Sludge (Sunstein): S-Tier Behavioral Designer’s Guide

You qualified for the program. You earned the refund. You wanted to cancel the subscription. None of that mattered, because somewhere between you and the thing you were entitled to, somebody had built a swamp. By the time you waded through page seven of the form, hunting for a documentation requirement that the agency had not bothered to list on page one, you had quietly decided it was not worth it. That decision is exactly what the swamp was designed to produce.

Cass Sunstein gave the swamp a name in 2022: Sludge. Not bureaucracy in the broad, ambient sense, but a specific category of friction, excessive or unjustified, that stands between people and the actions they want to take, the benefits they are owed, or the choices they have already made. Sludge is the dark twin of the nudge. A nudge is a small design move that helps the better choice happen. Sludge is a small design move that prevents it.

This is the framework every Octalysis-trained behavioral designer should run before shipping anything. Not because friction is always bad. Some friction is load-bearing: cooling-off periods on impulsive purchases, deliberate two-factor checks before sending money, the mandatory pause before a divorce. The interesting question is the one Sunstein keeps returning to: which of your frictions are doing real work, and which are just collecting time and dignity from your users without a defensible reason?

Below is the S-Tier Behavioral Designer’s Guide to Sludge: the academic backbone, the three categories of cost it imposes, the brain mechanisms that make it bite harder than economists expect, the design ethics line that separates protection from predation, and the Octalysis × Sludge mapping that turns the framework from an audit into a design surface. By the end you will be able to run a Sludge audit on your own product without consulting anyone else, and you will know which of your “ethical” frictions are actually predation in a button skin.

Speed Run Notes

• Sludge is the inverse of the nudge: excessive or unjustified friction that prevents beneficial action. Sunstein’s 2022 book gave the concept its formal definition and its audit methodology.
• Sludge imposes three costs by Sunstein’s taxonomy: learning costs, compliance costs, and psychological costs. Most product teams only measure the second. The third is where the real damage lives.
• Sludge is regressive. The same paperwork burden costs an attorney 20 minutes and a single mother of three an entire week of bandwidth, which is why eligibility uptake gaps survive even after benefits expand.
• Not all friction is sludge. The audit question is whether the friction is doing protective work the user would endorse on reflection, or extractive work disguised as protection.
• The Octalysis x Sludge mapping locates friction inside specific Core Drives. Sludge that breaks CD2 progress feedback feels different from sludge that weaponizes CD8 loss aversion, and the design fixes differ.
• The audit deliverable is not a memo. It is a per-flow inventory: every field, consent screen, and wait classified as protective, neutral, or extractive, with an owner assigned to remove the extractive rows.

Author Credibility: Yu-kai Chou

Yu-kai Chou created the Octalysis Framework after studying gamification since 2003 — years before the term entered mainstream vocabulary. As a Human-Systems Architect & Behavioral Designer, his framework has been applied by LEGO, Microsoft, Porsche, Coca-Cola, Salesforce, and MrBeast, impacting over 1.5 Billion Users.

Chou has taught the Octalysis methodology at Harvard, Stanford, Yale, Tesla, Google, BCG, and IDEO.

His work has been cited by Harvard, Stanford, MIT, Forbes, Wall Street Journal, Wired, US Department of Energy, NIST, NSF, NCBI, US Department of Education, ClinicalTrials.gov, and Google Scholar — with 3,700+ more academic publications. Explore his books here.

TEDxLausanne · Gamification to Improve Our World17 min

What Is Sludge?

Sunstein’s definition is narrow on purpose. Sludge is “excessive or unjustified frictions”: process burdens placed between a person and an outcome they want, which a reasonable observer would conclude do not earn their cost. The qualifier “unjustified” carries the whole load. Friction that protects people, prevents fraud, or forces a moment of deliberation is not sludge. Friction that simply taxes attention and time without a defensible payoff is. The job of a sludge audit is to walk a process step by step and force every friction to defend itself.

The concept entered the academic literature in 2019, when Sunstein published “Sludge Audits” in Behavioural Public Policy, but the construct had been building for years. Pamela Herd and Donald Moynihan’s 2018 book Administrative Burden documented how American public benefits programs had become systems of policymaking by deterrence: eligibility rules were generous on paper, but the application burden quietly produced the same outcome as denying the benefits in the first place. Sunstein’s 2022 book Sludge: What Stops Us from Getting What We Want and What to Do about It took the construct out of the public-administration silo and into general behavioral economics, then applied it to private-sector experiences with equal force: subscription cancellations, insurance claims, refund processes, healthcare prior authorization, college financial aid.

