2. 1. Affordances & Signifiers: How Things Communicate
This cluster of concepts, largely codified by Don Norman in The Design of Everyday Things (1988), deals with how objects communicate their purpose and usage to humans without instruction manuals.
| Concept | Design Definition | As a Mental Model | Example |
|---|---|---|---|
| Affordance | The actions an object makes possible, determined by the relationship between the object’s properties and the user’s capabilities. A chair affords sitting. A button affords pressing. Affordances exist whether or not they are perceived. | Every situation, tool, relationship, or environment affords certain actions and resists others. The question isn’t “what can this do?” but “what does this make easy or hard for me?” Change the affordances and you change the behavior. | An open-plan office affords spontaneous conversation but resists deep focus. A private office affords focus but resists collaboration. Neither is “better” — they afford different things. Choose the environment that affords the behavior you want. |
| Signifier | A perceivable cue that communicates what action is possible or expected. Distinct from affordance: the affordance is the actual possibility; the signifier is the signal. A flat plate on a door signifies “push” even though the door also affords pulling. | How you signal what’s possible is as important as what’s actually possible. If people can’t perceive the affordance, it doesn’t exist for them. Great capabilities with poor signifiers go unused. | A brilliant product feature hidden in a submenu (poor signifier) might as well not exist. A team that has an open-door policy but whose manager always looks busy (conflicting signifier) will never get approached. The affordance exists, but the signal says otherwise. |
| Norman Door | A door that gives false signals about whether to push or pull. Named after Don Norman. The canonical example of a signifier/affordance mismatch: a door with a handle (signifies pull) that must be pushed. | Any interface — physical, digital, or social — where the signals contradict the required action. When people consistently do the “wrong” thing, the design is wrong, not the people. A “Norman Door” moment is always a design failure, never a user failure. | An employee handbook that says “we value work-life balance” (signifier: take time off) while managers email at 11 PM and promote workaholics (actual affordance: overwork is rewarded). This is a Norman Door — the signal says one thing, the system does another. People will follow the system, not the signal. |
| Mapping | The relationship between controls and their effects. Good mapping: a stovetop where burner positions match knob positions. Bad mapping: four burners in a square, four knobs in a row — which controls which? | The connection between action and effect should be intuitive and spatial. When the relationship between what you do and what happens is unclear, errors multiply. Good mapping is invisible; bad mapping creates constant confusion. | In a codebase: function names that clearly map to what they do = good mapping. A function called “processData()” that also sends emails = bad mapping. In organizations: a reporting structure where it’s unclear who is responsible for what = bad mapping. |
| Visibility | The relevant parts of a system should be visible and convey the correct message. The state of the system and the available actions should be apparent without memorization. | People can only respond to what they can see. Hidden state, invisible progress, and unclear options create anxiety and errors. Make the important things visible. Hide the unimportant things. | A team using a shared Kanban board (visible state) vs. a team where tasks live in individual to-do lists (invisible state). The first team self-coordinates; the second requires constant status meetings to make state visible. Good visibility reduces the need for communication overhead. |
| Feedback | The system must confirm that an action has been received and show its result. A button that clicks, a progress bar that fills, a receipt that prints. Without feedback, users don’t know if their action worked. | After any action, people need to know: did it work? The absence of feedback creates anxiety and repeated attempts (double-clicking, re-sending emails, asking “did you get my message?”). Fast, clear feedback is one of the cheapest ways to improve any system. | A manager who receives a proposal and responds three days later (“looks good”) vs. one who immediately responds “received, will review by Thursday.” Same outcome, radically different experience. The second eliminated three days of anxiety with five seconds of feedback. |
| Conceptual Model | The mental model a user constructs about how a system works. It doesn’t need to be technically accurate — it needs to be functional. A thermostat: most people’s mental model (“higher setting = faster heating”) is wrong but works well enough for most contexts. | People operate on mental models, not reality. If their model is wrong but functional, correcting it is unnecessary. If their model leads to errors, the design must guide them toward a better model — not through documentation, but through the design itself. | Users think of “the cloud” as a magical, infinite storage space. Technically wrong. Functionally adequate for 99% of use cases. If you’re designing a cloud product, work with this model, not against it. Fighting users’ mental models is a losing battle. |
3. 2. Form & Function: The Eternal Tension
| Concept | Design Definition | As a Mental Model | Example |
