When Rebooting A Project, Throw Out The Bathwater But Keep The Baby

It may sound like common sense, but more often than not, we deny the flaws or scrap the project altogether.

When Rebooting A Project, Throw Out The Bathwater But Keep The Baby

[This is the second post in a series by Victor Lombardi, adapted from his forthcoming book Why We Fail: Real Stories and Practical Lessons from Customer Experience Failures. To read the first post, go here.]


The greatest achievement is not in never failing, but in rising again after you fall. — Vince Lombardi

As the 21st century approached, luxury automakers faced a technical design challenge: an increasing array of new in-car devices–phone, GPS navigation, digital radio, satellite radio, CD changers, Internet access, television, emergency notification, side and rearview cameras, and driving system settings–was turning dashboards into an unfathomable mess. Automakers were forced to rethink how drivers interface with cars.

BMW, never a company to shy away from bold designs, introduced the most radical solution: It removed almost all buttons, knobs, and switches that weren’t actually needed to drive the car and replaced them with a single dial. The dial, called the “Controller,” is located on the console between the two front seats and acts like a pointing device to select choices from a screen on the dashboard. This system is called iDrive and debuted in September 2001 on the top-of-the-line 7-Series, a $70,000 sedan ($88,000 in 2011 dollars).

As a concept, the design made sense. BMW could integrate several different subsystems into one system, and unforeseen future functions could be added via software, all controlled through the one central dial and screen.

“A total disaster,” wrote one journalist, summing up the massive backlash from customers and the media. Imagine trying to shop on by drilling down through menus with no way to search, all while driving down a highway. That was the iDrive experience when, say, trying to change the radio station. U.S. sales of the 7 Series declined from 22,006 cars in 2002 to 12,276 in 2008. Worst of all, the company was at risk of staining its storied reputation, as threatened in this 2009 Automobile magazine review:

The BMW 745Li should be the numero uno in this test, but its excellent performance at the test track and its even more impressive performance on challenging country roads are badly offset by iDrive. The iDrive system represents a layer of complexity that actually detracts from what ought to be a breathtaking driving experience. BMW bills itself as “the ultimate driving machine,” and our test car would have been all that and more without the iDrive.

Now, imagine you’re the BMW manager in charge of the iDrive project. What would you do next?


BMW could have scrapped iDrive, but instead they kept it and gradually revised the design. Over the next several years, they added back more physical buttons–some as shortcuts to basic functions like radio, CD, and navigation, and others that are programmable. They changed the software, too, replacing the on-screen circular navigation with hierarchical menus, like the iPod’s. Today, reviewers generally consider iDrive on par with the competition. Here’s what it looks like now:

Regardless of what BMW will tell you, that was one seriously painful product evolution. But the company succeeded in having the discipline over several years to discard the failed aspects of iDrive while keeping what worked.


I have worked with companies large and small and started two businesses of my own, experiencing my share of failure along the way. The most difficult challenge I’ve faced was taking on an established but endangered product business and trying to turn it around. I had some success but not without nights lying awake wondering if I was pulling the right levers. I wish I knew then what I know now from having studied other companies with failed customer experiences that managed to reverse course.

The lesson is this: Don’t throw out the baby with the bathwater, but make sure you throw out the bathwater.

What’s the bathwater? It’s the features that don’t work well or that barely anyone uses. It’s the part of the design that’s confusing, frustrating, or simply extraneous. It’s the stuff that customers don’t like, adds too much complexity to the product, and causes maintenance problems and customer-support pain, and maybe, like BMW, even harms your brand. It’s all dirty bathwater, and it should all go down the drain.


But don’t throw out the baby! The baby is the part that customers love.

This sounds obvious, of course, but unfortunately keep baby/discard water is not usually what happens. Usually when things go south we react in one of three ways:

1. We want to hold on to everything we’ve built. We’ve invested a chunk of our lives creating something, and we can’t bring ourselves to throw any of it away, until…

2. We come to a momentous resolution to make a big change and scrap everything to start over, or…

3. We keep what’s there and incrementally add features in the hope that some improvement will save the day.

It’s rarer to remove what doesn’t work and keep what does. It takes planning and discipline. We can’t generalize that keep baby/discard water is the right thing to do in every situation, but we should recognize when there’s baby and bathwater and act accordingly.

One more example: Plaxo is an online address book that tries to alleviate a pain most of us have: keeping online contact information up to date across our accounts and devices. But Plaxo’s mechanism for obtaining this information was infuriating. Say your friend Joe has you in his address book. Plaxo would send you an email saying, “Can you update your contact information for Joe?” It would do this periodically for every one of your friends using Plaxo, so although you weren’t a Plaxo customer, you were asked to do work to power the service. A lot of people thought this was a giant, annoying waste of their time. Sentiment analysis of Plaxo on the Internet was 73% hate/27% love.

No company can withstand that kind of public revulsion, and eventually Plaxo redesigned the product to keep the synchronization features customers loved (baby) while severely restricting the hate-provoking emails (bathwater). After this change, Plaxo was acquired by Comcast for a significant sum, rumored to be in the $150 million to $170 million range.

No product is perfect, and customers accept flaws. Maybe a car is not quite as fast as the competition, or a web service doesn’t have all the features of a competitor. But when the flaws hurt the customers’ experience, they can lash out at the manufacturer. While both BMW and Plaxo corrected their problems, it took years, years filled with frustrating experiences that hurt the companies’ brands.


So remember: Keep baby/discard water. And don’t take years doing it.