The cleanest way to grasp Sludge is by contrast with its sibling concept, the nudge. A nudge, in Thaler and Sunstein’s original 2008 formulation, is any aspect of choice architecture that alters behavior in a predictable way without forbidding options or significantly changing economic incentives. Defaulting employees into a 401(k) plan is a nudge. Sending a reminder text the day before a doctor’s appointment is a nudge. Both make the better outcome easier. Sludge does the opposite: it preserves the option but makes it harder to exercise, often by orders of magnitude, until the user gives up.

This is why “Sludge” is not a synonym for “bad UX” or “bureaucracy.” It is a specific category with its own diagnostic logic. A confusing app screen is bad UX. A confusing app screen that you encounter only when you try to downgrade your subscription is sludge. A long government form is bureaucracy. A long government form whose length cannot be defended by any specific information need is sludge. The audit question is always the same: what work is this friction doing, and would the user endorse it on reflection?

The Anatomy of Sludge: Three Categories of Cost

Sunstein’s central analytic move, borrowed and sharpened from Herd and Moynihan, is to break the cost of friction into three distinct categories. Most product teams measure only one. Understanding all three is what separates a sludge audit from a UX review.

Learning Costs: The Cost of Figuring Out What to Do

Before a person can comply with a process, they have to understand what the process requires. Learning costs are the time and cognitive effort spent figuring out eligibility, mapping the steps, finding the right form, decoding the jargon, and identifying the documentation needed. The Free Application for Federal Student Aid (FAFSA) is the canonical American example: before its 2024 simplification, the form was 108 questions long, and the average family spent six hours just understanding what was being asked. The Department of Education’s own research found that a meaningful fraction of eligible families never completed the form because they could not figure out the rules in the first place.

In private-sector products, learning costs hide in the help center. Every time a user has to leave the main flow to search “how do I cancel” or “where is the refund policy,” they are paying a learning cost the product has off-loaded onto them. The metric to watch is the ratio of help-center traffic to attempted-action traffic. A high ratio is not user incompetence. It is sludge.

Compliance Costs: The Cost of Actually Doing It

Once the user knows what is required, they still have to do it. Compliance costs cover the time, paperwork, signatures, uploads, phone calls, in-person visits, and follow-up steps that turn a permitted action into a completed one. The 2022 White House report on the “time tax” estimated that Americans spent more than 11 billion hours per year complying with federal paperwork — the equivalent of more than five million full-time jobs dedicated to filing forms with their own government.

Compliance costs are the only category most product teams measure, and they typically measure them badly. The standard funnel chart shows “steps completed” without ever asking what each step costs. A signup flow that requires uploading a photo of a driver’s license is not just “one more step.” For a user without a recent photo, decent lighting, or a smartphone with a working camera, it can be the friction that ends the relationship. Compliance costs are unevenly distributed by definition, and that uneven distribution is where Sludge does its most damage.

Psychological Costs: The Cost of Feeling Like You’re Doing It Wrong

This is the category product teams almost never quantify, and it is the one Sunstein argues matters most. Psychological costs are the dread, anxiety, embarrassment, stigma, and learned helplessness that accumulate during a sludge-heavy process. Filing for unemployment benefits requires admitting you are unemployed, sometimes to a recording, sometimes to a clerk, often to a webpage that calls you “the claimant.” Filing a health insurance claim requires becoming temporarily fluent in a billing dialect you neither speak nor want to learn. Disputing a charge requires being skeptical, repetitive, and slightly aggressive in a tone that does not come naturally to people who were raised to be polite.

Each individual instance is survivable. The accumulation is not. The same psychological tax that costs a tenured professor a few minutes of irritation can cost a single mother of three the last bandwidth she had that day. This is why sludge is regressive in a way that pure compliance-cost accounting misses entirely. The dollar value of a wasted hour is the same for everyone in a spreadsheet. The actual cost of that wasted hour is wildly unequal, and the curve bends against the people who most need the benefit at the end of the form.

What Sunstein Got Right

Three intellectual moves account for why Sludge stuck where dozens of “friction” papers before it did not.

The first move was naming the thing. “Bureaucracy” was too broad. “Friction” was the right concept but had been so thoroughly absorbed into product-design folklore that it had stopped meaning anything specific. By coining “Sludge” and pinning it explicitly to the nudge framework, Sunstein made the construct portable: it travels across government, healthcare, finance, and consumer software because everyone in those fields already had a working vocabulary for the nudge.

The second move was insisting on the audit as a deliverable. Most behavioral-science books leave the reader inspired but actionless. Sludge gave practitioners a process: walk through every step of every consequential flow, classify each friction, calculate its cost across all three categories, and present the inventory to a decision-maker who can remove rows. The 2022 OMB Memorandum M-22-10 turned this audit into federal policy by directing agencies to identify and reduce administrative burdens in public-facing programs. The construct went from book to operational protocol in less than a year, which is rare.

