Rennes, France, a Tuesday in November, 9:40 p.m. Samuel, head of e-commerce at a home-linen retailer, is watching the same video for the third time. It is an anonymized recording of a customer session. A shopper has filled her cart: a duvet cover, two pillowcases, a throw. 286 euros. She clicks “Order.” The screen asks for her email address and password.
She types a first password. Rejected. A second one. Rejected. She clicks “Forgot password,” waits about forty seconds, switches to her email tab, comes back to the site. The cursor hovers over the “Create an account” button. Then the tab closes.
Samuel calls his CRM manager.
“How many carts do we lose on the login screen?”
“I don’t know exactly. We don’t measure that screen.”
“What if we dropped mandatory account creation?”
“Marketing will never agree. No account, no customer data.”
Samuel hangs up. He has just put his finger on a friction nobody in the company sees, because nobody measures it. And he wonders whether he is even allowed to touch a screen that several departments consider their own.
Samuel’s story is fictional. The one that follows is real, and it reportedly brought in 300 million dollars.
A form with two fields and two buttons
January 2009. Jared Spool, founder of the consulting firm User Interface Engineering, publishes a short article with a cheeky title: “The $300 Million Button” (Center Centre). He describes a project for a large online retailer he never names.
The form in question is utterly ordinary. Two fields: email address and password. Two buttons: “Login” and “Register.” A “Forgot Password” link. It appears right after shoppers click “Checkout,” before they enter payment details (Center Centre).
The designers thought they were helping. A registered customer buys faster next time, since addresses and payment details are already stored. On paper, the logic holds.
Spool’s team runs usability tests. Participants get money and a shopping list, with one simple instruction: buy. First-time shoppers reject registration. One of them sums up the discomfort: “I’m not here to enter into a relationship. I just want to buy something.” (Center Centre)
Returning customers can’t remember whether they have an account. They try combinations, fail, get annoyed. Many suspect registration is a way to flood them with marketing. The irony: the registration form asked only for name, shipping and billing addresses, and payment information. Exactly what was needed to deliver the order anyway (Center Centre).
What the database had been saying for years
What happened in the test lab could have been brushed aside. A few grumpy participants, a sample too small. So the team went to the data.
45% of customers had multiple registrations, some as many as ten. About 160,000 people requested a password reset every day. And 75% of those who requested one never completed a purchase (Center Centre).
Read those numbers again. The data had been sitting in the company’s servers since day one. Nobody had looked at it from that angle, because nobody was asking. The form had become part of the furniture.
This is what I call invisible friction. I describe friction as one of the two levers of any habit change: you increase desire, you remove friction. A buying habit follows the same rules. Every field to fill in is a toll the customer pays with attention. When the toll exceeds the desire, the customer turns around (My book, chapter 5).
A button, and above all a sentence
The fix was small. The designers removed the “Register” button. In its place, a “Continue” button, with a message: you don’t need an account to buy, and you can create one during checkout to make future purchases faster (Center Centre).
Everyone remembers the button. Everyone forgets the sentence. Yet the sentence does most of the work. It answers the customer’s fear before it is voiced, and it keeps the account by moving it after the purchase. Registration goes from entry barrier to option, offered once the customer has what they came for.
The reported result: the number of customers making a purchase jumped 45%, with an extra 15 million dollars in the first month and 300 million in the first year (TheDomains). Spool says he found a voicemail from the CEO of this 25-billion-dollar retailer the week the new sales figures came in (CMSWire).
Why that number should make you suspicious
I have said it often: assumptions are dangerous, and beautiful stories even more so. This one ticks every box of the perfect corporate fairy tale. A tiny fix, a big round number.
Look at it critically. The retailer is never named. The figures come from the consultant who ran the project, and seventeen years later nobody can cross-check them. Nothing says whether there was a control group, or how much of the increase came from the season, promotions or the growth of online retail. Online, the story has even mutated: some versions describe an A/B test or a third “guest” button added next to the other two, which the original article does not say (Atticus Li).