|---|---|---|---|
| Form Follows Function | Coined by architect Louis Sullivan (1896). The shape of an object should be primarily based on its intended function. Ornament that doesn’t serve function is waste. The Bauhaus school and modernist design built entire movements on this principle. | Structure should emerge from purpose, not the other way around. Don’t start with how something looks, start with what it needs to do. Apply to organizations: the org chart should follow the work, not the other way around. | A startup that creates VP titles before having enough work for VPs has put form before function. An org that reorganizes around customer segments (function) rather than historical departments (form) is following this principle. Conway’s Law is the software version: system structure follows org structure. |
| Form Follows Emotion | Counter-argument to pure functionalism. Hartmut Esslinger (frog design): people don’t just use products, they feel about them. Form communicates identity, status, values, and belonging beyond mere function. | Function is necessary but not sufficient. People choose based on feeling, then rationalize with function. A purely functional solution that ignores emotional resonance will lose to a slightly less functional one that makes people feel something. | A plain-text resume (pure function) vs. a beautifully typeset resume (form + function). Both convey the same information. The second communicates “I care about craft” — an emotional signal. Apple products are rarely the most functional; they win on emotion. |
| Semantic Design | Objects communicate meaning through their form. A sports car “looks fast” even standing still. A bank building “looks stable.” The form carries semantic content beyond function. | Everything communicates, whether you intend it to or not. Your product, your office, your website, your email tone — all carry semantic meaning. The question isn’t whether you’re sending signals, but whether you’re sending the right ones. | A startup with a playful, colorful brand targeting enterprise security buyers is sending a semantic mismatch. The product may be excellent; the form says “we’re not serious.” Semantic design means aligning every signal — visual, verbal, behavioral — with what you want to communicate. |
| Skeuomorphism | Retaining visual cues from an older technology in a newer one. Early iOS: the Notes app looked like a yellow legal pad. The Contacts app looked like a leather address book. These ornamental references to physical objects helped users understand new digital interfaces by connecting them to existing mental models. | When introducing something radically new, anchor it to something familiar. Skeuomorphism is a bridge between the known and the unknown. It’s not dishonest — it’s a teaching tool. Once the new concept is understood, the skeuomorphic cues can be removed (Apple’s shift to flat design in iOS 7). | Calling a startup funding round a “seed” (agriculture metaphor) or describing cloud infrastructure as “servers” and “containers” are skeuomorphic language — borrowing from the physical world to explain the abstract. “Desktop,” “files,” “folders,” “trash can” — every computer interface is built on skeuomorphic metaphors. |
| Honest Design | Dieter Rams, Principle 6: “Good design is honest. It does not make a product more innovative, powerful or valuable than it really is. It does not attempt to manipulate the consumer with promises that cannot be kept.” | Don’t make things appear to be what they are not. Products, pitches, and promises that over-represent their capabilities create disappointment that destroys trust. Honest design under-promises, and the quality of the actual experience speaks for itself. | A SaaS landing page showing a beautiful dashboard mock-up that doesn’t reflect the actual product = dishonest design. Users sign up, see reality, feel deceived, churn. Contrast with products that under-promise visually but over-deliver on experience. The gap between expectation and reality is the design’s integrity. |
4. 3. Human Factors & Ergonomics: Designing for Bodies and Minds
| Concept | Design Definition | As a Mental Model | Example |
|---|---|---|---|
| Ergonomics | Designing products and environments to fit the human body and its capabilities. Not forcing the human to adapt to the tool, but shaping the tool to the human. Covers physical (body), cognitive (mind), and organizational (systems) ergonomics. | Adapt the system to the human, not the human to the system. If people are struggling, the system is wrong, not the people. This applies to software interfaces, organizational processes, management styles, and physical spaces. | An onboarding process that requires new hires to fill out 14 forms on their first day (“the system requires it”) has forced the human to adapt to the system. An ergonomic process would spread forms over the first week, pre-fill what’s known, and prioritize the human experience over the system’s needs. |
| Cognitive Load | The total amount of mental effort required to use a product. Working memory is limited (~4 items). Every option, label, and decision point consumes cognitive resources. Overloaded users make errors, feel frustrated, and abandon the task. | Attention is the scarcest resource. Every additional choice, piece of information, or step in a process depletes the user’s cognitive budget. The designer’s job is to spend this budget wisely and never waste it on unnecessary complexity. | A restaurant menu with 200 items creates decision paralysis (high cognitive load). A menu with 12 carefully curated items creates confident selection (low cognitive load). The same applies to product feature lists, meeting agendas, and strategy documents. Curation is a form of design. |