The third move was refusing to be ideological about whether friction is good or bad. Sunstein has spent his career being attacked from both directions: critics on the left have called nudges manipulative; critics on the right have called them paternalistic. Sludge sidesteps that fight by insisting on the question of justification. Some sludge is defensible. Some is not. The work is the case-by-case audit, not the political posture. That intellectual humility is what made the framework adoptable by both deregulatory conservatives interested in cutting red tape and progressive policymakers interested in increasing benefits uptake, often citing the same Sunstein paper.

The combined effect is a framework that is small enough to fit in a head, rigorous enough to survive academic peer review, and operational enough to ship as a checklist. That is a rare combination, and it is why Sludge has become the dominant analytic vocabulary for friction in choice architecture less than four years after the book’s publication.

Where Sludge Falls Apart

For all its analytic power, Sludge has three weak seams that any honest practitioner has to acknowledge before pulling out the audit checklist.

Critique 1: “Justified” Is a Judgment Call That Smuggles Politics into the Audit

The framework’s whole load-bearing distinction rests on whether a given friction is “justified.” But justification is not a property of the friction. It is a property of the values a society holds about whom that friction is supposed to protect, from what, and at what cost. A seven-day waiting period for a firearm purchase strikes a public-health researcher as obvious, defensible sludge and strikes a Second Amendment advocate as obvious, indefensible sludge. The audit methodology cannot resolve that disagreement; it can only document the friction and pass the contested cases to whoever has the authority to overrule one side or the other.

This is not a fatal flaw, but it is a quiet one. The Sludge literature tends to use clean examples (FAFSA, SNAP, subscription cancellation) where almost everyone agrees the friction is excessive. The hard cases (voter ID requirements, drug-test mandates for benefits, citizenship verification) are where the framework’s analytic neutrality dissolves into the political fight it was designed to step around. Pretending the audit produces “objective” findings on those cases is what makes deregulators and progressives both able to cite Sunstein when convenient.

Critique 2: Removing Friction Sometimes Just Reveals the Real Bottleneck

The implicit theory of change in a sludge audit is that if you remove the friction, the desired action happens. But many failures-to-act are not friction problems. They are motivation, eligibility, or information-design problems disguised as friction problems. Reduce the time cost of a benefits application to zero, and uptake will still leave gaps if the eligible population does not know the program exists, does not believe they qualify, or does not trust the institution administering it. The 2024 FAFSA simplification cut the form length dramatically, but completion rates initially fell because the rollout botched the launch logistics, proof that friction is only one of several reasons people do not finish things.

The implication for designers is that a sludge audit should always be paired with an upstream question: even if every friction in this flow disappeared, would the outcome change? Sometimes yes. Sometimes the outcome was being prevented by something the audit cannot see.

Critique 3: The Friction Balance Keeps Shifting as the Threat Environment Changes

Sunstein’s framework assumes the audit is a snapshot. In practice, the cost-benefit math on a given friction shifts every time the threat environment shifts. Two-factor authentication looked like aggressive sludge in 2010 and looks like a defensible baseline in 2026 after a decade of account takeovers. Verifying identity before issuing unemployment benefits looked like punitive sludge before the 2020 pandemic-era fraud waves and looks like prudent stewardship after. The audit cannot freeze a judgment; it has to be re-run as the world changes. Most organizations run it once, treat the findings as permanent, and discover three years later that they have either over-corrected (removing protective friction that turned out to matter) or under-corrected (leaving extractive friction in place because the original audit said it was justified at the time).

What’s Really Happening Inside the Brain

The reason small frictions produce outsized behavior change is not that the user is lazy or unmotivated. It is that the brain treats every friction as a costly act of executive function, and the executive-function budget is finite.

Sendhil Mullainathan and Eldar Shafir’s 2013 book Scarcity formalized what behavioral economists had been observing for a decade: cognitive bandwidth operates like a limited resource that depletes across the day. When someone is already under bandwidth pressure (financial stress, sleep deprivation, parenting load, a chronic health condition), each additional friction does not subtract a fixed amount of motivation. It subtracts a percentage of the bandwidth that remains. The same five-minute form costs a well-resourced person five minutes and costs a bandwidth-depleted person the entire remainder of their day.

The neurological underpinning is in the prefrontal cortex, specifically the regions that handle goal-directed planning, working memory, and inhibitory control. Functional MRI studies have shown that activities perceived as administratively burdensome activate the same dorsolateral prefrontal circuits as effortful arithmetic, and that those circuits fatigue measurably across sustained use. When the circuits are depleted, the brain defaults to the path of least resistance: typically, abandoning the task. This is not weakness of will. It is metabolic.

