Crowdsourcing is popular but it causes problems – too many ideas, not enough solutions being one of them. How do you do crowdsourcing properly? Klaus-Peter Speidel talks you through your options

While I’ve been using the term “Intelligent Crowdsourcing” for a while, I have never given a clear definition of what it is in written form, and, how exactly it is different from crowdsourcing simpliciter.

Jeff Howe, who came up with the term, defines Crowdsourcing as “the act of taking a job traditionally performed by a designated agent (usually an employee) and outsourcing it to an undefined, generally large group of people in the form of an open call.”

Crude Crowdsourcing

Crowdsourcing, in its most crude form distributes the power of contribution as widely as possible. No more privileges. Finally, innovation is being democratized.

And the approach indeed has many virtues: by eliminating biases that lead to unjustified preference for certain contributors, it can allow you to find solutions you had not been looking for: ideally, the best. If those who contribute ideas and vote on others’ ideas are also your customers – as in Dell’s Ideastorm, one of the most successful corporate crowdsourcing initiatives – market research and innovation is totally integrated – and free. If there are 118,590 people (as of April 11th, 2011) who like the idea of having OpenOffice preinstalled on all Dells, you know for sure there’s a customer want. Now if Dell doesn’t want to look tyrannic to a web 2.0 public, they really have to take the idea seriously – which could be seen as a downturn of Dell’s form of Crude Crowdsourcing (this is linked to the difference between Overt vs. Covert Crowdsourcing below). And Dell’s laconic – while sound – status update, “Dell is constantly reviewing options to offer to customers. As this is available for free today, we do not currently plan on offering this as an option to factory install”, isn’t sure to satisfy the people who first suggested the Open Office pre-install.

While this episode give a sense of how touchy Crude Crowdsourcing can be, many private companies and some public institutions have recognized the multiple values of crowdsourcing and launched their own initiatives. The principles of crowdsourcing have been applied to things as diverse as finding new names for products and services, create designs and to support innovation – which is what concerns us here.

Many organizations follow the early example of IBM’s Innovation Jams and regularly organize large innovation competitions to which employees across all functions contribute. IBM funded 10 ideas generated in its 2006 jam with $ 100 Million. In the last years, many companies have specialized in organizing crowdsourcing initiatives for others and many platforms that support crowdsourcing on the web or in a corporate intranet are available.

Intelligent Crowdsourcing

There are numerous problems with the crude form of crowdsourcing – where everyone contributes and all contributions are overt, especially when it comes to Research & Development. The needs and problems which give employees a hard time are generally so specialized that not everyone has something useful to say, so there is no point in trying to get everyone to contribute.

There are several reasons why you’ll want to avoid that everyone contributes in this kind of case: First of all, quantity: if you get too many ideas, it will immobilize too many people for too long to evaluate them. Second, quality: if contributors are not targeted, you will get many ideas that are besides the point, which will reduce the overall credibility of the initiative within your organization. On the contributors’ side, a lot of people will have contributed without being rewarded and you will make many people unhappy. If you reach out actively for solvers and solicit the wrong people too often, they’ll end up redflagging your organization.


If you get too many ideas, it will immobilize too many people for too long to evaluate them. If contributors are not targeted, you will get many ideas that are besides the point, which will reduce the overall credibility of the initiative within your organization.

Another point has to do with competition: if solutions to R&D and innovation challenges can be found, their value for an organization will be higher if the competitors don’t have access to the same solutions you are using. In an overt intiative like Dell’s everyone can follow the ideas with growing customer support: Asus can decide to implement an idea before Dell does.

This is where intelligent crowdsourcing offers a significant twist. The expression is voluntarily ambivalent: according to the scope of “intelligent”, we are either talking about intelligent crowds or about intelligent sourcing.

There is no intelligent crowd per se. Different problems and needs call for different crowds. Rather than trying to engage as many people as possible, intelligent crowdsourcing includes a phase where people with need-relevant expertise are identified.

Some companies have developed social or technological methods to do so. The social approach tries to tap into existing communities or networks (suppliers, university partners or existing networks of scientists, or users of products). Some companies use netnography to identify relevant crowds. This approach works best for lead-user identification and approaches that want to integrate innovation and consumer research.

The technological approach is based on exploiting data. My company hypios, which entices top experts in various fields to compete to solve problems in research and development, uses a special search motor to find problem/need-relevant experts based on different sources, e.g. the public activity traces they leave on the Internet. These potential solvers are then invited to submit a solution to a specific contest which fits their profile.

The tricky part is to keep the selection precise enough to get relevant people, but open enough to make sure, serendipity – the ability to find solutions you had not been looking for – is maintained.

This approach avoids many of the risks that are supposedly associated with Crowdsourcing. Gettting too many ideas on hand, but also getting no ideas at all: by actively reaching out to relevant people who have an interest in the problem being posed, you will almost certainly get contributions.

Crowdsourcing ideas or solutions?

Looking for solutions always implies a problem statement, however specific it may be. Ideas can be spontaneous creations and don’t imply that a need or problem has been defined.