Should you throw the case away? No. Put it back where it belongs: solid on the mechanism, shaky on the amount. And the mechanism has since been measured by others. The Baymard Institute, which studies checkout usability, surveyed 1,026 US adults who had shopped online in the previous three months: 18% said they had abandoned an order because they did not want to create an account (Baymard). Other research from the same institute, reported by Amazon Pay, finds that forgotten passwords stop 19% of account holders from completing checkout (Amazon Pay).
Seventeen years on, the friction is still there. Companies that keep it are not short of information. They lack the reflex to go looking for it.
Why nobody touches that screen
Go back to Samuel. His CRM manager has a point: without accounts, the company collects less data. Marketing defends its customer file, sales defends its conversion rate. Everyone has a good reason, and the form stays.
This is the scenario I see in most organizations. A login screen belongs to everyone, so it belongs to no one. Changing it takes a cross-functional decision, and cross-functional decisions are the ones silos bury best. In large companies, internal politics and turf wars weigh heavily among the obstacles to innovation (My book, chapter 2).
Add status quo bias. What exists feels legitimate simply because it exists. Nobody ever got fired for leaving a form in place. The cost of doing nothing is invisible, it shows up on no dashboard. The cost of a failed change has a name and an owner.
Finally, there is a stubborn belief that innovation must involve heavy technology, R&D budgets, artificial intelligence. The $300 million button shows the opposite. No algorithm, no investment. A label, a reassuring sentence, and the courage to remove something.
What Samuel can do on Monday morning
If you are in Samuel’s shoes, here is the approach I recommend. It takes four steps.
Step one: measure before you touch anything. How many visitors reach the login screen, how many leave, how many request a new password, how many complete their purchase after that request. Without this baseline, you will prove nothing.
Step two: watch your customers. Five recorded sessions are often enough to break a certainty. Spool invented nothing brilliant. He watched people buy, then checked in the database what those observations suggested.
Step three: test on a slice of traffic. One version with guest checkout, the current version, and enough time for the result to owe nothing to chance. The A/B test is the simplest tool to kill an assumption before it kills your conversion.
Step four: make account creation an option offered after payment, as a single checkbox, with a concrete benefit such as order tracking or easier returns. You will recover part of the accounts, this time from customers who chose to open one.
And rate yourself honestly, on a scale of 0 to 5: how much are you subject to status quo bias when it comes to the screens you see every day? I am not always at 5 myself.
The essentials
Remember four things:
- Friction is a hidden cost your customers pay first, and no standard dashboard shows it.
- The $300 million figure cannot be verified, but recent cart-abandonment research confirms the mechanism.
- The data that reveals friction often sits in your own servers, waiting for someone to query it.
- Before removing a step, measure. After removing it, measure again.
Sometimes innovating means removing an obstacle everyone had stopped seeing. Innovate or agonize, the choice is yours!
A corporate event, a seminar, an executive committee meeting, or a management committee meeting?
Philippe’s keynotes on innovation and AI are groundbreaking, you’ve been warned!
Complement a powerful keynote with innovative and impactful workshops.
A keynote inspires and raises awareness; workshops transform!
References
- (My book) https://philippeboulanger.com/book/
- (Center Centre) https://articles.centercentre.com/three_hund_million_button/
- (TheDomains) https://www.thedomains.com/2009/01/16/the-300-million-button
- (CMSWire) https://www.cmswire.com/cms/web-content/web-design-the-300-million-button-003809.php
- (Atticus Li) https://atticusli.com/blog/posts/300m-button-friction-jared-spool-checkout-ux/
- (Baymard) https://baymard.com/research-articles/current-state-of-checkout-ux
- (Amazon Pay) https://pay.amazon.com/us/blog/the-baymard-report-series-how-forcing-sign-ups-drives-down-sales







