Affinnova, Evolver of Consumer Products, Evolves Itself
Creating a new diet soda, deodorant, or dish soap is a notoriously chancy business: consumers are unpredictable, and experience and intuition can only get a product designer so far. But what if you could apply the power of evolution to product development—subjecting various ideas to generation after generation of Darwinian competition, with consumers themselves as the agents of selection, picking which attributes of each product survive into the next generation?
This exact idea was one of the hottest trends in product development—in the year 2000. That was when a group of entrepreneurs and researchers from MIT, with backing from Cambridge, MA-based Flagship Ventures, founded a startup called Affinnova with the goal of applying a new form of mathematics called evolutionary algorithms to the traditionally expensive and time-consuming process of market-testing of new products. The company grabbed a few big clients such as Procter & Gamble, and even attracted imitators like Cambridge’s Icosystem. But eight years on, the field of evolutionary algorithms still hasn’t produced a smash-hit startup; after three rounds of venture financing totaling some $25 million, Affinova is only now approaching profitability.
But the company’s roughest years may be over. While Affinnova (pronounced ” AFF-i-NO-va”) doesn’t release actual sales figures, it says sales in the first half of 2008 were up by more than 60 percent compared to the first half of 2007. Its client base grew by one-third in the same period, and the company is in position to bring in $25 million in revenues next year, according to chief marketing officer Steve Lamoureux.
This growth can be attributed largely to changes made since the arrival of CEO Waleed Al-Atraqchi in 2005, Lamoureux says. (Al-Atraqchi replaced David Andonian, who went on to found Dace Ventures.) “Waleed is a turnaround guy,” he says. “There was an underlying asset that was being underleveraged in the business world. We were top-loaded with PhDs making advances in evolutionary algorithms, but we were putting fewer resources into making that technology understandable and adoptable. One of the things Waleed has done is to balance that out. We’ve taken the technology and made it really streamlined and sticky, and as a result we get a lot of repeat business.”
So, what exactly has Affinnova been doing to evolve its own strategy? Lamoureux and Al-Atraqchi filled me in about two of the company’s initiatives last week when I visited their offices in Waltham. The first, it turned out, was to stop helping companies come up with new product ideas, and start helping them pick between the ones they already had.
When I first became aware of Affinnova around 2002, the company was already a couple of years old, and had been focusing largely on helping designers dream up new product ideas that stood out from the crowd. In one example, the company’s Web-based software created thousands of variations on an old standby–the plastic water bottle—and asked participants in online focus groups to pick the ones with the most exciting combination of shapes, contours, and label colors. The process had the potential to arrive at radical combinations that no human designer would have bothered (or dared) to try, but that tested well with consumers anyway.
But Al-Atraqchi says that when he arrived at Affinnova, “I said, in a sense, let’s not confuse ideation with optimization. Most studies say that companies don’t lack for ideas. Which ones are the best, and what to do with them, are really the big issues, and that’s the problem we should try to solve.”
The company has found a niche, according to Lamoureux, by helping consumer manufacturers bringing out new products find the sweetest spot within a fairly narrow range of pre-determined options. Dannon, for example, came to Affinnova a few years ago for advice on the best way exploit the low-carb craze then sweeping the nation. The company knew it wanted to bring out a new brand of low-carb yogurt; the only question was how to package it. Should the label be teal, as Dannon’s experts thought, or some other color? Should the package hold 6 ounces of yogurt or 8? Should the product be called “Carb Control” or “Carb Smart” or “Carb Oh”?
Affinnova tested combinations of all of these options and more, and was able to determine not only which overall product concept was most appealing to consumers, but which elements were most important to their choices. The winner: a red package branded Carb Control, which went on to become one of the year’s top new food products, earning Dannon $75 million in its first year on shelves. A competing product, Yoplait Ultra, came out in a teal package and failed within a year.
Affinnova has refined its evolutionary process to the point that it can create competitive concepts even for products no one has heard of. As an exercise, Lamoureux paid a design firm to come up with three concepts for the packaging for a hypothetical new product—a cleaner for stainless steel kitchen appliances. The three designs ranged from bright and cheery (targeted at general consumers) to elegant and understated (targeted at older, more affluent consumers). In Affinnova’s tests on real consumers, participants gravitated to a somber black package that emphasized the product’s roots in professional kitchens. And when the company tested that design against an existing product from Pledge, 34 percent of consumers said they’d take home the hypothetical Affinnova brand, while 49 percent said they’d buy the Pledge product.
“I’m thrilled with that outcome,” says Lamoureux. “It means I got 70 percent of Pledge’s market share without doing any advertising. Pledge has 50 years of brand-building behind it, and I made up this product three months ago.”
But as powerful as evolutionary tools have become, there’s still a problem: companies aren’t hiring Affinnova until near the end of their product development process, when they have a few solid concepts that they’ve already run through a painstaking process called “monadic” testing. So Affinnova’s second initiative is an effort to get product designers using its services earlier—after the “ideation” stage, but before monadic testing.
Monadic testing companies take concepts for new products or packages and show them to consumers, then score their reactions and compare those scores to historical databases. Concepts that reach a certain benchmark—scoring, say, in the top 50 percent of all products ever tested—can go forward in the development process. Monadic testing is a well-established art, with more than a decade’s worth of product scores stacking up in databases, so companies aren’t going to skip it. But “right now we have to wait around for this process to end, for the concept to be identified, before we can help with all the other stuff,” Lamoureux explains. “It’s irregular business, because we’re waiting for something to come over the wall. From our business perspective, there would be a strategic advantage to being at the beginning of the process, to help them get more ideas over the wall.”
That’s the role Affinnova envisions for a new service called Power Screener. It’s a quick-and-dirty version of the company’s trademark evolutionary optimization process, intended to help companies pre-screen variations on their basic product concepts before they send them for monadic testing.
Affinnova is marketing the tool as a time- and money-saver. “Every time you do a monadic test, it costs $7,000 and takes two months,” says Lamoureux. “On average it takes 25 monadic tests to find one concept that meets your benchmarks. That’s $175,000 to get one concept that works. With our system, you can go through Power Screen for $15,000 and get two to five top concepts that you can then present for monadics, so you are at a total cost of $60,000. For the Procter & Gambles of the world, that can save you millions. And by eliminating the wait [for repeated monadic tests] you can cut months out of your whole go-to-market strategy.”
Lamoureux says the company is putting the finishing touches on Power Screener right now and will start selling service this fall. If it’s successful, it could help companies bring new brands to market even faster, and give each one a better chance of succeeding. But there’s one thing manufacturers will always have to do for themselves, says Lamoureux: make decent products that actually improve consumers’ lives.
“A better package can increase your trial potential, and to the degree that the package and the message and the concept help, we can maximize that; we can put the best lipstick on the pig,” Lamoureux says. “But if the product stinks, it’s still going to fail in the marketplace.”
Trending on Xconomy
By posting a comment, you agree to our terms and conditions.