The dopaminergic system compounds the effect. Dopamine is not just the reward chemical; it is the motivational fuel that the brain spends to initiate goal-directed action. Frictions that delay reward decrease the perceived value of the goal at the moment of action initiation, an effect quantified by Kable and Glimcher’s 2007 work on temporal discounting. A benefit that arrives in six weeks after a form is processed is, neurologically, worth significantly less in the moment of “should I fill out this form?” than the same benefit would be if it arrived in six minutes. This is why same-day enrollment systems consistently outperform same-week enrollment systems by margins that are difficult to explain from a pure rationality model.

The clinical implication is that the most expensive frictions are the ones that occur at low-bandwidth moments, on slow-payoff outcomes, for populations whose baseline executive-function budget is already taxed. The Sludge audit’s emphasis on psychological costs is not a soft, “let’s be kind” addition. It is a recognition that the brain costs the friction differently depending on who is paying.

Sludge vs. Other Theories

Sludge sits inside a dense neighborhood of friction-adjacent constructs. Understanding the borders matters because the audit prescriptions differ.

Sludge vs. Nudge

The cleanest distinction. A nudge alters choice architecture to make the better outcome easier. Sludge alters choice architecture to make the better outcome harder. The same default-enrollment rule is a nudge into a retirement plan and is sludge out of an email subscription list. The mechanism is symmetrical; the directionality and the user’s interest are what flip the label. A designer fluent in both should be able to identify when their product is using nudges to help and sludge to retain.

Sludge vs. Administrative Burden

Herd and Moynihan’s Administrative Burden (2018) is the public-administration ancestor of the Sludge construct. The vocabulary overlaps almost completely. The three-cost taxonomy (learning, compliance, psychological) is theirs. The difference is scope. Administrative Burden was specifically a critique of how American public benefits programs use procedural friction as a tool of de facto policymaking. Sludge generalized the construct to the private sector and added the audit methodology as a portable deliverable. The two frameworks are best read together: Administrative Burden for the deep public-administration case studies, Sludge for the operational checklist.

Sludge vs. Dark Patterns (Deceptive Patterns)

Harry Brignull coined “dark patterns” in 2010 to describe interface design choices that deceive users into actions they would not otherwise take. The terminology was rebranded to “deceptive patterns” in 2023 because the word “dark” had picked up unintended racial connotations. Sludge overlaps with deceptive patterns on a meaningful subset of cases (the cancel-the-subscription maze is both), but the constructs aim at different things. Deceptive patterns describe interface designs intended to mislead. Sludge describes process burdens that block legitimate action regardless of whether the user was misled. A perfectly honest, perfectly clear, but excessive form is sludge but not a deceptive pattern. A misleadingly labeled button is a deceptive pattern but not (necessarily) sludge. The FTC’s 2022 “Bringing Dark Patterns to Light” staff report and the EU’s Digital Services Act treat the two constructs as related but distinct enforcement targets.

Sludge vs. Friction-as-Feature

Some friction is intentional and load-bearing. The discipline of “frictionful design” (Aaron Shapiro, others) argues that the right amount of friction at the right moment is what separates a thoughtful product from a fast one. A confirmation dialog before a destructive action is friction. So is a cooling-off period before a major purchase, a typing-delay before a public post, the requirement to fill in your name before deleting an account. These exist precisely to prevent users from doing what they will regret. Sludge audits are not anti-friction; they are anti-unjustified-friction. The audit question is whether the friction exists to protect the user or to retain them against their interest. Same mechanism, opposite intent.

Sludge in the Real World

Four sectors are where the construct has done most of its real-world work in the last five years. Each illustrates a different pattern of sludge accumulation and a different shape of fix.

Government Services: The Time Tax

The 2022 White House report “Tackling the Time Tax” (the public-facing version of OMB Memorandum M-22-10) calculated that Americans collectively spent more than 11 billion hours each year complying with federal paperwork, the equivalent of more than five million full-time jobs. The Internal Revenue Service alone accounts for over 1.7 billion of those hours through tax-filing compliance. The Department of Health and Human Services accounts for hundreds of millions more through Medicare, Medicaid, and SNAP applications. The Department of Education’s pre-2024 FAFSA accounted for tens of millions of hours each application cycle.

The pattern across these programs is consistent: the friction was almost never designed in. It accumulated through legislative add-ons, agency self-protection, and forty years of “one more verification step won’t hurt.” The IRS’s Direct File pilot, launched in 2024, demonstrated that federal income-tax filing for simple returns could be completed in well under an hour without third-party software. That capability had been technically possible for a decade. It was a sludge problem, not a technology problem.

