Feeling Lucky? The $100 Million Button Google Refused to Remove

Most product teams remove anything that appears to reduce efficiency or revenue. Google kept one of those things on its homepage for years, quite intentionally.

That decision matters because it reveals a discipline most companies barely understand. They know how to optimize a feature. They know how to defend a metric. They know how to A/B test a button until every soft edge gets sanded off. What they often do not know is when a product should feel slightly inefficient on purpose.

Google’s homepage became iconic partly because it respected cognitive ease. It did not assault users with portals, promos, widgets, and clutter. Then right beside the utilitarian Search button, it kept something mischievous: I’m Feeling Lucky. That is a weird move if your only question is revenue per click. It is a smart move if your real question is what the product should feel like when a human touches it.

That is the broader lesson. A company can protect a strange little feature not because it wins the local spreadsheet, but because it strengthens the global relationship. When teams obsess over the nearest metric instead of the top business metric, they often delete the exact moments that make a product memorable.

And once those moments are gone, they are surprisingly hard to reconstruct. You can always add another conversion tweak later. It is much harder to rebuild a feeling of distinctiveness after years of sanding the product down into pure efficiency. That is why these decisions deserve more seriousness than they usually get inside growth meetings.

⚡ Speed Run Notes

• I’m Feeling Lucky worked because it turned search from a pure utility action into a tiny emotional gamble.
• The feature was not strong because it was random. It was strong because the downside stayed low while the curiosity stayed alive.
• Google Doodles reinforced the same behavioral signature at the brand level: this homepage might still surprise you.
• The hidden win was not direct monetization. It was affection, trust, distinctiveness, and repeat visitation.

Table of Contents

• About Yu-kai Chou
• The Button That Seemed to Cost Google Money
• The Behavioral Design Behind I’m Feeling Lucky
• Google Doodles and Brand-Level Curiosity
• What This Reveals About Core Drive 7
• Modern Lessons for Product Teams
• Frequently Asked Questions
• Related Reading

About Yu-kai Chou

Yu-kai Chou is an S-Tier Behavioral Designer, keynote speaker, and creator of the Octalysis Framework. His work has influenced products and communities reaching over 1.5 billion users through clients, audiences, and research citations that include LEGO, Microsoft, Porsche, Tesla, Stanford, Harvard, Google, and more than 3,700 academic publications.

Verify: Wikipedia · Google Scholar · Wikidata · LinkedIn

The Button That Seemed to Cost Google Money

Google’s homepage became iconic partly because of its restraint. While most portals in that era kept shoving more content onto the screen, Google made the opposite bet. It trusted that a cleaner interface would feel faster, calmer, and more intelligent.

Then it did something even more interesting. It kept a second button that looked, at least on paper, economically irrational. I’m Feeling Lucky sent users straight to a likely destination instead of making them dwell on the search results page where the ads lived.

In a narrow meeting, that button is easy to attack. It appears to waste monetizable attention. In a better meeting, the button looks like brand theater in the best sense. It signals that the company is confident enough to occasionally skip the monetization ritual. That kind of confidence quietly creates status, because users sense when a product behaves like it is leading rather than pleading.

It also adds a touch of productive tension. The best products are not always the ones that remove every bit of resistance. Sometimes a little bit of chosen uncertainty creates meaning. That is the same logic behind meaningful friction. The point is not to annoy the user. The point is to make an interaction feel like a tiny experience instead of a sterile transaction.

That is why the button mattered. It made search feel less mechanical and more alive. The click carried a tone, not just a function.

Most executives are not trained to protect tones. They are trained to protect funnels. That is why features like this are so vulnerable. In a spreadsheet review, a whimsical button looks soft, sentimental, and expensive. In real user psychology, it can become one of the clearest signals that the company still understands delight.

That signal compounds. Users rarely sit down and announce, “I trust this product more because it preserved one odd little flourish.” They simply absorb the feeling. Over time, enough of those feelings become brand identity. And once identity is established, the product gets credit for qualities that are much harder for competitors to clone with raw feature parity alone.