| Fitts’s Law | The time to reach a target is a function of the distance to it and its size. Larger, closer targets are faster to hit. This governs button sizes, menu placement, and tool accessibility in physical and digital design. | Make important things big and close. Make dangerous things small and far away. Reduce the effort to do the right thing and increase the effort to do the wrong thing. This is physical choice architecture. | The “Buy Now” button is large and prominent (big, close). The “Cancel Subscription” link is small and buried (small, far). This is Fitts’s Law applied as business strategy. In a team: the easiest-to-reach communication channel becomes the default, regardless of policy. |
| Hick’s Law | Decision time increases logarithmically with the number of choices. 2 options = fast decision. 10 options = slow. 50 options = paralysis. Each additional option has diminishing returns on value but increasing cost on decision time. | More options is not always better. Beyond a threshold, additional options decrease satisfaction (paradox of choice). Curate ruthlessly. If you must offer many options, structure them hierarchically so each decision point has 3–7 choices, not 50. | A pricing page with 3 plans (Good, Better, Best) converts better than one with 7 plans. A startup pitch with 3 key metrics is more convincing than one with 12. A meeting agenda with 3 topics gets through all of them; one with 8 topics finishes none. |
| The 95th Percentile | In ergonomic design, accommodate the 5th to 95th percentile of users. Not the average user (who doesn’t exist) but the range. Designing for the average fits nobody well. Adjustability covers the range better than a single “average” setting. | There is no average user, customer, employee, or reader. Design for the range, not the mean. Adjustability and flexibility beat one-size-fits-all solutions. If you can only design for one persona, pick the most constrained one — what works for them usually works for everyone. | A desk at a fixed 29-inch height “fits” the average person but is uncomfortable for the 5th percentile (short) and 95th percentile (tall). An adjustable desk fits everyone. Same for meeting length: a fixed 1-hour meeting is too long for quick syncs and too short for deep dives. Adjustable format > fixed format. |
5. 4. Emotional Design: The Three Levels of Experience
Don Norman’s Emotional Design (2004) argues that design operates on three processing levels, and all three must be addressed for a product to succeed.
| Level | Design Definition | As a Mental Model | Example |
|---|---|---|---|
| Visceral (Instinct) | The immediate, pre-conscious response to how something looks, feels, sounds, or smells. First impressions. Gut reactions. Happens in milliseconds before rational thought engages. Driven by evolved preferences: symmetry, color, proportion. | First impressions are formed before rational evaluation. People decide if they like something within seconds, then spend the rest of the time rationalizing their instinct. If the visceral response is negative, you may never get the chance to demonstrate substance. | A website that takes 4 seconds to load creates a negative visceral response before any content is seen. A pitch deck with poor typography signals “amateur” before a single word is read. A product with satisfying packaging creates positive anticipation before unboxing. The visceral level is the gate — fail here and nothing else matters. |
| Behavioral (Use) | The experience of actually using the product. Function, usability, performance, reliability. This is where the product either delivers on the visceral promise or betrays it. Evaluated unconsciously during use: does it feel right, respond correctly, work reliably? | After the first impression, substance takes over. Does it actually work? Is it pleasant to use? Does it do what was promised? The behavioral level is where trust is built or destroyed. Visceral attracts; behavioral retains. | A beautiful app that crashes constantly: visceral = great, behavioral = terrible. Users leave. A plain-looking tool that works perfectly every time: visceral = mediocre, behavioral = excellent. Users stay and eventually love it. The ideal is both, but if forced to choose, behavioral wins long-term. |
| Reflective (Identity) | The conscious, post-use evaluation. What does owning/using this say about me? Does it align with my self-image? The story I tell about the product. The reflective level is cultural, personal, and the most variable across users. | People don’t just use products; they use products to construct identity. “I’m the kind of person who uses [X]” is one of the most powerful forces in consumer behavior. The reflective level is why branding exists: it’s not about the product, it’s about what the product says about you. | A developer using Vim/Neovim (reflective: “I’m a serious programmer”). Someone driving a Tesla (reflective: “I’m tech-forward and eco-conscious”). A company using “we’re a Notion shop” as an identity marker. The reflective level explains why technically inferior products can win if they carry stronger identity signals. |
The Hierarchy and Its Implications
The three levels interact:
- Visceral gates behavioral: If first impressions fail, the product never gets used enough for behavioral evaluation.
- Behavioral gates reflective: If the product doesn’t work well, no amount of brand identity saves it long-term.