Healthcare: Prior Authorization and the Bandwidth Cliff

American health insurance has built one of the densest sludge fields in the modern economy. Prior authorization, the requirement that a clinician obtain insurer approval before performing a procedure or prescribing a medication, accounts for an estimated 14 hours of physician and staff time per provider per week, according to repeated American Medical Association surveys. The downstream effect lands on patients, who experience treatment delays that have been documented to produce measurable harm in oncology, cardiology, and behavioral health.

The 2024 Centers for Medicare & Medicaid Services prior-authorization rule attempted to compress decision times and require electronic submission, but the underlying friction remains. From the user’s seat, a denied claim that requires a phone call, an appeal letter, a re-submission, and a thirty-day wait is functionally identical to a denial. Sludge converts a “covered benefit” into an effective exclusion without any policy change actually being made. This is the sludge-as-policy-by-other-means pattern that Herd and Moynihan described, implemented at scale by private payers rather than public agencies.

Subscriptions: The Cancel-Flow Maze

The consumer-software industry built much of its retention economy on the asymmetry between sign-up and cancel flows. Sign-up takes thirty seconds and one click. Cancellation takes a chat agent, a hold queue, a “we have a special offer for you” deflection, and sometimes a physical phone call. The Federal Trade Commission’s “Click-to-Cancel” rule (proposed 2023, finalized October 2024, briefly in force from January 2025, then vacated by the Eighth Circuit in July 2025 on procedural grounds, with the FTC signaling re-promulgation) attempted to require symmetry: if you can sign up online, you must be able to cancel online, in the same number of steps or fewer. The rule’s existence — and the litigation that erased it — is itself a confession that the underlying sludge had become extractive at the level of an entire industry’s business model, and that the political fight over how much to formalize the audit is still live.

The interesting design question is not whether the regulation is sufficient. It is whether a product team would have caught their own cancel-flow sludge in a self-audit without the regulation forcing it. Most would not, because the cancel flow was meeting its KPI: retention. From inside the company, sludge that improves a metric you are paid to optimize is invisible. The audit only works when an outside lens forces the question of whether the metric itself is defensible.

Civic Participation: Voter Registration as a Designable Surface

Voter registration is the most politically contested sludge surface in the United States, and it is the one where Sunstein’s analytical neutrality is most strained. Different states require different combinations of documentation, deadlines, in-person verification, and address-update procedures. Same-day registration states have measurably higher registration rates than registration-deadline states, and automatic voter registration states (Oregon, California, others) have rates higher again. The friction structure is doing work. Whether that work is “fraud prevention” or “turnout suppression” is the political fight Sludge cannot adjudicate but can at least quantify.

What the audit can clarify is that the rhetorical framing of “we are just asking for ID” obscures the variance in compliance costs across populations. A working-age, employed citizen with a state driver’s license pays a few minutes of compliance cost. A retired non-driver, a college student attending school out of state, or a recently moved tenant without a current utility bill pays a substantially higher cost. The audit does not have to take a position on whether that variance is acceptable; it just has to make the variance visible.

The Elephant in the Room: Good Sludge

The instinct after reading a Sludge audit is to declare war on all friction. That instinct is wrong, and Sunstein is the first to say so. The harder design problem is not removing friction; it is keeping the friction that protects users while removing the friction that exploits them.

Good sludge has three signatures. First, it occurs at a high-stakes decision point where reversibility matters. The four-business-day waiting period before a gun purchase, the mandatory cooling-off period before certain contracts can be voided, the typing-confirmation before account deletion. All three create friction precisely because the action they precede is hard to undo. Second, the friction is calibrated to the irreversibility of the action it precedes; a two-second loading spinner before “Send Email” would be absurd, but a four-second hold-to-confirm before “Delete Account Permanently” is calibrated. Third, the friction’s existence and purpose are transparent to the user; they understand why the system is slowing them down and would, on reflection, endorse the slowdown.

Bad sludge inverts each of those signatures. It occurs at low-stakes points where reversibility is easy. Its magnitude is grossly disproportionate to the decision (a fifteen-minute cancel flow for a five-dollar-a-month subscription). And its purpose is opaque or actively hidden. The user is not told they are being slowed down, only that the system “needs additional verification.”

The audit prescription is therefore not “remove all friction” but “make the friction’s contract with the user explicit.” A friction that the system can defend out loud is probably justified. A friction the system would be embarrassed to defend out loud is sludge. The test is whether the company would publish, on its homepage, the actual reason for every step in its cancel flow. The cancel flows that would survive that test are the cancel flows that should have existed all along.

The deeper point, and the one most product teams resist, is that the friction inventory belongs in a public-facing document, not an internal Confluence page. When the friction has to defend itself to the user, the indefensible frictions tend to disappear on their own. Transparency is the dominant counter-design move against bad sludge, and it is also the move companies are least willing to make voluntarily.

