How The World Works | Friction

A world without friction sounds efficient until you try to walk across it.

Your foot pushes backward against the ground. The ground pushes you forward. A tyre grips the road. A brake pad slows a wheel. Fingers hold a glass because surfaces resist sliding.

Friction wastes energy.

Friction also makes controlled movement possible.

That paradox is worth carrying out of physics and into the rest of the world.

A form can have friction. A market can have friction. Changing banks can have friction. Finding trustworthy information can have friction. A bureaucracy can create friction. A safety check can deliberately add friction. A poorly designed interface can create so much friction that a technically available route becomes practically unusable.

Not all these uses are literal physics. They are analogies built around one powerful structural idea:

Movement across a boundary usually has a cost.

Quick Read

In physics, friction is a contact force that opposes relative motion or attempted motion between surfaces.

In economics, organisations and everyday systems, “friction” is often used more broadly for the costs that resist movement, exchange or change: time, search, paperwork, uncertainty, retraining, negotiation, coordination, switching, authentication and other forms of effort.

The crucial insight is that friction is not automatically bad.

Too much friction can block useful action.
Too little friction can remove safety, deliberation, grip or accountability.
Good design reduces unnecessary friction while preserving friction that protects the system.

A mature question is therefore not “How do we remove all friction?”

Where should movement be easy, where should it be difficult, and who pays the resistance?

The One-Sentence Answer

Friction works by imposing resistance at an interface, so movement requires additional force, time, information, coordination or cost; the same resistance can protect useful boundaries or become wasteful when it no longer serves a legitimate purpose.

Start With the Literal World

It helps to begin with physics because the original concept keeps our metaphors disciplined.

OpenStax defines friction as a force that opposes relative motion between systems in contact. Static friction acts when surfaces are not sliding relative to one another; kinetic friction acts when they are.

Push a heavy box gently. It does not move. Push harder. It still may not move because static friction rises to oppose the attempted motion up to a limit. Exceed that limit and the box begins sliding. Once moving, the frictional behaviour changes.

This creates a wonderfully useful intuition:

Starting movement and maintaining movement are not always the same problem.

Human systems often behave similarly—not because paperwork obeys Newton’s laws, but because initiation can require a threshold of effort that ongoing activity does not.

Friction Is Not Merely Waste

If friction only wasted energy, engineers would eliminate it everywhere.

They do not.

Without sufficient tyre-road friction, a vehicle cannot accelerate, steer or brake effectively. Without friction under a shoe, walking becomes difficult. A knot holds partly because contact resists sliding. A screw depends on frictional forces to stay in place.

So the better principle is:

Resistance is bad only relative to the movement you are trying to achieve.

That sentence becomes extremely important when we move into institutions.

The Systems Translation

In a human system, friction is a useful umbrella term for effort added between intention and completion.

intention → interface → resistance → additional effort → completion, delay, abandonment or error

The resistance can be physical, cognitive, informational, financial, administrative, social, technical or legal.

A locked door creates physical friction. A password creates security friction. A confusing form creates cognitive friction. A search process creates information friction. A cancellation call creates administrative friction. Incompatible file formats create technical friction. A rule requiring two approvals creates governance friction.

These should not be collapsed into one undifferentiated number. The label helps us notice resistance; analysis still has to identify its source.

Transaction Costs: The Price of Making Exchange Happen

Economics provides one of the richest versions of the friction idea.

A simple market story imagines buyers and sellers meeting, agreeing on a price and exchanging goods.

Real exchange requires more.

Find a counterparty.
Compare alternatives.
Verify quality.
Negotiate terms.
Write or understand a contract.
Transfer payment.
Monitor performance.
Resolve disputes.
Enforce the agreement if something goes wrong.

Those activities consume resources.

Ronald Coase made transaction costs central to understanding why firms and institutions exist. Oliver Williamson developed transaction-cost economics further, examining how different governance structures handle exchange and conflict.

The deeper lesson is that production cost is not the whole cost of economic activity. Coordination has a price too.

Search Friction: Before You Can Choose, You Must Find

Imagine trying to buy the best product in a market where prices are hidden, quality is uncertain and every seller must be contacted separately.

The market may contain excellent options. Yet discovering them is costly.