- Reflective overrides both: A product with strong identity value (luxury, status, tribe membership) can survive poor visceral and behavioral scores — up to a point.
In practice: invest minimum viable effort in visceral (clear the gate), maximum effort in behavioral (earn retention), and strategic effort in reflective (create word-of-mouth and loyalty).
6. 5. Error Prevention & Forgiveness: Designing for Failure
| Concept | Design Definition | As a Mental Model | Example |
|---|---|---|---|
| Poka-Yoke (Error-Proofing) | Japanese manufacturing concept from Shigeo Shingo (1960s). Design the process so that errors are physically impossible or immediately detected. A USB plug that only fits one way. A gas nozzle that doesn’t fit in a diesel tank. The design prevents the error, not the user’s attention. | One of the most powerful design mental models. Don’t rely on users being careful. Design so that the wrong action is impossible, not just discouraged. Constraints that prevent errors are infinitely more reliable than training, warnings, or willpower. | A deployment pipeline that requires passing tests before merging (poka-yoke: can’t deploy broken code). Contrast with a policy document saying “please run tests before deploying” (relying on human attention). A database schema that enforces NOT NULL (poka-yoke) vs. documentation saying “this field should always have a value” (hope). |
| Forcing Function | A constraint that forces the correct behavior. A microwave that stops when you open the door. A car that won’t start unless the brake is pressed. An ATM that returns your card before dispensing cash (forces you to take your card). | When a behavior is critical, don’t request it — force it. A forcing function makes the desired behavior the only possible path. This is stronger than incentives, training, or culture. | Code review required before merge (forcing function). Contrast with “we encourage code review” (suggestion). A checklist that must be completed before a surgery begins. A meeting that can’t be booked without an agenda. Each forces the desired behavior structurally. |
| Undo / Reversibility | Allow users to reverse their actions. Ctrl+Z is the most beloved design pattern in computing. When errors are easily reversed, users explore more confidently and recover faster from mistakes. | Systems that allow reversal encourage experimentation and reduce the cost of errors. Systems that don’t allow reversal create fear and paralysis. Make as many actions as possible reversible, and clearly mark the irreversible ones. | Gmail’s “Undo Send” (5-second window to reverse). Git’s branching model (experiment freely, revert if wrong). An organization that treats a restructuring as a permanent, irreversible decree vs. one that says “we’ll try this for 3 months and adjust” (designed reversibility). |
| Progressive Disclosure | Show only what’s needed at each stage, revealing complexity gradually as the user needs it. Basic options first; advanced options behind a “More” button. Reduces initial cognitive load while preserving full capability for power users. | Don’t dump everything on people at once. Reveal complexity in layers, starting with the simplest and most common. This applies to product interfaces, onboarding flows, meeting structures, and teaching. | A new hire’s first day: they need to know where to sit, how to get lunch, and who their manager is (layer 1). They do NOT need the full corporate strategy, benefits enrollment, and 401k options on day one (layer 5). Most onboarding failures are progressive disclosure failures — too much, too soon. |
| Confirmation Dialog | Before an irreversible or high-consequence action, ask the user to confirm. “Are you sure you want to delete this?” A speed bump before a cliff. The friction is intentional. | For high-stakes, irreversible actions, add intentional friction. But be careful: too many confirmations trains users to click “yes” reflexively (confirmation fatigue), which is worse than no confirmation at all. Reserve confirmation for truly dangerous actions. | Deleting a production database: requires typing the database name to confirm (strong confirmation). Sending an email: no confirmation needed (low stakes, reversible). A startup deciding to pivot: should require a “confirmation dialog” (deliberate, structured discussion) not an impulsive Slack message. |
7. 6. Constraints & Trade-offs: The Art of Choosing
| Concept | Design Definition | As a Mental Model | Example |
|---|---|---|---|
| Design Constraints | The non-negotiable boundaries within which a design must operate: manufacturing cost, material availability, safety regulations, size limits, weight limits, user capabilities. Constraints are not obstacles — they are the problem definition. | Constraints are generative, not restrictive. The most creative solutions emerge from the tightest constraints. Unconstrained design produces bloat, indecision, and mediocrity. Embrace constraints; they tell you what problem you’re actually solving. | Twitter’s 140-character limit (now 280) was a constraint that created a new form of communication. Haiku’s 5-7-5 syllable constraint creates poetry. A startup’s limited budget forces focus on the one feature that matters. Remove the constraint and the focus disappears. |
| CMF (Color, Material, Finish) | The three surface-level decisions that disproportionately affect perception: what color is it, what is it made of, and what texture does its surface have. CMF choices communicate quality, brand, price point, and target user without changing function at all. | Surface presentation creates perception, independent of substance. The “CMF” of your work — how it looks, what it’s packaged in, what quality signals it carries — affects reception as much as the content itself. This isn’t superficial; it’s communication. | The same business proposal in a Google Doc (cheap CMF) vs. in a well-formatted PDF with company branding (premium CMF). Same content, different reception. The same meal on a paper plate vs. on ceramic. CMF is the reason packaging design is a multi-billion-dollar industry. |