I sometimes use crowdsourcing “solutions” and “ideas” as synonyms. But while they look similar from a high-level perspective, they are quite different on a practical level. Looking for solutions always implies a problem statement, however specific it may be. Ideas can be spontaneous creations and don’t imply that a need or problem has been defined. In terms of process, ideas are most useful at the entry of the innovation funnel. At later stages of the R&D process, you usually need workable solutions for very specific problems.

To source ideas, it can be enough to specify topics. You’ll get more submissions and they will be less targeted than if you formulate problems, so you need more time for evaluation. But what you get can be more surprising and radical. If you only source ideas and implement communications strategies and incentives, the risk of getting no submission is extremely low. Ideas are more common than expertise.  But the risk of getting too many uninteresting proposals is high.

The number of ideas you will get depends on how much you advertise, but also on how precise your brief is. If you look for “green technologies”, which is pretty much what the GE Ecomagination challenge did, you’ll get more ideas than if you look for technologies or methods for a more precise estimate of energy consumption.

Idea Overload, Covert Crowdsourcing, Groupthink

As noted above, creating a topic-related crowd limits the risk of getting too many irrelevant ideas which make evaluation difficult.

There’s another advantage: as you start with a personal invitation where you ask people to become members rather than just setting up a website, people who are unsatisfied with the way you manage the contest are likely to address you directly rather than going public. Most people still don’t perceive going public with a complaint as the standard but as the last way out when they can’t find anyone to treat their concern. One main reason for doing an overt competition, where everyone can comment and vote on everyone else’s competition, is that it makes it easier to sort through solutions. By selecting a relevant crowd, you limit the number of irrelevant submissions. You can thus go covert and let your team do the selection.

This also avoids the risk of your competitors stealing an interesting idea before it’s implemented or protected.

Another advantage of covert crowdsourcing is that it impedes groupthink, the tendency of members of a group to converge towards a few ideas and to minimize dissent.  While overt submissions allow participants to build on each other’s ideas, reading through existing submissions may limit the way they think about the problem. Furthermore, public verdict can kill some ideas too early in overt crowdsourcing.

In covert mode, participants will have developed them to a higher degree before anyone gets to see them.

Aggregate for Sustainability

For R&D problem-solving, where problems tend to be highly specific and require high degrees of specialization the risk of not getting contributions is higher than getting too many contributions. SMEs or corporations that are not consumer brands like Lego or Dell, have found it hard to generate enough outside interest for problems. The solution of using an intermediary who aggregates needs for different companies, creates and maintains a reputation as a trusted third party, builds and manages a community, develops dedicated tools and methods to implement this approach to open innovation in corporate culture, can be a way to avoid too many orphan problems on a deserted corporate platform. Intermediaries will probably become even more relevant in the future.

Here’s why:

The number of portals is already high and most Fortune 500s still lack one. Engaging researchers and inventors will be even more of an issue when more companies launch their own initiatives trying to convince innovators to work with them. The researchers I know are definitely not people who will make their already busy days more busy reading through each of the 20+ innovation needs posted on platforms like Innovate with Kraft, Procter & Gamble’s Connect&Develop, RB’s Idealink, or the nicely named newcomer Pearlfinder at Beiersdorf and the many others.

Problem-Statement and Confidentiality

Clearly stating what you want to achieve with solutions and telling people which methods you have already explored or cannot use and why, will help you get submissions that are to the point. Making your problem-formulation open and general will allow people from diverse backgrounds to make interesting contributions.

On a corporate portal, problems are by definition branded. Intermediaries, on the other hand, allow you to make your need/problem anonymous, so you don’t have to get communications or marketing involved. However there are reasons to think that confidentiality is often badly understood. As Kevin McFarthing, who implemented Reckitt Benckiser’s corporate innovation portal Idealink underlines: “You should always ask yourself if it would surprise our competitors that we are posting this problem? If the answer is yes, then think twice. But in most cases the answer will be no.” This is because whole industries are often looking for the same solutions. The rationale for being the first to make a problem public is high: if you are the first to get a problem in front of a larger and more diverse crowd of experts than everyone else, there is a good chance that your company will be the first to get new solutions.

By Klaus-Peter Speidel

About the author:

Photograph by Sandrine Photographie(s)Klaus-Peter co-founded hypios in 2008. Relying on semantic web tools as well as Artificial Intelligence, hypios now has a broadcast reach of more than 700.000 Solvers. Today, several Fortune 500s from different industry sectors (like consumer-goods, aerospatial engineering, mining and others) regularly rely on hypios to solve their R&D problems.

Klaus-Peter lives and works in Paris, teaches a class at Sciences Po, and was an invited speaker at PDMA, Palomar 5 (Deutsche Telekom), The Masters of Innovation in European Business, as well as other companies and organizations. He regularly writes and talks about expert- and crowdsourcing, innovation and open problem-solving. Klaus-Peter’s recent publications include “Problem Description in Open Problem Solving. How to overcome cognitive and psychological roadblocks.” in A Guide to Open Innovation and Crowdsourcing, edited by Paul Sloane at Paul Kogan.