Search friction includes the time and effort required to locate alternatives and obtain enough information to compare them.

Search engines, comparison sites, directories, indexes, catalogues, maps, standardised labels and transparent pricing all reduce certain forms of search friction.

But lowering search friction can create new problems. If ranking systems are poor, manipulated or opaque, people may move quickly toward the wrong option.

Speed is not the same as epistemic quality.

Information Friction: Finding Is Not Understanding

A document can be easy to access and difficult to use.

Technical vocabulary, poor structure, missing context, inconsistent terminology and ambiguous instructions create information friction even when the file is one click away.

This matters because digitisation can remove physical friction while leaving cognitive friction untouched.

A paper form becomes an online form. The queue disappears. But the questions remain confusing, the required evidence remains unclear, and errors still send users back to the beginning.

The interface became faster. The route did not necessarily become easier.

Administrative Friction: The Work Around the Work

Administrative friction appears when people must perform procedural work to access, deliver or change something.

Some of this work is necessary. Identity must be verified. Public money needs accountability. Safety-critical changes need review. Records must be kept.

But procedures accumulate.

A form is added after one incident. An approval is added after another. A legacy database requires duplicate entry. A rule is never removed when the original risk disappears. Different departments ask for the same information because their systems cannot exchange it.

Eventually the organisation spends surprising amounts of energy moving through itself.

Sludge: Friction That Has Lost Its Justification

Behavioural science has popularised the term sludge for excessive or unjustified friction that makes it harder for people to do what they want or need to do.

The OECD has documented governments using “sludge audits” to identify psychological, temporal, financial and administrative burdens in public services.

The important word is unjustified.

A security check may be friction without being sludge. A waiting period may be deliberate safety friction. A second approval may protect against fraud or catastrophic error.

The audit should ask what purpose each barrier serves rather than assuming every barrier is waste.

Switching Friction: Why Leaving Is Harder Than Arriving

Switching costs appear when moving from one provider, technology, standard or routine to another requires extra work.

Data must be transferred.
New software must be learned.
Contacts must be informed.
Contracts must be ended.
Files may need conversion.
Equipment may be incompatible.
Benefits accumulated in the old system may be lost.
There may be uncertainty about whether the new system will work.

Switching friction helps explain why users remain with systems even when alternatives exist.

The OECD has noted that data portability and interoperability can reduce switching costs in digital markets by allowing users to carry information to another provider rather than rebuilding their digital history from zero.

This is a direct bridge from friction to competition and path dependence.

Coordination Friction: When Everyone Must Move Together

Some changes are hard not because any individual step is difficult, but because many actors must coordinate.

A standard changes. Suppliers must adapt. Software must update. Users need training. Regulators may need new rules. Old and new versions must coexist temporarily. Documents must be revised. Support staff need new scripts.

The friction is distributed across relationships.

This is why large systems can remain on an old path even when nearly everyone agrees the new destination is better. Agreement about destination does not eliminate coordination cost.

Cognitive Friction: The Cost Inside the Head

A choice can be physically easy and mentally expensive.

Compare two forms.

The first asks five clear questions using familiar language.

The second asks the same five things but uses unfamiliar terms, unclear categories, exceptions, nested instructions and a warning that mistakes may delay the application.

The number of clicks is identical.

The friction is not.

Cognitive friction matters because people differ in background knowledge, language, confidence, working memory, fatigue and prior experience. A process that feels simple to its designer can be exhausting to its receiver.

Social Friction: When the Cost Is Embarrassment, Permission or Status

Some routes are difficult because taking them changes a social relationship.

Asking for help may require admitting uncertainty. Challenging a senior colleague may risk status. Reporting a problem may require crossing hierarchy. Leaving a group may create social cost. Negotiating a price may feel culturally uncomfortable.

These costs are real even when they do not appear in accounts.

A system that depends on people reporting bad news must therefore design for social friction, not merely provide a reporting button.

Geographical Friction: Distance Is Not Just Kilometres

Distance creates cost through travel time, transport, borders, terrain and connectivity.

But two places ten kilometres apart can have very different effective distance depending on roads, public transport, checkpoints, congestion or digital connectivity.

Infrastructure lowers geographical friction by converting difficult movement into reliable routes.

