IrrationalLabs Profile picture
Behavioral design for better products and a better world. Join the community: https://t.co/saBzFbv0z6
Sep 7, 2022 7 tweets 3 min read
Hop in, we're designing for human behavior! 🚗 We decreased auto loan defaults by 69% and doubled automatic payments. Here's how.

You go to a car dealership on your day off and find your new ride🏎️ It's not the Mustang that teenage you dreamed of, but hey, in this economy? 🧵 You get financing and a call from a bank. Then the problems start.

💲Your loan is due each month on the day that you bought the car, even though it’s 8 days from your payday. They’ll send you a reminder, but you don’t read your mail very often.

Could this be easier? 🤔
May 3, 2022 9 tweets 5 min read
Ever feel like you want to throw your phone across the room? No? Just us? 😳

If you’re like us (constantly fighting distraction), this might explain why. 🧵 People receive an average of 60-80 daily notifications on their phones.

While notifications have become a standard part of our lives, a constant barrage of information is also linked to ⬆️ anxiety, depression, and attention disorders.
Dec 9, 2021 20 tweets 5 min read
Knowing these top 16 critical behavioral ideas could change your work—and your life. 🚨

We cover all of these in our bootcamps. And we wrote an essential primer. Read on! (THREAD 👇) There are a lot of behavioral hacks, but we think these 16 are essential. They cover a few major categories:

+ Perceived Costs and Benefits
+ Attention and Effort
+ Cognitive Heuristics
+ Risk and Uncertainty
+ Choice Architecture
+ Norms and Influence

Learn all 16 below 👇
Dec 7, 2021 10 tweets 3 min read
How do you know if someone will pay you back?

Behavioral Scientist Allison White summarizes a recent paper from Wang, Drabeck and Wang that tries to discern the best predictor of loan repayment using "hard" and "soft" information about borrowers. (THREAD) "Hard Information" describes financial metrics like assets, mortgages, and income.

"Soft Information" includes demographic characteristics like age, education, gender + race, as well as social networks, video interviews, profile pictures, and prior borrowing stories. /2
Feb 3, 2021 7 tweets 2 min read
We designed an intervention that reduced shares of flagged content on TikTok by 24% via a large scale RCT, thread 👇1/7 See article: techcrunch.com/2021/02/03/tik… 2/7