bottleneck. Add capacity. Redesign the process. This is where AI automation fits. But only after you've identified the right constraint, squeezed what you can from it manually, and aligned everything else around it.

5. Repeat

Once you fix the biggest constraint, a new one emerges. That's not failure. That's progress. The system improves each time the constraint moves.

One company identified hidden handoffs in their process using this approach, achieved a 33% reduction in rework and a 10% increase in throughput. Another saved $1.2 million annually with 70% fewer errors by remapping and digitising workflows. A government programme returned $4.50 for every $1 invested in lean process improvement, saving $33 million total.

The most effective audit tool ever invented was a chalk circle on a factory floor. No software. No framework. Just standing still long enough to see what you've stopped seeing.

• Map friction first, then adaptations, then dependencies. In that order.
• Find the single constraint before trying to improve anything else
• Never automate a process you haven't audited. Automation magnifies whatever's already there.
• Involve the people doing the work. They already know where the problems are. The audit's value is creating a structure where that knowledge can be acted on.
• Start small. One constraint, one fix, one measurement. Then repeat.

If you want to run a friction audit on your operations but aren't sure where the constraint is, book a free audit. We'll stand in the circle with you.

Sources

1. SixSigma.us: TIMWOODS and the 8 Wastes of Lean
2. IBM Garage: Value Stream Mapping
3. Lean Enterprise Institute: Theory of Constraints
4. Yale SOM: Cass Sunstein on Sludge Audits
5. Gemba Academy: Genchi Genbutsu and the Chalk Circle
6. Harvard DataSmart: Results Washington Lean Programme
7. McKinsey: Common Pitfalls in Transformations
8. Havior: Operational Inefficiencies Are Costing You More Than You Think