        <div>
        .className
        </div>
        padding:
        onClick()
    

NATIONAL GEOGRAPHIC LEARNING

💻

Unit 4

Website Design
on a Giant Scale

Data, Intuition & Billions of Users

Lead-in 01

Have you ever been annoyed by a website or app? What bothered you? 😤

Facebook, YouTube, Google — any tiny design change affects billions of people. How do you design for the whole world?

📊

Data

Tells you what, not why

💡

Intuition

Fills the gap

🧪

Testing

Proves your idea

🔄

Change Carefully

Users resist change

Meet the designer who built for Facebook, YouTube, and Google — and learn what she discovered.

Reading 02

Skimming Task ⏱️

Read the article quickly (90 seconds). Answer three questions:

👤

WHO?

Who is Margaret Gould Stewart, and what has she designed?

📖

WHAT TWO LESSONS?

What are the two main lessons Stewart teaches?

⭐

HOW?

How did YouTube change its five-star rating — and why?

✅ WHO: A designer who worked at Google, YouTube, and Facebook | TWO LESSONS: (1) Look Beyond Data; (2) Introduce Change Carefully | HOW: Most users only used 1 or 5 stars, so they simplified to thumbs up/down to match real behavior

Reading 03

Intro — Stewart's Background

Margaret Gould Stewart has designed for some of the giants of the Internet, including Google, YouTube, and Facebook. Here are two lessons she has learned from her experience designing for Internet users.

'Giants' is a metaphor borrowing the image of colossal creatures to convey size, power, and dominance. Unlike 'big websites', 'giants' implies not just scale but overwhelming influence on the world. It also carries connotations of something slightly intimidating. The metaphor frames the design challenge as epic: even experienced designers like Stewart face uniquely enormous pressures working at this level.

Reading 04

Stewart's Two Lessons

'Has learned' is present perfect, indicating that the learning happened in the past and the result is still relevant now. The lessons are not historical curiosities — they are currently valid insights. This tense choice positions Stewart's experience as ongoing wisdom, not outdated advice. It also adds credibility: the learning came from real sustained experience over time.

Reading 05

The Photo Reporting Problem

For a long time, Facebook had a tool that allowed people to report photos as spam or abuse. But of the cases reported, only a small percentage of the photos were actually offensive. One of the designers on the team felt there probably was a reason for this, so he studied the cases carefully. He found that in most cases users just didn't like the photos of themselves their friends had posted, and wanted them taken down. To enable people to report cases like these, the Facebook team added a new feature. This feature allowed people to message their friends to ask them to take the photo down. But only 20 percent of people used the function.

'For a long time' establishes a problem state that existed for a significant period — implying it was unchallenged. The simple past 'had' signals that this situation is now over — it's a narrative setup that promises change. Together, the phrase creates a classic story arc: flawed system existed for a while → someone noticed → things changed. The reader anticipates a turning point.

Reading 06

Only Small % Were Offensive

'Actually' is an emphasis adverb signaling that the reality contradicts what one might expect. If photos were reported as spam/abuse, one would assume they were genuinely offensive — but 'actually' reveals the gap between user behavior and user intention. This is the central insight of the section: data (reports) looked like one thing, but the reality (emotion, social awkwardness) was something quite different. 'Actually' is the linguistic marker of a discovered mismatch.

Reading 07

A Designer's Curiosity

'Felt' (vs 'knew') expresses a hunch or intuition, not confirmed knowledge — which is appropriate at this stage of investigation. 'Probably' is a probability adverb that hedges the certainty. Together they capture the designer's mindset: acting on curiosity, not certainty. This is the point: good design requires following intuitions even without proof. The language models intellectual humility and open inquiry — a key theme of the text.

Reading 08

The Real Reason — Embarrassment

'Just' here is a downplayer/minimizer — it reduces the emotional gravity to something relatable and ordinary. The author uses it to humanize the issue: users weren't filing legal complaints, they were experiencing everyday social discomfort. 'Didn't like' is deliberately informal and universal — everyone has felt this. The informal register invites the reader to empathize with users rather than judge them.