A bridge does not shorten the map. It shortens the journey.

This is one reason transport networks can reorganise economies and cities: they change the cost surface of space.

Temporal Friction: Delay Changes Behaviour

Time itself can be friction.

A process that takes two weeks may produce a different outcome from the same process completed in two minutes.

Some people abandon the route. Needs change. Opportunities expire. Memory fades. Uncertainty grows. Other systems continue moving while approval waits.

But delay can also be protective. Cooling-off periods, review windows and staged approvals can prevent impulsive or dangerous action.

Again, the problem is not delay itself. It is whether the delay serves the system’s purpose.

Interface Friction: The Boundary Between Two Systems

Many failures happen at interfaces rather than inside components.

Two departments each work well but exchange information poorly. A hospital system and insurer use incompatible formats. A student understands an idea but cannot translate it into the answer format required by an examination. A website contains the right information but hides it behind navigation nobody can predict.

Interface friction is especially dangerous because each side can appear locally successful.

The failure lives between them.

See How Interfaces Work.

Friction Accumulates Along a Route

One small inconvenience rarely destroys a system.

Ten small inconveniences can.

Imagine a service with a 95% completion probability at each of ten independent stages. The end-to-end completion rate would be roughly 60%, not 95%.

Real processes are not usually independent in that neat mathematical way, but the lesson is valuable: friction compounds across routes.

A confusing instruction leads to a missing document. The missing document leads to a rejected form. The rejection requires a phone call. The phone call has a long wait. The user misses the call-back. A minor problem becomes route abandonment.

Local optimisation can therefore miss end-to-end friction.

The Last-Mile Problem

A system has not succeeded merely because it produced an output.

The output must reach and be usable by its receiver.

A government benefit exists but the application is inaccessible. A lesson is excellent but the student cannot decode the vocabulary. A product reaches a warehouse but not the customer. A warning is issued but buried in technical language. A data standard exists but no converter is available.

The final interface can carry more friction than the entire upstream system expects.

This is why receiver-side measurement matters.

Friction and Inequality

The same friction does not cost everyone the same amount.

A process requiring a weekday appointment is easier for someone with flexible work. A digital-only route is easier for someone with reliable access and confidence. A legal document is easier for someone with expertise. A long delay is easier for someone with savings. A confusing form is easier for someone who can pay a professional to complete it.

Friction therefore has distribution.

A designer who measures only average completion time can miss the fact that one group is carrying most of the burden.

Good systems examine who drops out, who needs help, who pays extra and who never reaches the endpoint.

Friction Can Be a Business Model

Sometimes an organisation benefits when entering is easy and leaving is hard.

Sign-up takes two clicks. Cancellation requires a phone call. Data enters easily but exports poorly. A free trial converts automatically unless the user notices a deadline. A loyalty system accumulates benefits that disappear on exit.

Not every asymmetric process is malicious. Security, fraud prevention or contractual requirements may justify differences.

But asymmetry deserves inspection when the party controlling the friction benefits financially from the receiver failing to cross it.

Friction and Defaults

Defaults and friction are close companions.

A default determines what happens if no active alternative is selected. Friction determines how difficult it is to choose differently.

A benign default with an easy opt-out may preserve meaningful choice. The same default with a hidden, time-consuming exit route can become far more powerful.

Choice architecture therefore has at least two dimensions:

Where does the person start?
What does it cost to move?

See How The World Works | Defaults.

Friction and Path Dependence

Switching friction is one of the engines that makes history persistent.

An early technology gains users. Users accumulate data and skills. Complementary products develop. Moving to an alternative now requires conversion, retraining and coordination.

The friction of switching reinforces the old path.

This is why data portability, open standards and interoperability can reopen options: they do not erase history, but they reduce the cost of leaving it.

See How The World Works | Path Dependence.

Friction and Externalities

Transaction costs help determine whether people affected by an externality can bargain.

If one neighbour and one business can cheaply negotiate over noise, a private agreement may be possible. If millions of people are affected by dispersed pollution, the coordination friction becomes enormous.

The existence of an externality is one problem. The friction involved in resolving it is another.

See How The World Works | Externalities.

Friction and Markets

Perfectly frictionless markets are useful theoretical benchmarks.