| DFM (Design for Manufacturing) | Design the product with manufacturing constraints in mind from the start. Every curve, angle, and material choice must be producible at the target cost, volume, and quality. A beautiful design that can’t be manufactured is a sketch, not a product. | Design with the delivery mechanism in mind. A strategy that can’t be executed is a wish. A product that can’t be built at scale is a prototype. A plan that can’t be staffed is a fantasy. DFM thinking means always asking: “Can we actually make this?” | A product roadmap full of features that would require 3x the current engineering team = not designed for manufacturing. A go-to-market plan that assumes distribution channels that don’t exist = not designed for manufacturing. Always design the plan to be producible, not just the product. |
| The Iron Triangle | Good, fast, cheap — pick two. In product development, you can optimize for quality, speed, or cost, but never all three simultaneously. Changing one always affects the others. | In any endeavor, there are irreducible trade-offs. Pretending they don’t exist leads to failure on all three dimensions. The mature response is to explicitly choose which dimension to sacrifice and communicate that trade-off clearly. | A startup must choose: ship fast with lower quality (speed + cheap), ship high quality slowly (quality + cheap), or ship fast and high quality with expensive talent (quality + speed). Saying “we want all three” isn’t ambition — it’s denial of physics. |
| Universal Design | Design products and environments to be usable by the widest range of people, without requiring adaptation. The 7 principles: equitable use, flexibility, simple and intuitive, perceptible information, tolerance for error, low physical effort, appropriate size and space. | Designing for the margins improves the experience for everyone. Curb cuts were designed for wheelchairs but benefit parents with strollers, delivery workers with carts, and travelers with suitcases. Closed captions were for the deaf but are used by everyone in noisy gyms and quiet libraries. | Making your product accessible to users with disabilities (screen readers, keyboard navigation, high contrast) simultaneously improves the experience for power users (keyboard shortcuts), users in non-ideal conditions (bright sunlight), and SEO (structured content). The “edge case” design improves the mainstream experience. |
8. 7. User Behavior & Desire Paths: How People Actually Act
| Concept | Design Definition | As a Mental Model | Example |
|---|---|---|---|
| Desire Paths | Unofficial trails worn into grass by pedestrians taking shortcuts instead of following paved walkways. The path people actually take vs. the path the designer intended. Universities famously observe desire paths before paving walkways — letting human behavior determine the design. | One of the most important design mental models. Observe what people actually do, not what they say they’ll do or what you designed them to do. When behavior deviates from design, the design is wrong. Pave the desire paths. | Users are using your search bar to navigate (instead of the menu) = desire path. Pave it: make search the primary navigation. Employees are using Slack DMs instead of the official project management tool = desire path. Pave it: accept that Slack is the real communication channel and integrate with it, rather than fighting it. |
| Desire Path Inversion | Instead of waiting for desire paths to emerge, deliberately create conditions that make the desired behavior the natural path. Place the walkway where you want people to walk and make it the most attractive, easiest option. | Don’t just observe behavior — shape the environment to make the desired behavior the path of least resistance. This is the difference between reactive design (observing desire paths) and proactive design (creating them). | A company that wants knowledge sharing: instead of mandating wiki updates (paving an arbitrary path), they make the daily standup include a “TIL” (Today I Learned) slot and auto-post it to a Slack channel (creating a desire path where sharing is the easiest option). |
| Learned Helplessness in Design | When users encounter too many confusing or broken interfaces, they stop trying to learn new ones. They develop a default assumption that new things won’t work and revert to familiar patterns regardless of improvement. | Repeated bad experiences create a defensive posture where people stop engaging with new options, even beneficial ones. If your users/employees/customers have been burned before, your first design challenge is rebuilding trust, not showcasing features. | A team whose previous three project management tools were abandoned after 2 months will be resistant to the fourth, no matter how good it is. The tool isn’t the problem; the learned helplessness is. Address the pattern, not just the product. |
| Subway Map Thinking | Harry Beck’s 1931 London Tube map sacrificed geographic accuracy for usability. Stations are evenly spaced, lines are straightened, and only topology (connections) is preserved. It’s “wrong” geographically but perfect for its purpose: navigating the system. | The best representation is not the most accurate one — it’s the one that best serves the user’s task. Simplification that removes irrelevant details and amplifies relevant ones creates clarity. Don’t show what’s true; show what’s useful. | An architecture diagram that shows every microservice and database (geographically accurate) vs. one that shows the 5 main flows a new engineer needs to understand (subway map). The second is “wrong” but far more useful. Good documentation is subway map design. |
9. 8. Simplicity & Complexity: The Rams Tradition
Dieter Rams’s ten principles of good design, developed in the late 1970s while at Braun, remain the most influential articulation of design values. Several of them function as standalone mental models.