The Behavioral Design Behind I’m Feeling Lucky

The wording does a lot of the work. A label like Top Result or Direct Search would frame the feature as a mere shortcut. I’m Feeling Lucky frames it as a playful wager. The user is not only trying to find an answer. The user is also testing what kind of outcome Google thinks is good enough to surprise them with.

That is where the modern reading of Core Drive 7: Unpredictability, Curiosity, and Mystery becomes useful. Good curiosity design is not chaos. It is a controlled question mark. There is enough uncertainty to feel interesting, but not so much uncertainty that the user feels punished for clicking.

The feature also behaves like a very gentle Hook Model. There is a cue, a low-effort action, a variable reward, and a subtle reason to come back. The loop is soft, but that softness is precisely why it works. Google did not need to squeeze the user. It only needed to create the sense that one click might lead somewhere unexpectedly delightful.

There is also a small ownership dynamic here. The user chooses the gamble. That chosen gamble creates a touch of ownership and possession, because the outcome feels partly authored by the user’s decision rather than imposed by the interface. That detail sounds minor, but it changes the emotional texture of the moment.

Most companies miss this because they talk as if every user is motivated by the same thing. They are not. Some people want certainty. Some want speed. Some want surprise. Some want to feel clever for discovering something. That is why product teams need a richer language for different types of motivation instead of pretending every click is purely utilitarian.

That distinction matters in practical design reviews. If a team only asks whether a feature reduces friction, it will misread features that are trying to create wonder, suspense, or delight. Those experiences are not defects just because they are harder to describe in operational language. They are often the reason the product becomes memorable in the first place.

Another overlooked point is that curiosity does not need to be loud. Some of the most effective curiosity mechanics are tiny. A label. A hover state. A strange invitation. A gentle shortcut. When a designer understands that, they stop assuming that every engaging mechanic must be dramatic. Sometimes the smartest emotional design move is a whisper.

Google Doodles and Brand-Level Curiosity

The same instinct appeared in Google Doodles. Doodles turned the homepage into a place where something unexpected might show up even if the visitor arrived with a purely functional goal. Sometimes the surprise was visual. Sometimes it was playful. Sometimes it became a miniature interactive experience.

That matters because it shows curiosity operating at the brand level, not just at the button level. Google was teaching users that the homepage was not merely a utility surface. It was also a place that occasionally winked back.

When surprise is used this way, it strengthens intrinsic retention. Users do not only return because the product solves a task. They also return because the product has built an emotional expectation around discovery. That is a deeper form of loyalty than simple habit automation because it lives in anticipation, not only routine.

Notice what Google did not do. It did not flood the interface with novelty. It did not make every visit unstable. It rationed surprise carefully. That restraint is what kept the delight clean instead of exhausting. A curious brand knows when to surprise and when to disappear.

This is where many product teams get unpredictability wrong. They think more novelty automatically means more engagement, so they stack popups, streak gimmicks, random rewards, and forced discovery into the same surface. The result is not delight. The result is noise. Curiosity only feels elegant when it is supported by restraint.

Google, at its best, understood pacing. The homepage did not beg. It invited. That invitation was powerful because the default experience was already useful. Surprise becomes much more persuasive when it grows on top of competence instead of trying to compensate for incompetence.

That pacing lesson is applicable far beyond search. Education products, shopping flows, productivity tools, and SaaS dashboards all face the same temptation to become relentlessly instrumental. The teams that resist that temptation intelligently are often the ones that feel most alive. They leave just enough room for the product to express a personality instead of behaving like a form field with a logo.

What This Reveals About Core Drive 7

Good Core Drive 7 design is not about keeping people off balance. It is about offering structured uncertainty with low downside and meaningful intrigue.

• The stakes are light enough that trying feels safe.
• The framing is playful instead of stressful.
• The reveal creates delight, insight, or exploration.
• The unpredictability reinforces the product instead of distracting from it.

That last point is critical. If the surprise feels bolted on, users experience it as gimmickry. If the surprise feels native to the product’s identity, users experience it as charm.