How to Apply Sludge with the Octalysis Framework

Sludge gives behavioral designers an inventory tool. The Octalysis Framework supplies the missing layer: a map of which Core Drives a given friction is sabotaging, which lets the audit’s findings translate into specific design fixes rather than a generic “reduce friction” mandate.

Run any flow through the eight Core Drives and you will discover that “friction” is a category collapse. Different frictions damage different drives, and the design fix that restores Core Drive 2 (CD2): Development & Accomplishment is different from the fix that restores Core Drive 8 (CD8): Loss & Avoidance. The Octalysis lens forces the audit to specify which drive the sludge is breaking, which makes the remediation actionable.

Core Drive 1 (CD1): Epic Meaning & Calling tends to be relatively immune to sludge damage when present. A user who believes the action they are taking serves a meaningful purpose will tolerate substantially more friction than a user who experiences the action as transactional. This is why naturalization paperwork, however heavy, has higher completion rates than equally heavy commercial paperwork. The CD1 elevation absorbs the cost. The design implication is that where you cannot remove the friction, you can sometimes elevate the meaning of the action, and the second move sometimes substitutes for the first.

Core Drive 2: Development & Accomplishment is where most sludge does its first damage. The drive depends on visible progress feedback, and almost every sludge-heavy flow breaks the progress feedback loop. The user fills out page three of seven and has no idea whether they are 40% done, 30% done, or about to be told that everything they entered on page two needs to be redone. The fix is a visible, accurate, monotonically advancing progress indicator that does not lie about completion percentage. The cost of building it is trivial. The fact that it is so often absent is itself a diagnostic — the team is not measuring CD2.

Core Drive 3: Empowerment of Creativity & Feedback is the drive most damaged by black-box validation. When a user submits a form and waits seven business days for a yes/no answer with no intermediate feedback, the system has stripped them of agency for the duration of the wait. The fix is real-time validation at the field level, instant error messaging, and confirmation that the input was received and processed. The brain experiences instant, specific feedback as a small reward; it experiences delayed, vague feedback as a small punishment.

Core Drive 4 (CD4): Ownership & Possession is the drive sludge can erode most invisibly. A user who has filled out half of a multi-page form has, in a real sense, invested labor that belongs to them. If the form does not save partial progress, the system has confiscated that investment. The fix is auto-save on every meaningful state change, with a clear “you can come back and finish later” affordance. The CD4 erosion is what produces the disproportionate frustration users report on long forms that lose state.

Core Drive 5: Social Influence & Relatedness is where sludge produces the regressive effects that the audit’s psychological-cost column tries to capture. When the cost of completing a flow varies by population (by income, by parenting load, by chronic-condition status, by language fluency), sludge produces exclusion. The user who cannot complete it is not the user who lacks motivation; it is the user who lacks bandwidth. The fix is either reducing the flow’s universal cost or, where that is not possible, providing assisted-completion pathways (call lines staffed with humans, in-person navigators, multilingual support) so the flow does not silently sort by bandwidth.

Core Drive 6 (CD6): Scarcity & Impatience can be weaponized through sludge in dangerous ways. The “limited-time offer” that is technically still available but requires a fifteen-minute verification process to claim is using sludge to construct artificial scarcity. The user who gives up has been denied the offer not by the offer’s terms but by the operational architecture around it. Sludge that weaponizes CD6 is one of the most ethically suspect patterns in modern consumer design, and one of the easiest to find with an audit.

Core Drive 7 (CD7): Unpredictability & Curiosity is the drive sludge degrades by making the process itself unpredictable in unproductive ways. A user does not mind a surprise that delights them; a user does mind a surprise that says “you also need to upload a notarized copy of this third document we never mentioned.” The fix is upfront disclosure of every step, every required document, and every expected wait time. CD7 should fire on the outcome, not on the process. When the process is the source of unpredictability, sludge is the cause.

Core Drive 8: Loss & Avoidance is the drive sludge exploits most reliably. Cancellation flows are built on CD8: every step is designed to remind the user what they will lose if they continue. The “are you sure? you’ll lose your progress, your data, your discount, your community” sequence is sludge using CD8 against the user’s own decision. The design fix is symmetry: if losing the account requires three confirmation screens, so should signing up. The asymmetry is the diagnostic.

The synthesis: a sludge audit run through the Octalysis lens does not produce a generic friction inventory. It produces a per-drive damage report: which drives are being broken, which are being weaponized, and which design moves restore each drive specifically. That specificity is what turns the audit from a complaint into a design surface.

