For Successful Innovation, Sell Imperfect Products

If you had tasted the very first PowerBar, you probably would have gagged.
It was a chewy, tasteless mass of goo: barley malt, nuts, brown rice, molasses, and nutrients blended into a taffy-like bar, packed in a metallic wrapper with a crude logo.

Created by Brian Maxwell in 1986 for a few thousand dollars, PowerBar received mixed initial results. Serious athletes liked it because it met an important need—quick energy during competition—but the taste and consistency needed work.

Maxwell knew his product wasn’t optimal. But by getting feedback on an early version, he was able to modify it, changing the package and marketing strategy to build a following among athletes and weekend warriors. PowerBar eventually became a $150 million business, creating the $1 billion energy bar category. In 2000, Nestlé bought the brand for about $350 million.

Maxwell’s approach of purposely launching an imperfect product has been adopted by many well-heeled organizations recently as a new innovation tactic. Rather than making big upfront research and development investments in new ideas, these companies instead start with an MVP, or minimum viable product, a term coined by Eric Ries, author of The Lean Startup.

The concept behind the MVP is to find the fastest, most cost-effective way to build a salable product that delivers on your basic idea and induces a reaction from prospective buyers. Even if many don’t buy, this approach generates specific feedback on individual features, providing a blueprint for the ultimate design.

Apple builds great, enjoyable products largely because they live by this approach. “It’s hard to design products by focus groups,” co-founder Steve Jobs once said. “People don’t know what they want until you show it to them.”

Apple’s run of success has not been a matter of luck or genius but a result of the company’s process of testing a wide range of early prototypes and analyzing user feedback before selecting the best ones. Every one of the iPad’s design features is there because it beat out less successful options during early-stage testing.


Many corporations test products during the development cycle but not early or often enough with actual customers. R&D is the most expensive step in the innovation process. If you wait until you have something close to a finished product before getting user feedback, large sunk costs will make your team reluctant to employ significant changes.

Procter & Gamble has a great R&D organization that spends about $2 billion per year (2.3 percent of sales). But recently they’ve been out-innovated in key categories by Estée Lauder (beauty care), Reckitt Benckiser (household products), and L’Oréal (personal care), companies that invest at much lower rates but move new products more quickly to market.

The first Nike shoes were made on a waffle iron by University of Oregon track coach Bill Bowerman. While not a thing of beauty, they allowed Bowerman and Nike co-founder Phil Knight to discover the ideal running shoe design and turn their efforts into a huge new business.

To this day, Nike uses athletes to help develop and test new products during the design stage, giving the company a competitive advantage over its peers.


Rather than just conducting research studies, however, you need to go further to find out which products and features will entice consumers actually to spend their hard-earned money.

Starbucks is a master of this. Every new product concept—from a new flavor of Frappuccino to an updated store design—is tested in a handful of stores while the idea is still new. Fast and frequent testing leaves competitors unable to figure out what the coffee giant is really up to.

When Starbucks launched its line of instant coffee, Via, many pundits questioned whether the product would succeed. They didn’t know Via had already survived several rounds of store testing that measured its real sales. From this testing, Starbucks ensured it had the right product, package, branding, and marketing program. The company knew exactly what to expect from Via, now a $180 million brand.

Lowering product development costs is an important benefit of early market testing. Since more than 90 percent of new products fail the first time they’re launched, smart companies have discovered they can improve their success rate simply by testing an early version.

Paul Graham, an entrepreneur turned venture capitalist, started a Silicon Valley incubator called Y Combinator to help startups raise seed money. Graham has lowered average product development costs of his portfolio companies to less than $100,000. At this low cost, Y Combinator can fund almost 100 startups per year. In less than five years, Graham has had more than two dozen successes. Some have gone public; others, such as Dropbox and Airbnb, have raised subsequent private funding at valuations of more than $1 billion.

Graham is not any better at picking winners than other early-stage investors, but he makes a lot more investments and keeps his costs low. He pushes every company he invests in to get a minimum viable product to market quickly.

“Launching helps you find out how you suck,” Graham told Dave McClure and other investors at a 2009 Facebook fbFund meeting. “Until you launch, you might suck, but you won’t know why.”


If you’re running a big company, getting your costs anywhere near “Paul Graham levels” can prove difficult. An early-stage PowerBar would never get out the door: R&D would want a better formula; packaging would need a more elaborate wrapper; a design agency would have to create the right label; manufacturing and sales would have to weigh in; and the corporate lawyers would insist on a detailed risk analysis.

Companies able to get around this red tape generally have dedicated innovation teams working independently, or they outsource to build and test products.

Coca-Cola has established a Venturing & Emerging Brands unit that made early-stage investments in Honest Tea and Zico Coconut Water, which have become two of the beverage industry’s hottest new soft drinks. If the traditional Coke organization had tried to develop and launch these products from scratch, they might still be in the R&D lab.

Not all of these projects work out. In fact, for every PowerBar, there are nine failures. But the beauty of the MVP model is you can experience hundreds of failures and walk away worry-free, knowing you just saved millions in R&D, marketing, and manufacturing costs.