| Rams Principle | What It Means | As a Mental Model | Example |
|---|---|---|---|
| “As little design as possible” | Rams’s 10th principle. Concentrate on the essential aspects. Back to purity, back to simplicity. “Less, but better” — Weniger, aber besser. | The best work is the minimum necessary to achieve the goal. Everything beyond that is waste that degrades the result. The courage is in what you remove, not what you add. This is the hardest design discipline: restraint. | A product with 5 well-designed features beats one with 50 mediocre features. A strategy deck with 3 clear slides beats one with 40 dense slides. An email with 2 sentences beats one with 8 paragraphs. In every domain, editing is the highest form of design. |
| “Good design is unobtrusive” | Rams’s 5th principle. Products should be like tools — neither decorative objects nor works of art. Their design should be neutral, restrained, and leave room for user self-expression. | The best systems are invisible. They don’t demand attention; they enable action. When you notice the design, something has gone wrong. Infrastructure, processes, and tools should be felt, not seen. | Good typography is invisible: you read the words, not the font. Good project management is invisible: work flows without people thinking about process. Good management is invisible: the team succeeds and feels it was their own doing. The highest compliment is “it just works.” |
| “Good design is long-lasting” | Rams’s 7th principle. It avoids being fashionable and therefore never appears antiquated. Unlike fashionable design, it lasts many years — even in today’s throwaway society. | Optimize for durability over trendiness. Things designed for the current moment become obsolete when the moment passes. Things designed on fundamental principles remain relevant because the fundamentals don’t change. | A Braun clock from 1968 still looks modern. A product designed around “Web 2.0 gradients” in 2008 looks dated. Code written with clean, fundamental patterns ages better than code that uses every fashionable framework. Writing in clear prose ages better than jargon-laden buzzwords. |
| “Good design is thorough down to the last detail” | Rams’s 8th principle. Nothing must be arbitrary or left to chance. Care and accuracy in the design process show respect towards the user. | Quality is fractal: it must be consistent at every scale. A product that is beautiful on the surface but sloppy in the details signals that the creators didn’t really care. People sense this, even if they can’t articulate what’s wrong. | A startup with a beautiful website but 404 errors, a gorgeous app with typos in error messages, or a polished pitch with a poorly formatted financial model. The detail mismatch creates cognitive dissonance. Quality must be consistent — a chain is only as strong as its weakest link. |
| Essential Complexity vs. Accidental Complexity | Fred Brooks (1986): essential complexity is inherent in the problem itself. Accidental complexity is introduced by the solution. Good design minimizes accidental complexity; it cannot reduce essential complexity. | Before simplifying, distinguish between complexity you can’t avoid (the problem is genuinely hard) and complexity you introduced (your solution is poorly designed). Most systems suffer from 80% accidental complexity. Attack that first. | Tax software is complex because taxes are complex (essential). Tax software that requires entering the same information three times is complex because it’s poorly designed (accidental). An org chart with 7 levels of hierarchy for a 50-person company is accidental complexity. The work requires maybe 3 levels. |
10. 9. Obsolescence & Longevity: Designing for Time
| Concept | Design Definition | As a Mental Model | Example |
|---|---|---|---|
| Planned Obsolescence | Deliberately designing a product to have a limited useful life, forcing replacement. Coined by Brooks Stevens (1954). Three types: functional (ceases to work), quality (degrades to unacceptable level), and desirability (becomes “unfashionable”). | Consider: what is the intended lifespan of what you’re building? Some things should be permanent (infrastructure, principles, culture). Some should be deliberately temporary (experiments, campaigns, interim processes). Problems arise when you build temporary things that become permanent, or permanent things that become obsolete. | A “temporary” spreadsheet process that becomes the company’s billing system for 5 years = accidental permanence of a temporary design. A monolithic codebase built to “last forever” that needs complete rewriting after 3 years = accidental temporariness of a permanent design. Be explicit about intended lifespan. |