Real markets have search cost, information asymmetry, contract cost, transport, taxes, delays, switching costs, regulation and enforcement.

These frictions shape which transactions happen at all.

A profitable exchange in a zero-friction model may never occur if finding the other party costs too much. A competitive market may become sticky if customers cannot compare prices. A superior entrant may fail if switching costs keep users attached to incumbents.

Friction is therefore not an afterthought around the market. It helps determine the market’s structure.

Friction and Organisations

Why do firms exist instead of every worker contracting independently with everyone else for every task?

Coase’s famous answer points toward the cost of using the market itself. Finding prices, negotiating separate contracts and coordinating repeated exchange can be expensive.

Organisations can reduce some of those transaction costs by bringing activity under a common governance structure.

But organisations create internal friction too: hierarchy, meetings, reporting, politics, bureaucracy and slower adaptation.

There is no frictionless organisational form. The design problem is comparative:

Which arrangement handles this kind of coordination with the lowest relevant total cost and acceptable safeguards?

Friction and Safety

Some of the most important friction in civilisation is intentional.

We require confirmation before deleting important data. High-risk changes need review. Aircraft use checklists. Financial systems use authentication and settlement controls. Laboratories use safety procedures. Courts require process. Organisations separate certain duties so one person cannot silently complete every sensitive step.

These controls slow action.

That is partly the point.

Safety friction creates time or effort for error detection, verification and accountability.

A system obsessed with speed can accidentally optimise away its brakes.

The Two Kinds of Friction Error

Designers usually make one of two opposite mistakes.

Error 1: Too much friction

Useful action becomes expensive. People abandon the route. Staff create workarounds. Informal channels replace formal ones. Only experts can navigate the process. Errors rise because users must remember too many steps.

Error 2: Too little friction

Irreversible actions become too easy. Mistakes propagate instantly. Fraud becomes cheap. Impulsive decisions receive no pause. Dangerous changes bypass independent review. Systems become tightly coupled and fast-moving without containment.

Maximum convenience is therefore not the same as maximum quality.

The Brake-Pedal Principle

A brake creates friction so that a vehicle can stop where the driver intends.

That is a useful design metaphor.

Good friction should be placed where uncontrolled momentum creates meaningful risk.

Before irreversible deletion.
Before sending sensitive information.
Before high-cost commitments.
Before actions that affect many other people.
Before changes that are difficult to roll back.

Routine, low-risk actions should not have the same braking force as rare, high-risk ones.

This is where risk-based design becomes more intelligent than universal bureaucracy.

The Lubrication Principle

Physical systems reduce unwanted friction through lubrication, bearings, smoother surfaces and better alignment.

Human systems have equivalents.

Clear instructions.
Single sign-on.
Pre-filled information that the user can verify.
Interoperable standards.
Shared identifiers.
Well-designed search.
Transparent prices.
One point of contact.
Automatic data transfer with consent.
Templates that remove repetitive work.
Common vocabulary.

The analogy should not be stretched too far, but the design idea is sound: reduce resistance at interfaces where resistance serves no useful function.

Digitisation Does Not Automatically Remove Friction

Moving a bad process online can produce a fast bad process.

A paper form may become a web form while preserving duplicate fields, ambiguous questions, unnecessary approvals and incompatible databases.

The OECD has explicitly warned that digitalisation can entrench friction when the underlying process is poorly designed.

Technology should therefore follow process diagnosis rather than substitute for it.

Ask first why the step exists. Then decide whether to digitise, simplify, combine, automate or remove it.

Workarounds Are Friction Sensors

When people repeatedly bypass a formal process, the immediate reaction is often to blame non-compliance.

Sometimes that is correct.

But workarounds can also be diagnostic.

Staff use spreadsheets because the official system cannot answer a practical question. Customers call a human because the automated route loops. Teachers create unofficial materials because the official interface is too slow. Users screenshot information because export is impossible.

A workaround is evidence that somebody found the official friction too expensive relative to an alternative route.

The correct response is to inspect both: Is the workaround unsafe, and what failure in the formal route made it attractive?

Friction Can Move Rather Than Disappear

A redesign may make one user’s route easier by transferring work elsewhere.