There is also an underappreciated White Hat layer here. Unpredictability becomes more powerful when it serves something larger than compulsion. The moment the company starts using uncertainty only to squeeze more attention, the relationship gets ugly fast. If you are trying to design Epic Meaning into a seemingly meaningless product, then the surprise has to support a value the user can respect, not merely an urge the company can exploit.

Google mostly stayed on the right side of that line here. The button felt generous, not extractive. It suggested, “we think this jump might actually be good for you.” That is a very different emotional posture from the dark patterns many modern products use. It feels like hospitality instead of extraction.

Generosity is easy to overlook because it rarely announces itself with a dashboard spike. But users can feel the difference between a product that is nudging them toward a good surprise and a product that is cornering them into another monetizable behavior. That feeling changes the moral temperature of the brand.

And when the stakes do rise, the design discipline changes. High uncertainty with high downside is a very different emotional cocktail than low uncertainty with low downside. The same motivational tool can feel delightful in one context and punishing in another, which is why risk calibration matters so much whenever a team borrows game mechanics for a serious product.

Modern Lessons for Product Teams

First, do not evaluate every feature only by direct monetization. Some features create value indirectly by making the product feel more confident, playful, and worth returning to. Those indirect effects are often the ones that quietly improve everything else.

Second, understand that surprise is not a substitute for substance. A weak product cannot hide behind whimsy forever. But a strong product can absolutely use surprise to deepen affection.

Third, protect the odd little choices that signal identity. In many organizations, those are the first things to get cut because they are the hardest to defend in finance language. That does not mean they are weak. It often means they are doing invisible work.

Invisible work is still work. It may be soft in attribution but very hard in consequence. Once a brand starts feeling interchangeable, every future acquisition, retention, and pricing conversation becomes more expensive because the product has lost some of the emotional leverage it once had for free.

Fourth, design teams should ask a more strategic question: what is our equivalent of a lucky button? In other words, where do we deliberately refuse to become purely extractive? That question can reveal a lot about whether the brand is building trust or merely harvesting attention.

Finally, remember that a product can look less efficient and still be more effective in the long run. The teams that understand this are often the ones that keep their magic while everyone else keeps sanding theirs off.

I think this is one of the biggest strategic differences between product teams that become beloved and product teams that become replaceable. Replaceable teams optimize what is visible on the dashboard. Beloved teams also protect the emotional texture the dashboard cannot see yet. They understand that delight, trust, and identity are not fuzzy extras. They are compounding assets.

If you are leading a product team, it is worth asking which small choices on your surface are doing invisible brand work right now. Which ones make the experience feel more human? Which ones give users a reason to smile, trust, or linger? And which ones are at risk of getting deleted because nobody has a clean ROI slide for them? Those questions usually reveal more strategic truth than another week of local button optimization.

Frequently Asked Questions

Why did Google keep the I’m Feeling Lucky button?

Because the feature did more than route users. It made Google feel playful and confident while giving people a low-stakes curiosity-driven shortcut. That emotional framing mattered as much as the click path itself.

How does this connect to Octalysis?

It is a clear example of Core Drive 7 at work. The feature creates a small moment of uncertainty that invites exploration without making the user anxious. That balance is the whole art here.

Why do Google Doodles matter in this analysis?

They show the same design philosophy expressed at the brand level. Users learned that Google’s homepage might occasionally surprise them in an enjoyable way. That expectation became part of the brand over time.

Is unpredictability always good in product design?

No. It works best when the stakes stay low, the surprise is welcome, and the uncertainty reinforces delight instead of frustration. Otherwise it turns manipulative very quickly, and users notice.

What is the practical lesson for product teams?

Do not judge every feature by immediate monetization alone. Some features strengthen loyalty by shaping how the product feels, and that emotional layer can be more durable than the local metric you were tempted to optimize.

Related Reading

• The Octalysis Framework: Complete Gamification Framework
• How the 8 Core Drives Interact
• Level II and III Octalysis: The Advanced Framework
• Why Dopamine Addicts Make Terrible Life Choices
• Core Drive 7: Unpredictability & Curiosity

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