| Patina | The way a material ages. Leather develops a patina that many find beautiful. Cheap plastic degrades unattractively. Good materials age gracefully; cheap materials deteriorate. Some products are designed to look better with use (wabi-sabi), not worse. | Does your work age well or poorly? Some things gain character and value with time (good writing, well-maintained relationships, compound skills). Others degrade (trendy tactics, shallow networks, narrow expertise). Build things that develop a beautiful patina rather than an ugly one. | A well-maintained open-source project gains contributors, documentation, and trust over time (positive patina). A growth-hacked product built on dark patterns accumulates negative reviews, regulatory risk, and user resentment over time (negative patina). Choose strategies that age gracefully. |
| Repairability | Designing products so they can be fixed when components fail, rather than replaced entirely. Modular construction, standard fasteners, available spare parts, published repair guides. The “right to repair” movement. | Build things that can be repaired, not just replaced. This applies to relationships (capacity for repair after conflict), teams (ability to recover from setbacks), and products (users can fix issues themselves). Repairability is a form of resilience. | A codebase where a broken module can be fixed or replaced without touching the rest = repairable. A codebase where fixing one bug creates three more = unrepairable. A team culture where conflict is addressed directly and resolved = repairable. A culture where conflict festers into resentment = unrepairable. |
11. 10. Manufacturing Reality: Where Design Meets Physics
| Concept | Design Definition | As a Mental Model | Example |
|---|---|---|---|
| Bill of Materials (BOM) | The complete list of raw materials, components, and quantities needed to build one unit of a product. The BOM is the ground truth: it tells you exactly what the product is, not what you wish it were. | For any plan, project, or goal, there is a bill of materials: the actual resources, skills, time, and components required. The BOM is the antidote to wishful thinking. Write it down. If you can’t enumerate the BOM, you don’t understand what you’re building. | A product launch BOM: engineering time (X sprints), design assets (Y pages), marketing copy (Z pieces), QA (W hours), customer support preparation. If any component is missing, the launch is incomplete. Most launch failures are BOM failures — someone forgot a component. |
| Minimum Order Quantity (MOQ) | The smallest number of units a manufacturer will produce. Below the MOQ, per-unit costs are prohibitively high. The MOQ creates a threshold: you must commit to scale or pay a premium for small batches. | Many activities have implicit MOQs. Hiring has an MOQ (you can’t hire 0.3 of an engineer). Content marketing has an MOQ (one blog post doesn’t work; you need a sustained volume). Trust-building has an MOQ (one gesture isn’t enough; you need consistent behavior over time). | A company that tries SEO with 3 blog posts (“SEO doesn’t work for us”) hasn’t met the MOQ. SEO’s MOQ is probably 50–100 quality pages before meaningful traffic. A founder who attends one networking event = below the MOQ for relationship-building. Meeting the MOQ is the prerequisite for seeing results. |
| Injection Molding Thinking | Injection molding has massive upfront cost (the mold) but near-zero marginal cost per unit. The economics are terrible at low volume and incredible at high volume. This creates a commitment: once you invest in the mold, you’re committed to high volume to amortize the cost. | Some strategies have high fixed costs and low marginal costs. They only make sense at scale. Before committing, ask: am I confident I’ll reach the volume needed to justify the fixed investment? If not, use a higher-marginal-cost but lower-fixed-cost approach (3D printing instead of injection molding). | Building a custom platform (high fixed cost, low marginal) vs. using off-the-shelf tools (low fixed cost, higher marginal). At 10 users, use Airtable. At 10,000 users, build the custom platform. The switch point is your “injection molding decision.” Most startups invest in the mold too early. |
| Assembly Sequence | The order in which parts are assembled matters. Some sequences are efficient; others are impossible (you can’t install an internal component after the enclosure is sealed). The sequence constrains the design. | The order of operations matters. Some things must come first because they’re prerequisites for everything else. Trying to do step 5 before step 2 creates rework. Map the dependency graph and work in the right order. | Hiring a sales team before having a product (wrong assembly sequence). Building a complex feature before validating the core value proposition. Scaling marketing before product-market fit. Each is an assembly sequence error — the right parts in the wrong order. |
12. 11. The Design Process: How to Think Like a Designer
| Concept | Design Definition | As a Mental Model | Example |
|---|---|---|---|
