How Rich Barton Built Expedia and Zillow — Audacious Goals, Provocation Marketing, and More
Executive Summary
Rich Barton discusses creating consumer products that people talk about, choosing a memorable name, learning from a failed Microsoft project and coaching company founders. Selected product, marketing and organization passages only; health, politics and unrelated personal material omitted. Company histories and marketing judgments are his account, not guarantees. Metadata uses the linked YouTube version; sections are untimed. Short quotes credited to The Tim Ferriss Show (https://tim.blog/podcast); official transcript: https://tim.blog/2025/04/19/rich-barton-transcript/
Key Takeaways
Give people a useful reason to talk
Barton calls his approach provocation marketing: a feature matters enough that people have strong reactions and discuss it. He connects that reaction to a useful consumer product, not outrage for its own sake. Ask what people gain from the feature before treating attention as proof of value.
Use changing data as repeatable material
He describes Zillow's marketing team distributing fresh data that consumers and the press care about, with Glassdoor as another example. A changing information stream can support repeated interest. The editorial lesson is to make useful data understandable rather than expect a launch announcement to carry every conversation.
Make the name distinctive
His naming rules include rare letters, a small number of syllables and positive associations. These are his practical preferences, not scientific guarantees of brand success. Test whether people can hear, remember and repeat the name before assuming an unusual spelling makes it effective.
Understand the trade-off of an invented word
Barton says a literal category name is easy to explain while an invented word takes more effort to define. He likes the latter because the company can build a distinctive meaning around it. The benefit comes with a communication job; a clever name does not explain itself.
Learn from a big swing without ignoring patterns
He recalls an unsuccessful software-and-book bundle at Microsoft and a manager who encouraged his next idea. He says organizations should support innovation but also notice repeated failures. Separate the experiment's lesson from the person's worth while still reviewing whether the same problem is happening again.
Recognize the data collection problem
In his Glassdoor account the team needed enough trustworthy user-submitted salary information to make the product useful. He describes a give-to-get approach after early surveys. A data product needs a reason for people to contribute as well as a reason for others to read it.
Choose engaged company-building advisers
Barton describes private-company boards as advisers and coaches who help with strategy, fundraising and recruiting, rather than run the company. He distinguishes public-company duties. Choose relevant experience and engagement, not only a recognizable name; his account is not governance or legal advice.
The Conversation Breakdown
Give people a useful reason to talk
⏱ Untimed topicBarton calls his approach provocation marketing: a feature matters enough that people have strong reactions and discuss it. He connects that reaction to a useful consumer product, not outrage for its own sake. Ask what people gain from the feature before treating attention as proof of value.
Use changing data as repeatable material
⏱ Untimed topicHe describes Zillow's marketing team distributing fresh data that consumers and the press care about, with Glassdoor as another example. A changing information stream can support repeated interest. The editorial lesson is to make useful data understandable rather than expect a launch announcement to carry every conversation.
Make the name distinctive
⏱ Untimed topicHis naming rules include rare letters, a small number of syllables and positive associations. These are his practical preferences, not scientific guarantees of brand success. Test whether people can hear, remember and repeat the name before assuming an unusual spelling makes it effective.
Understand the trade-off of an invented word
⏱ Untimed topicBarton says a literal category name is easy to explain while an invented word takes more effort to define. He likes the latter because the company can build a distinctive meaning around it. The benefit comes with a communication job; a clever name does not explain itself.
Learn from a big swing without ignoring patterns
⏱ Untimed topicHe recalls an unsuccessful software-and-book bundle at Microsoft and a manager who encouraged his next idea. He says organizations should support innovation but also notice repeated failures. Separate the experiment's lesson from the person's worth while still reviewing whether the same problem is happening again.
Recognize the data collection problem
⏱ Untimed topicIn his Glassdoor account the team needed enough trustworthy user-submitted salary information to make the product useful. He describes a give-to-get approach after early surveys. A data product needs a reason for people to contribute as well as a reason for others to read it.
Choose engaged company-building advisers
⏱ Untimed topicBarton describes private-company boards as advisers and coaches who help with strategy, fundraising and recruiting, rather than run the company. He distinguishes public-company duties. Choose relevant experience and engagement, not only a recognizable name; his account is not governance or legal advice.
Core Frameworks & Ideas
Provocation marketing - guest label
A useful feature can prompt strong reactions and unprompted conversation; attention alone does not prove value.
Naming rules - discussed approach
Compare distinctiveness, syllables and positive associations while testing whether people can repeat the result.
Give to get - discussed approach
Create a contribution incentive for a data product instead of assuming useful information will arrive on its own.
Actionable Next Steps
- ✓Write the useful consumer benefit of your most talked-about feature.
- ✓Test candidate names aloud with someone unfamiliar with them.
- ✓Identify one changing data point your audience wants to understand.
- ✓Review an unsuccessful experiment for a lesson and a repeatable problem.
Notable Quotes
“pick the super rare letters, and pick them because they’re very distinctive.”
“So if the word could be a good dog name, you’re onto something, like you can call for it.”
“When you have a really provocative feature that you know people are going to feel emotional about one way or the other and they’re going to talk about it, you’re on to something.”
The One Thing to Remember
Build a useful feature people want to discuss, and make the organization able to learn from bold attempts.