Customers receive a one-click return, but warehouse staff face complicated inspection work. A teacher gets automated marking, but students must enter answers in a rigid format. A central team standardises reporting, but local teams spend hours translating their work into the template.

Friction has not necessarily been removed.

It has been relocated.

This is why end-to-end system design matters. Measure total route cost, not just the visible interface.

Friction and Education

Learning contains productive friction.

Retrieving an answer from memory is harder than rereading it. Writing an explanation is harder than recognising the correct option. Solving a novel problem is harder than copying an example.

That difficulty can be useful because learning often requires effortful processing, retrieval and error correction.

But educational friction can also be useless. A student cannot understand a question because the instructions are poorly written. Homework requires five platforms with separate logins. A worksheet overloads decoration while hiding the concept. Feedback arrives so late that the learner no longer remembers the reasoning.

The teaching job is not to make learning frictionless.

It is to keep the difficulty attached to the capability we actually want to build.

Do not make the route easy by removing the learning. Make the learning reachable by removing unrelated friction.

Friction and Change

People often underestimate change because they compare two end states and ignore the migration path between them.

Current system A costs $100 per year. New system B costs $80. Therefore switching appears obvious.

But migration requires data conversion, retraining, temporary duplication, testing, downtime risk and compatibility work. The first-year cost may be far higher.

This does not mean “never change.” It means the transition has its own economics.

See How Change Works.

Friction and Capacity

Friction consumes capacity.

If staff spend half their day reconciling incompatible systems, that is capacity unavailable for the main job. If citizens spend hours navigating forms, that time is unavailable elsewhere. If machines lose energy to unwanted friction, more input is needed for the same output.

Removing unnecessary friction can therefore increase effective capacity without adding more people, machines or money.

But if the removed friction was a safety control, apparent capacity can rise while risk rises faster.

Capacity and safety must be read together.

Friction and Trust

Trust can reduce transaction friction.

If parties trust that commitments will be honoured, less effort may be needed for verification, monitoring and defensive contracting.

But blind trust can remove useful checks.

Reliable systems therefore convert personal trust into trustworthy structure: clear standards, audit trails, enforceable commitments and transparent processes. That allows friction to fall without making the system dependent on optimism.

See How Trust Works.

Friction and Power

The party who controls the interface can often decide where friction sits.

An employer chooses how complicated reimbursement is. A platform chooses how easy data export is. A government chooses how many steps a permit requires. A school chooses how parents access information. A company chooses whether cancelling is as easy as joining.

Friction placement is therefore a form of power.

The ethical question is not merely how much friction exists, but whether the people bearing it had any role in the design and whether the burden is proportionate to the purpose.

The Friction Budget

Every process spends some of the receiver’s time and attention.

Think of that as a friction budget.

If identity verification genuinely requires effort, spend the budget there. If a safety warning genuinely needs attention, make it clear. Do not simultaneously waste the same person’s attention on decorative warnings, duplicate fields, unexplained jargon and unnecessary navigation.

When everything is difficult, nothing receives appropriate attention.

Good design concentrates friction around meaningful decisions and removes it from routine passage.

The Friction Audit

Define the desired movement. What is trying to move—person, data, money, object, decision, skill or authority?
Find the interface. Where does resistance appear?
Name the friction type. Physical, cognitive, informational, administrative, financial, social, legal, technical or temporal?
Measure the burden. Time, money, steps, errors, abandonment, retraining or uncertainty?
Find the purpose. What risk or coordination problem is this friction supposed to solve?
Test necessity. Does the barrier still serve that purpose?
Check distribution. Who carries most of the burden?
Check asymmetry. Is entering much easier than leaving, and why?
Trace workarounds. What informal routes have people created?
Check accumulation. Do small frictions compound across the whole journey?
Check relocation. Would simplification here merely move work somewhere else?
Redesign selectively. Remove sludge, preserve safeguards, then measure the receiver again.

When the Friction Lens Fails

“Friction” is a metaphor outside physics, and metaphors can overreach.

Not every disagreement is social friction. Not every cost is transaction friction. Not every slow process is badly designed. A legal review may be slow because the underlying question is genuinely complex. Learning may be difficult because the subject is difficult. A border may impose delay because legitimate inspection takes time.