Reading 09

First Fix — Message Feature

'To enable...' is a to-infinitive phrase expressing purpose (purpose adverbial). It answers the question "why did the team add a new feature?" Purpose infinitives are extremely common in instructional and expository writing because they make cause-and-effect explicit. They are more concise than 'because they wanted to enable...' or 'in order that people could report...' — signaling the writer's economy with language.

Reading 10

Message a Friend

The chain — allowed → to message → to ask — mirrors the multi-step social interaction: the feature enables → user initiates contact → user makes a request. Each infinitive represents a dependent step. This nested infinitive structure is economical but also captures how one enablement leads to another. The parallel between grammar and social process is rare and insightful — the form enacts the content.

Reading 11

Only 20% Used It

The short sentence creates a dramatic pause — a contrast against the longer explanatory sentences before it. The brevity signals failure. The word 'only' is a minimizer that underscores disappointment: 20% is not enough. Structurally, this sentence sets up the next paragraph's deeper investigation. Short sentences after long ones create narrative rhythm — the punch of the short lands harder because of what came before.

Reading 12

Communicating Emotions — The Real Need

The team worked on the case further — it spoke to communications experts and studied rules of polite language. It discovered that users didn't just want to tell their friends to take the photo down — they wanted to tell their friends how the photo made them feel. So the team made a small change. People could select a message to explain why they didn't like it, such as, "It's embarrassing." This small change had a huge impact — 60 percent of people who reported photos used the function. Surveys showed that people on both sides of the conversation felt better as a result.

A tech company consulting communications experts (not engineers) is surprising — it suggests the problem is fundamentally human and social, not technical. This is the lesson: UX design is not just about code or data, but about understanding human emotions and social norms. The move challenges assumptions that tech problems have tech solutions. It positions humanities expertise as essential to great design.

Reading 13

What Users Really Wanted

The not…but structure creates a sharp correction: the first 'want' is replaced by a deeper one. The contrast reveals that users aren't primarily practical (take it down) but emotional (express how I feel). This is an insight about human psychology: people need to be heard, not just helped. The structure forces the reader to revise their assumption — a rhetorical strategy of expectation + correction.

Reading 14

Small Change Made

The short sentence acts as a pivot point in the narrative — a moment of suspense before the reveal. It also uses 'so' as a consequence connector, making cause-effect clear: discovery → action. 'Small' is strategic: it sets up the contrast with "huge impact" in the next sentence. By minimizing the change first, the author amplifies the surprise of the outsized result. This is classic rhetorical understatement.

Reading 15

Small Change, Huge Impact

The direct-speech example grounds the abstract concept in a specific, relatable moment. Readers immediately picture a real social situation — an unflattering photo, an awkward conversation. This is the power of exemplification in writing: abstract ideas (expressing emotional context) become concrete through specific instances. The quote-within-example also creates a shift to first person, briefly putting the reader inside the user's perspective.

Reading 16

60% Used the New Feature

The sentence uses antithesis — 'small' vs. 'huge' — to dramatize the result. This contrast is rhetorically effective because it defies expectation: a minimal change should produce a minimal result. The 60 percent statistic (compared to the original 20 percent) provides the concrete evidence that backs up the claim. Together, the rhetorical contrast and the data create a compelling argument for the power of empathy-driven, emotion-aware design.

Reading 17

Both Sides Felt Better

Emphasizing 'both sides' demonstrates that the new design created a win-win outcome: the person upset about the photo felt heard; the friend receiving the message felt respected. This is a core principle of ethical design: solutions that only benefit one party are inferior. The author uses this detail to argue that empathy-driven design — not just data — creates mutually beneficial outcomes. It's also more persuasive evidence than a statistic alone.

Reading 18

Data Can Help, Not Lead

While data about how people are using a product can help designers make decisions, it isn't always as simple as following the numbers. Other factors such as intuition, research, and testing of design are equally important. As Stewart points out, "Data can help you make a good design great, but it will never make a bad design good."

'While' here is a concessive conjunction — it acknowledges the value of data before introducing a limitation. This structure (while A is true, B is also true) is very common in academic and analytical writing because it shows nuanced, balanced thinking. It avoids the extremes of "data is useless" or "data is everything." The author uses it to introduce the main lesson: data is one tool among many, not the final authority.

