Executive Briefings

For Key Business Decisions, It Makes Sense to Follow the Crowd

If there's one core principle that underlies a democratic society, it has to be this one: trust the crowd. The more people who line up behind an idea, the more chance it has of being the best one. There's deep wisdom buried within the collective mind that can't be accessed by individuals or small numbers of "experts."

So goes the theory behind crowdsourcing, which is often incorrectly viewed as a dressed-up version of "majority rules." In reality, the concept stems from the idea that elements of a solution are often distributed among a large population of supposed non-experts. By casting the net widely enough, one can piece together the answer to any number of tough problems.

Crowdsourcing works its magic by inviting large groups of people to make wagers, using real or virtual money, in support of their opinions. Ideally, the system will have a recursive aspect, in which the future opinions of the most successful bettors are weighted more heavily, resulting in more accurate answers down the line.

The practice has been deployed with varying degrees of success for such purposes as guessing future Oscar winners and victorious political candidates. But its most promising application to date has been in the business world, where a growing number of companies are relying on some version of the "crowd" to guide critical decisions such as how to design a new product, where to stock inventory and how best to allocate merchandise among retailers. The implications for supply-chain managers, who face risk and uncertainty every day of their working lives, are especially compelling.

One of the more fascinating applications of crowdsourcing can be found in the mission of Crowdcast, the four-year-old company about which I've written before. While not the first to bring the concept into the corporate world, Crowdcast has come up with one of its most interesting variations. Under the original model, employees within a client organization were each given a virtual $10,000, which they could use to bet on any question posed by management. The company maintained an ongoing "leaderboard" of the most successful bettors over time, although their true identity wasn't disclosed.

So went the initial approach. But just a few years of experience with the model has prompted Crowdcast to tweak it somewhat, even if that has meant departing from classic crowdsourcing theory. The most obvious change of the last several years has been the emergence of social media. Crowdcast founder and chief executive officer Mat Fogarty says the company has added the ability for participants to engage in conversations and rate one another's wagers. They can also explain the rationale behind their predictions.

An even bigger change revolves around the definition of "crowd." Crowdcast started out with the idea that everyone in an organization should be eligible to participate. But that approach frequently led to "too much noise within the general population, and not that much intelligence about specific numbers," says Fogarty. So Crowdcast sought to shrink the universe of players. In essence, it has replaced the wisdom of crowds with the wisdom of teams. Only those individuals with intimate knowledge of a given project, for example, can wager on its outcome.

One instance highlighted the need for a fresh approach. The ship date for a new software product had slipped five times, and 2,000 employees across the company were betting it would slip again, despite having no inside information on the initiative. Then the project team decided to cut a number of features, and got the product out on time.

In crowdsourcing theory, says Fogarty, the "dumb money" eventually gets washed out of the system. What Crowdcast learned was that most companies don't have the time for that to happen. Hence the new emphasis on "team intelligence." "We can get much better," he says, "by presupposing some of the noise we're going to collect."

Experience has taught Crowdcast additional lessons. An effort to seek feedback from external partners and customers didn't work because those individuals lacked the necessary expertise and level of commitment. "They would come in quickly, leave and never come back again," says Fogarty. To keep bettors with losing records from getting discouraged and pulling out of the game, Crowdcast had to pay off "millionaires" in prizes at the end of each year, then restart everyone with $10,000 apiece. In addition, the company became sensitive to the political implications of generating intelligence that doesn't come from the top. In the case of one client, where a sales forecast by the crowd proved more accurate than the official number, the head of sales shut the system down. Top executives hate to be bested by underlings.

What's next? Fogarty wants to extend the Crowdcast model to the evaluation of softer elements, such as new-product concepts. Interestingly, the tool hasn't been used much in the area of supply-chain risk, which would appear to be an ideal subject for applying crowd intelligence. Supplier stability is an area where groups should be able to offer valuable insight.

Other providers are offering their own takes on crowdsourcing. The Progress Group, a logistics and supply-chain consultancy out of Alpharetta, Ga., is experimenting with a concept it calls crowd engineering, for improving warehouse processes. The idea, says partner Steve Mulaik, is to closely observe individuals as they go about their assigned tasks. Using video and special software that allows for viewing in slow motion, analysts can identify the most efficient workers, learn the secrets of their success, then apply those techniques to everyone in the warehouse.

In an apparel distribution center, for example, one worker appeared to know where her next picking location was, before the system told her. Turns out she was assuming that the next site was adjacent to the previous one, which turned out to be true 88 percent of the time. Says Mulaik: "It was something no one could really see - something we had to dig into the process to uncover."

I might quibble with Mulaik's use of the word "crowd" to describe the system, but there's something inherently democratic about learning from workers themselves, rather than some high-priced process guru. In the supply-chain arena, at least, there's more to crowdsourcing than a trendy buzzword. Without getting too New-Agey about it, think of it as a way for each of us to tap our inner expert.

Watch my video interview with Mat Fogarty of Crowdcast here. And watch editor emeritus Jean Murphy's conversation with Steve Mulaik of The Progress Group here.