Practical Steps: Running a Sludge Audit on Your Own Product

The audit is portable. Whether you are running a federal program, a SaaS dashboard, or a consumer-app onboarding flow, the steps are the same.

1. Pick one consequential flow and list every step end to end. Not the happy path the spec describes. The actual user path including the help-center searches, the error states, the back-button retries, the email re-confirmations. The list should be longer than you expect. Most teams cannot produce this list on first request, which is itself a finding.
2. Classify each step as protective, neutral, or extractive. Protective steps prevent harm the user would endorse preventing. Neutral steps neither help nor hurt. Extractive steps slow the user down for reasons that benefit the company at the user’s expense. Write the classification next to each step. Disagreements at this stage are the most valuable part of the audit.
3. Quantify the three cost categories for each step. Learning cost (minutes spent figuring out the step), compliance cost (minutes spent executing it), psychological cost (qualitative: frustration, anxiety, embarrassment, abandonment risk). You will not get the third one exact; that is fine. The act of writing it down forces the conversation.
4. Map each extractive step to its damaged Core Drive. Is this step breaking CD2 progress feedback? Eroding CD4 ownership? Weaponizing CD8 loss aversion? The mapping tells you which design fix to apply. Generic “reduce friction” prescriptions fail because they do not name the damaged drive.
5. Assign an owner to remove each extractive step within a defined timebox. Audits without owners are just complaints. The owner is the person whose KPIs change when the step disappears, and the timebox is whatever the next release cycle accommodates. Without these two, the audit becomes a wallpaper artifact.
6. Publish the inventory. Internal-only audits get watered down. Public inventories produce the transparency that converts indefensible frictions into deleted frictions. The publication is the enforcement mechanism. Most companies will not do this voluntarily, which is what regulation is for.
7. Re-run the audit on a fixed cadence. The threat environment, the user base, and the product all change. An audit is a snapshot; the snapshot needs refreshing. Annually is a defensible minimum for most products; quarterly is better for high-volume consumer flows.

Sludge Was the Beginning, Not the End

Sunstein’s contribution was to make a recurring phenomenon legible. The construct names something practitioners had been bumping into for decades without a shared vocabulary, and the audit gives that vocabulary a deliverable. What Sludge did not do is solve the deeper question: who decides what counts as justified friction, and what stops them from being captured by the institutions they are auditing?

The next decade of work on choice architecture has to grapple with that question seriously. The audit is necessary. It is not sufficient. The institutions that produced the sludge in the first place are usually the institutions asked to remove it, and that institutional self-audit is the design problem that the framework itself cannot solve. Where it is solved, it is usually solved by external pressure: regulators, journalists, users with enough leverage to make the cost of the sludge visible to the people who built it.

For behavioral designers, the actionable takeaway is the inversion. Every flow you ship is a sludge inventory in waiting. The only question is whether you run the audit before you ship it or whether the audit gets run on you by someone less friendly to your business model. The Octalysis lens makes the audit specific. Sunstein’s framework makes it portable. The discipline is doing both, every cycle, before someone else does it for you.

Frequently Asked Questions

What exactly is “Sludge” in Sunstein’s framework?

Sludge is excessive or unjustified friction that prevents people from taking actions they want to take, claiming benefits they are entitled to, or completing choices they have already made. Sunstein formalized the construct in his 2019 paper “Sludge Audits” and expanded it in the 2022 book Sludge: What Stops Us from Getting What We Want and What to Do about It. The qualifier “unjustified” is doing critical work. Friction that protects the user or prevents real harm is not sludge.

How is Sludge different from a nudge?

Both are design choices that alter behavior without changing economic incentives or restricting options. A nudge makes the better outcome easier to choose; sludge makes the better outcome harder to reach. Defaulting employees into a retirement plan is a nudge. Making it take fifteen minutes to cancel a subscription is sludge. The mechanism is symmetrical. The user’s interest decides which label applies.

What are the three costs Sunstein identifies?

Learning costs (the effort to figure out what is required), compliance costs (the effort to actually do it), and psychological costs (the dread, anxiety, embarrassment, and learned helplessness that accumulate during the process). The taxonomy comes from Pamela Herd and Donald Moynihan’s 2018 book Administrative Burden. Most product teams measure only compliance costs. Psychological costs are where the largest damage usually lives.

Why is Sludge described as regressive?

Because the same friction does not cost everyone the same amount. A well-resourced user pays a small bandwidth fee for a complicated form. A bandwidth-depleted user (financially stressed, sleep-deprived, parenting a sick child, navigating a chronic condition) pays a percentage of the limited bandwidth that remains. Sendhil Mullainathan and Eldar Shafir’s 2013 research on scarcity formalized why the curve bends against the people who can least afford it.

What is a Sludge audit?