Reading 19

Intuition and Research Matter Too

'Equally' makes a strong comparative claim — it doesn't say intuition is somewhat useful, but that it is precisely as important as data. This challenges the tech industry's typical emphasis on data-driven decisions. The three-item list — intuition, research, testing — uses tricolon structure, a persuasive pattern that feels complete and authoritative. Listing them as parallel gives each equal rhetorical weight. The sentence is a direct rebuttal to data-only thinking.

Reading 20

Stewart's Famous Quote

The quote uses antithesis: good design → great (positive progression) vs. bad design → still bad (data can't rescue it). The parallel structure — "make a ___ design ___" — makes the contrast crystal clear. This is an aphorism: a short, memorable statement of principle. The use of 'never' (absolute negation) strengthens the second half. The quote is memorable precisely because it is structurally elegant — form mirrors function: the design of the sentence demonstrates good design.

Reading 21

YouTube's Rating Problem

At one time, YouTube was looking for ways to encourage more people to rate videos. When Stewart and her team looked into the data, they found that most people were only using either the highest rating (five stars) or the lowest rating (one star). Almost no one was using two, three, or four stars. So the team decided to simplify the rating — it gave users a choice between good or bad: thumbs up or thumbs down.

The past continuous suggests ongoing, active searching — a process that was in progress. It creates a sense of background activity before the discovery. 'Encourage' is significant: users were not being forced, but nudged. This sets up the tension: even a well-designed 'encouragement' system can fail if it doesn't match actual user behavior. The continuous aspect frames the team as engaged, curious, and problem-solving — not reactive.

Reading 22

Extreme Ratings Only

Extreme ratings reveal emotional engagement bias: online, people are most motivated to react when they feel strongly. Moderate feelings ('it was okay') rarely motivate action. This is sometimes called the 'J-shaped distribution' of reviews. Psychologically, it suggests that rating systems designed for nuance may not match human motivation patterns online. The data told a clear behavioral story — but the design team had to interpret what behavior meant, not just record it.

Reading 23

Middle Ratings Unused

'Almost no one' is a near-universal quantifier — it stops just short of saying 'nobody' but implies near-total absence of middle-range use. This is important data: if two, three, and four stars were used by even a significant minority, removing them would be controversial. But when 'almost no one' uses them, removing those options loses virtually nothing while simplifying the system. The sentence provides the statistical justification for the redesign that follows.

Reading 24

Simplify: Thumbs Up or Down

The colon introduces an illustration — it says "here is what that choice looks like." The design decision directly mirrors the insight: since users were already choosing between extremes, make the system reflect that binary reality. This is elegant design thinking: redesign to match behavior, not to force new behavior. The colon grammatically connects principle to implementation — a mini-demonstration of the lesson itself.

Reading 25

Preparing Users for Change

YouTube tried to prepare people for this change by sharing data about how the five-star rating system wasn't being used as intended. It announced that it was going to change the system to match user behavior. When the change was made, it was still frustrating for some users as they had become attached to the old design. However, because of the preparatory steps taken earlier, it was easier for YouTube to get users to accept the change. This experience shows that even when huge websites try to manage change carefully, it's impossible to completely avoid negative responses. Any changes — even small improvements — need to be introduced carefully.

YouTube uses transparency as a change management strategy — showing users the data that motivated the decision. By making the evidence public, they invited users to understand (not just accept) the change. This is effective because it respects users' intelligence and makes them feel included in the process. However, as the next sentence shows, even transparency doesn't eliminate resistance — habit is powerful. The strategy reduces but cannot eliminate friction.

Reading 26

Announcing the Change

'Going to' signals a planned, definite future action — not a possibility, but a decision already made. The announcement is strategically important: by making the change public in advance, YouTube gives users time to mentally prepare. This is a key principle of change management — surprises generate more resistance than anticipated changes. The choice of 'match user behavior' is also clever framing: it positions the change as serving users, not the company.

Reading 27

Attachment to Old Design

Past perfect ('had become') indicates that the attachment formed before the change was made — it was already established, deeply rooted by the time of the change. This is crucial: the problem wasn't that users disliked the new design rationally; they were emotionally connected to the old one. The tense choice highlights the temporal gap between habit formation and design change — and why that gap makes change difficult.