- Robert J. Bowman, SupplyChainBrain

Comment on This Article

If there's one core principle that underlies a democratic society, it has to be this one: trust the crowd. The more people who line up behind an idea, the more chance it has of being the best one. There's deep wisdom buried within the collective mind that can't be accessed by individuals or small numbers of "experts."

So goes the theory behind crowdsourcing, which is often incorrectly viewed as a dressed-up version of "majority rules." In reality, the concept stems from the idea that elements of a solution are often distributed among a large population of supposed non-experts. By casting the net widely enough, one can piece together the answer to any number of tough problems.

Crowdsourcing works its magic by inviting large groups of people to make wagers, using real or virtual money, in support of their opinions. Ideally, the system will have a recursive aspect, in which the future opinions of the most successful bettors are weighted more heavily, resulting in more accurate answers down the line.

The practice has been deployed with varying degrees of success for such purposes as guessing future Oscar winners and victorious political candidates. But its most promising application to date has been in the business world, where a growing number of companies are relying on some version of the "crowd" to guide critical decisions such as how to design a new product, where to stock inventory and how best to allocate merchandise among retailers. The implications for supply-chain managers, who face risk and uncertainty every day of their working lives, are especially compelling.

One of the more fascinating applications of crowdsourcing can be found in the mission of Crowdcast, the four-year-old company about which I've written before. While not the first to bring the concept into the corporate world, Crowdcast has come up with one of its most interesting variations. Under the original model, employees within a client organization were each given a virtual $10,000, which they could use to bet on any question posed by management. The company maintained an ongoing "leaderboard" of the most successful bettors over time, although their true identity wasn't disclosed.

So went the initial approach. But just a few years of experience with the model has prompted Crowdcast to tweak it somewhat, even if that has meant departing from classic crowdsourcing theory. The most obvious change of the last several years has been the emergence of social media. Crowdcast founder and chief executive officer Mat Fogarty says the company has added the ability for participants to engage in conversations and rate one another's wagers. They can also explain the rationale behind their predictions.

An even bigger change revolves around the definition of "crowd." Crowdcast started out with the idea that everyone in an organization should be eligible to participate. But that approach frequently led to "too much noise within the general population, and not that much intelligence about specific numbers," says Fogarty. So Crowdcast sought to shrink the universe of players. In essence, it has replaced the wisdom of crowds with the wisdom of teams. Only those individuals with intimate knowledge of a given project, for example, can wager on its outcome.

One instance highlighted the need for a fresh approach. The ship date for a new software product had slipped five times, and 2,000 employees across the company were betting it would slip again, despite having no inside information on the initiative. Then the project team decided to cut a number of features, and got the product out on time.

In crowdsourcing theory, says Fogarty, the "dumb money" eventually gets washed out of the system. What Crowdcast learned was that most companies don't have the time for that to happen. Hence the new emphasis on "team intelligence." "We can get much better," he says, "by presupposing some of the noise we're going to collect."

Experience has taught Crowdcast additional lessons. An effort to seek feedback from external partners and customers didn't work because those individuals lacked the necessary expertise and level of commitment. "They would come in quickly, leave and never come back again," says Fogarty. To keep bettors with losing records from getting discouraged and pulling out of the game, Crowdcast had to pay off "millionaires" in prizes at the end of each year, then restart everyone with $10,000 apiece. In addition, the company became sensitive to the political implications of generating intelligence that doesn't come from the top. In the case of one client, where a sales forecast by the crowd proved more accurate than the official number, the head of sales shut the system down. Top executives hate to be bested by underlings.

What's next? Fogarty wants to extend the Crowdcast model to the evaluation of softer elements, such as new-product concepts. Interestingly, the tool hasn't been used much in the area of supply-chain risk, which would appear to be an ideal subject for applying crowd intelligence. Supplier stability is an area where groups should be able to offer valuable insight.

Other providers are offering their own takes on crowdsourcing. The Progress Group, a logistics and supply-chain consultancy out of Alpharetta, Ga., is experimenting with a concept it calls crowd engineering, for improving warehouse processes. The idea, says partner Steve Mulaik, is to closely observe individuals as they go about their assigned tasks. Using video and special software that allows for viewing in slow motion, analysts can identify the most efficient workers, learn the secrets of their success, then apply those techniques to everyone in the warehouse.

In an apparel distribution center, for example, one worker appeared to know where her next picking location was, before the system told her. Turns out she was assuming that the next site was adjacent to the previous one, which turned out to be true 88 percent of the time. Says Mulaik: "It was something no one could really see - something we had to dig into the process to uncover."

I might quibble with Mulaik's use of the word "crowd" to describe the system, but there's something inherently democratic about learning from workers themselves, rather than some high-priced process guru. In the supply-chain arena, at least, there's more to crowdsourcing than a trendy buzzword. Without getting too New-Agey about it, think of it as a way for each of us to tap our inner expert.

Watch my video interview with Mat Fogarty of Crowdcast here. And watch editor emeritus Jean Murphy's conversation with Steve Mulaik of The Progress Group here.

- Robert J. Bowman, SupplyChainBrain

Comment on This Article