A systematic walk-through of a consequential flow, step by step, classifying each friction as protective, neutral, or extractive, quantifying the learning / compliance / psychological cost of each, and assigning an owner to remove the extractive steps within a defined timebox. The audit is meant to be a deliverable, not a memo. Without owners and a timebox, it is wallpaper.

Is there such a thing as good sludge?

Yes. Friction at high-stakes decision points that protects users from regret is good sludge: gun-purchase waiting periods, account-deletion confirmations, cooling-off periods on major contracts. The test is whether the user would, on reflection, endorse the slowdown. If yes, the friction is protective; if no, it is extractive. The audit’s job is to make the distinction case by case, not to declare friction good or bad in general.

How is Sludge related to “dark patterns” (deceptive patterns)?

The two constructs overlap on a meaningful subset of cases but are not identical. Deceptive patterns describe interface designs intended to mislead users (confusing button labels, hidden costs, false urgency). Sludge describes excessive process burden regardless of whether the user was misled. A perfectly honest but excessively long form is sludge but not a deceptive pattern. A short but misleadingly labeled flow is a deceptive pattern but not necessarily sludge. The FTC and the EU treat them as related but distinct enforcement targets.

How does Sludge interact with the Octalysis Framework?

Sludge gives you the inventory. Octalysis tells you which Core Drive each friction is damaging. The same generic complaint (“this flow is too long”) becomes diagnostically specific when you ask whether the length is breaking CD2 progress feedback, eroding CD4 ownership, weaponizing CD8 loss aversion, or all three. The design fix differs by drive. Running an audit through the Octalysis lens converts a friction list into a remediation plan with named owners per drive.

References

• Sunstein, C. R. (2022). Sludge: What Stops Us from Getting What We Want and What to Do about It. Cambridge, MA: MIT Press.
• Sunstein, C. R. (2019). Sludge Audits. Behavioural Public Policy, 6(4), 654-673.
• Herd, P., & Moynihan, D. P. (2018). Administrative Burden: Policymaking by Other Means. New York: Russell Sage Foundation.
• Thaler, R. H., & Sunstein, C. R. (2008). Nudge: Improving Decisions about Health, Wealth, and Happiness. New Haven: Yale University Press.
• Mullainathan, S., & Shafir, E. (2013). Scarcity: Why Having Too Little Means So Much. New York: Times Books.
• Kable, J. W., & Glimcher, P. W. (2007). The neural correlates of subjective value during intertemporal choice. Nature Neuroscience, 10(12), 1625-1633.
• Madrian, B. C., & Shea, D. F. (2001). The power of suggestion: Inertia in 401(k) participation and savings behavior. Quarterly Journal of Economics, 116(4), 1149-1187.
• Bertrand, M., Mullainathan, S., & Shafir, E. (2004). A behavioral-economics view of poverty. American Economic Review, 94(2), 419-423.
• Brignull, H. (2010). Dark Patterns: User Interfaces Designed to Trick People (subsequently rebranded “deceptive patterns” in 2023). deceptive.design.
• Federal Trade Commission (2022). Bringing Dark Patterns to Light: An FTC Workshop Staff Report. Washington, DC: FTC.
• Office of Management and Budget (2022). Memorandum M-22-10: Improving Access to Public Benefits Programs Through the Paperwork Reduction Act. The White House.
• The White House (2022). Tackling the Time Tax: How the Federal Government Is Reducing Burdens to Accessing Critical Benefits and Services. Office of Information and Regulatory Affairs.
• American Medical Association (annual survey series). Prior Authorization (PA) Physician Survey. Chicago, IL: AMA.
• Centers for Medicare & Medicaid Services (2024). Final Rule: CMS Interoperability and Prior Authorization Rule (CMS-0057-F). Federal Register.
• Soman, D., & Cheema, A. (2011). Earmarking and partitioning: Increasing saving by low-income households. Journal of Marketing Research, 48(SPL), S14-S22.
• Chou, Y. (2015). Actionable Gamification: Beyond Points, Badges, and Leaderboards. Octalysis Media.

Related Reading

• The Octalysis Framework — The eight Core Drive lens used to map which drive each friction is damaging in the audit above.
• Nudge Theory (Thaler & Sunstein) — The construct’s mirror image and the prior work the Sludge framework deliberately inverts.
• Loss Aversion (Kahneman & Tversky) — The Core Drive 8 mechanism most often weaponized by cancellation sludge.
• The Scarcity Mindset (Mullainathan & Shafir) — Why the same friction costs different populations radically different amounts.
• The Behavioral Framework Library — The full catalog of behavioral models analyzed through the Octalysis lens.

WOULD YOU LIKE YU-KAI CHOU TO WORK WITH YOUR ORGANIZATION?