The word helps locate resistance. It does not explain the resistance by itself.

Always descend one level further and identify the actual mechanism.

Do Not Optimise Away the Brake

Modern design culture often celebrates seamlessness.

One click. Instant. Automatic. No waiting. No paperwork.

For many routine tasks, this is excellent.

But civilisation also depends on pauses.

A surgeon checks the patient and procedure before operating. An engineer reviews a high-risk change. A court hears both sides. A bank authenticates a transfer. A researcher submits work to scrutiny. A teacher asks a student to show working rather than only producing an answer.

These moments create friction because error matters.

The mature system is not frictionless.

It is friction-literate.

How Friction Connects to the Rest of the World

Defaults: friction determines how hard it is to leave the starting state.
Path dependence: switching friction helps inherited systems persist.
Externalities: transaction friction affects whether outside costs can be bargained or corrected.
Markets: search, bargaining, transport and enforcement costs shape exchange.
Competition: switching costs can make customers less mobile.
Power: interface owners often decide who bears friction.
Trust: reliable institutions can reduce verification costs.
Standards: compatibility lowers translation friction.
Interoperability: shared interfaces reduce migration and exchange costs.
Safety: deliberate friction can stop dangerous momentum.
Capacity: unnecessary friction consumes usable resources.
Change: migration friction determines how feasible reform really is.

Questions a Reader Can Now Ask

What is trying to move?
Across which boundary?
What exactly resists it?
Is the resistance physical, informational, administrative, cognitive or social?
Who pays the time or effort?
What useful purpose does the friction serve?
What risk appears if we remove it?
Why is starting easier or harder than continuing?
Why is joining easier or harder than leaving?
Which users face disproportionately high friction?
Where do people create workarounds?
Can a standard, adapter, better interface or clearer instruction reduce friction safely?
Did we actually remove work, or merely move it elsewhere?

Frequently Asked Questions

Is all friction bad?

No. Physical friction lets us walk and brake. Human systems also use deliberate friction for safety, verification, accountability and reflection. The target should be unnecessary friction, not all resistance.

What is transaction friction?

It is the cost around making exchange happen: searching, comparing, negotiating, contracting, monitoring and enforcing. Economists generally discuss these as transaction costs.

What is sludge?

Sludge is excessive or unjustified friction—such as unnecessary paperwork, confusing processes or avoidable psychological burden—that obstructs useful action without a proportionate purpose.

Why do switching costs matter?

Because a better alternative is not enough if moving to it requires large conversion, retraining, data transfer or coordination costs. Switching friction can protect incumbents and reinforce path dependence.

What is the best way to reduce friction?

First identify the friction’s purpose. Remove duplication, ambiguity and unnecessary steps; improve interoperability and information; simplify low-risk routes; keep safeguards where error matters; then measure whether the receiver’s outcome actually improved.

Research Basis and Further Reading

OpenStax, “Friction”, for the physical foundation: static and kinetic friction and resistance to relative motion.
Ronald H. Coase, Nobel Prize materials on transaction costs and the institutional structure of the economy.
Oliver E. Williamson, “Transaction Cost Economics: The Natural Progression”, Nobel Prize lecture.
OECD, “Fixing frictions: ‘sludge audits’ around the world”, on unnecessary administrative and psychological burdens.
OECD, Data Portability, Interoperability and Competition, on switching costs and competition in digital markets.

The Larger Idea

There is a temptation, once you learn to see friction, to declare war on it.

Remove the queue. Remove the form. Remove the password. Remove the review. Remove the delay. Remove the difficult question. Make everything immediate.

Then the world slides.

Some things become beautifully easy.

Other things become impossible to stop.

The deeper intelligence is not frictionlessness. It is placement.

Put grip under the foot. Put brakes on dangerous momentum. Put clear interfaces between systems. Remove paperwork that proves nothing. Remove repeated data entry. Remove hidden exit barriers. Preserve verification where error can hurt people. Make routine movement smooth and consequential movement deliberate.

That is what the physical world has been teaching us all along.

The best systems do not eliminate resistance. They put resistance exactly where it earns its cost.

How The World Works | Friction — Why Movement, Exchange and Change Are Never Free