| Empathy First | Design thinking begins with empathy: deeply understanding the user’s actual experience, needs, frustrations, and context. Not what you think they need — what they actually experience. Observation > assumption. | Before solving any problem, understand it from the perspective of the person experiencing it. Your mental model of their experience is almost certainly wrong. Go observe. Ask open questions. Sit in their chair, literally or metaphorically. | A product team that redesigns the checkout flow based on analytics data (they see drop-off) vs. one that watches 20 users try to check out and sees them confused by the shipping options (they understand why). Both identify the same problem; only the second finds the right solution. |
| Diverge Then Converge | The “double diamond” design process: first diverge (generate many ideas without judgment), then converge (select and refine the best ones). Trying to diverge and converge simultaneously produces mediocre results. | Separate the creative phase (generate options) from the evaluative phase (choose among them). Brainstorming while simultaneously critiquing kills creativity. Choosing without first generating options leads to premature commitment. | A strategy meeting where someone proposes an idea and it’s immediately debated = premature convergence. A meeting with 15 minutes of “all ideas welcome, no evaluation” followed by 15 minutes of “now let’s evaluate” = proper diverge-then-converge. The process feels slower but produces better outcomes. |
| Prototype to Think | In design, you don’t think your way to a good solution and then build it. You build your way to understanding. A prototype is a thinking tool, not a pre-production artifact. The purpose is to learn, not to ship. | When stuck, build something — anything. A rough prototype, a sketch, a spreadsheet model, a draft. Thinking in the abstract hits diminishing returns quickly. Making something concrete reveals problems and possibilities that pure thought never finds. | A product discussion that goes in circles for three meetings (“what if we…?”) vs. one person who builds a clickable prototype in an afternoon and the team reacts to something concrete. The prototype answers more questions in 30 minutes than the discussion answered in 3 hours. |
| Kill Your Darlings | The willingness to discard a beloved design element if it doesn’t serve the whole. Coined by Faulkner (writing) but deeply practiced in design. Sunk cost in a beautiful component that doesn’t fit is still sunk cost. | Don’t fall in love with your solution. Fall in love with the problem. A beautiful feature, a clever idea, or an elegant implementation that doesn’t serve the user is waste. The ability to discard your own work is the mark of a mature creator. | A feature that took 3 months to build but confuses users in testing. The sunk cost fallacy screams “ship it anyway.” The design mindset says: cut it. An argument you spent hours crafting but that doesn’t persuade your audience: cut it. The work served its purpose (you learned); now discard the artifact. |
| Design for the Extremes | Design for the extreme users first, and the mainstream will be well-served. If a kitchen tool works for someone with arthritis, it works beautifully for everyone. If software is accessible to a blind user, it’s exceptionally well-structured for all users. | The edge cases reveal the weaknesses in your design. If your system works for the most demanding, unusual, or constrained user, it works for everyone. The mainstream user benefits from the robustness created by designing for the extremes. | OXO Good Grips: designed for people with arthritis (extreme case), became the best-selling kitchen tools for everyone. Curb cuts: designed for wheelchairs, used by every parent with a stroller and traveler with a suitcase. The extreme case is the design brief. |
| Wabi-Sabi | Japanese aesthetic concept that finds beauty in imperfection, transience, and incompleteness. A cracked ceramic bowl repaired with gold (kintsugi). A weathered wooden surface. The beauty is in the evidence of time, use, and life — not in pristine perfection. | Not everything needs to be polished to be valuable. Some things are more beautiful, authentic, and trustworthy because of their imperfections. The pursuit of perfection can destroy the humanity, warmth, and character that makes something resonate. | A hand-written note vs. a perfectly typeset card. A startup founder who openly shares their failures vs. one who presents a polished facade. A product with a few rough edges but genuine usefulness vs. a pixel-perfect product that solves no real problem. Wabi-sabi says: ship it imperfect, and let the patina of real use give it beauty. |
How to Use This Alongside the Mechanical Engineering Models
The mechanical engineering mental models answer: how does this system behave? How will it break? What are the physics?
The industrial design mental models answer: how will people interact with this system? What will they perceive, feel, and do? How do we make it work for humans, not just in theory?
Together, they form a complete toolkit: engineering tells you what’s possible, design tells you what’s desirable, and the intersection — what is both possible and desirable — is where great products, teams, and systems live.