Reading 28

Preparation Made the Difference

'Because of the preparatory steps' is a causal prepositional phrase — it directly links prior action to a better outcome. The use of 'easier' (comparative adjective) is honest: it doesn't claim the change was painless, only that it was less painful. This is a nuanced concession that strengthens the argument: preparation cannot guarantee smooth acceptance, but it measurably reduces friction. The sentence validates YouTube's transparency strategy with a positive, evidence-based result.

Reading 29

Change Always Brings Some Negative Responses

'Completely' is a degree adverb that subtly softens the claim: the author isn't saying negative responses can never be reduced — only that they can never be entirely eliminated. This is precise, honest argumentation. It acknowledges that preparation helps (as the previous sentence shows) while insisting there is always some residual resistance. The use of 'impossible' + 'completely' together creates a nuanced absolute: the outcome is unavoidable in some measure, however well you manage the process.

Reading 30

Final Lesson

The dashes enclose a parenthetical interrupter that intensifies the claim. 'Even small improvements' is a scalar argument: if even minor positive changes can cause disruption, then change in general demands care. This challenges the naive assumption that users always welcome improvements. The dashes make the parenthetical feel like a real-time thought — the author catching themselves adding important nuance. The sentence ends the article with a universalizable rule: always manage change carefully.

Language 31

Gerund as Subject

When an action becomes the topic — gerunds lead the sentence

Designing Testing Following -ing as subject

A) "Designing for billions requires both data and empathy." — gerund phrase as subject
B) "Testing the interface with real users revealed unexpected patterns." — same structure, different context
C) ❌ "To design for billions requires..." — grammatically ok, but less natural in academic prose
D) RULE: Gerund subjects emphasize the ongoing process over the agent or result — academic writing prefers this for describing activities and methods

Why it matters: Gerund subjects are extremely common in academic and technical writing. They shift focus from "who does it" to "the activity itself." When Stewart says things like "testing of design," the gerund form positions design-testing as an established practice worth studying in its own right — not just what one person does. Practice: Rewrite "We need to test user behavior" using a gerund subject.

Language 32

Concessive "Even When"

Acknowledging the best case — then showing it's still not enough

even when even if even though concessive clause

A) "Even when huge websites try to manage change carefully, it's impossible to avoid negative responses." — text example
B) "Even if you prepare users, some resistance is inevitable." — hypothetical concession
C) ❌ "When huge websites try, they can avoid negative responses." — no concession, different meaning
D) RULE: even when/if = scalar concession — "I grant you the best possible scenario, and my point still stands"

Contrast: 'Although' is neutral concession. 'Even when/if' is stronger — it pushes the conceded scenario to an extreme to make the conclusion more powerful. Use 'even when' when you want to argue that your point holds under the most challenging conditions. This is a rhetorical escalator — the harder you make the concession, the more convincing your conclusion.

Language 33

Reporting Verbs — Degrees of Certainty

How we introduce other people's ideas in academic writing

points out found showed felt

A) "As Stewart points out..." — present tense = still valid, authoritative opinion
B) "He found that in most cases users just didn't like..." — past tense = specific research finding
C) "He felt there probably was a reason..." — verb of mental state = subjective interpretation, not proven fact
D) RULE: says/points out/argues = opinion still held; found/showed = evidence discovered; felt/thought = subjective, hedged

Academic writing tip: Choosing the right reporting verb controls how readers receive information. 'Points out' implies the speaker believes the claim is objectively true. 'Argues' implies it's a contestable position. 'Found' implies empirical evidence. Always match your reporting verb to the epistemological status of the claim — how certain is this knowledge?

💻

LESSON COMPLETE

📊

Data Alone Isn't Enough

Numbers show what, not why

💡

Intuition + Research

Understanding emotions matters

🔄

Manage Change Carefully

Habit makes change painful

🧑‍🤝‍🧑

Both Sides Win

Empathy-driven design works

"Data can help you make a good design great,
but it will never make a bad design good."

— Margaret Gould Stewart

Website Designon a Giant Scale