The Industry Website

Web's Largest listing of crowdsourcing and crowdfunding events


Web's Largest Directory of Sites

2,968 crowdsourcing and crowdfunding sites

"Towards a New Taxonomy" by Jeff Howe

"Towards a New Taxonomy" by Jeff Howe

When I coined the word crowdsourcing five years ago, the phenomenon itself hardly existed. Oh sure, there were a few glaring examples: Wikipedia, of course, and that massive hurly burly of user-generated content known as MySpace. NASA, I recall, was already experimenting with using volunteers to measure asteroid craters. But the original article was, in a sense, an act of prognostication. I was saying, "This is how things will be." It was a gamble, and it happened to pay off.  


A lot changed between 2006 and 2008, when my book was published. In writing the book I faced an entirely new challenge. I no longer had to predict the future; I had to analyze and interpret the present. Books force you to create taxonomies by their very structure: Chapters are subdivisions of a larger subject. 

The problem was that examples of crowdsourcing now proliferated across every domain of human endeavor. When asked what couldn't be crowdsourced, I used to answer: a restaurant. Then one day a Washington Post reporter called to ask me what I thought about a new--you guessed it--crowdsourced restaurant. My answer to that question these days? Everything can be crowdsourced. I figure it's best to play it safe.

In trying to bring some coherence to this dizzying variety of online collaborations I came up with four basic categories:

1) What the Crowd Knows: A large share of what we think of as crowdsourcing involves the attempt to solve problems by tapping into what people already know. This is nearly synonymous with what's often thought of as the "wisdom of the crowds," which as I always say, is just one type of crowdsourcing. To quote the economist, F.A. Hayek, society is best able to solve its problems "not by the acquisition of more knowledge, but by the utilization of knowledge which is ... widely dispersed among individuals." 

What Hayek couldn't know is that the Internet would come along, allowing us to gather up much of that previously dispersed information. In this category I included firms like InnoCentive, which broadcasts scientific problems to a large network of scientists, in this category, as well as efforts like the Netflix Prize, which awarded $1 million to the team of researchers who improved their recommendation engine by 10 percent.

2) What the Crowd Creates: wants you to make something then sell it to them for very a little bit of money. So do logo design firms like CrowdSpring and 99Designs. Ditto "journalism" startups like Helium and, for that matter, AOL's Patch. If there was one crucial distinction I drew in my book, it was between forms of crowdsourcing that rely on what people already know (see above), and what people can make. As we'll see, I still believe that drawing this line is the most important aspect of any crowdsourcing taxonomy, but it needs to be refined.

3) What the Crowd Thinks: What do Yelp, "American Idol," Amazon, and The New York Times have in common? They all rely on the crowd to either explicitly (Yelp) or implicitly (The Times, through its "most-emailed" bar) pass judgment in order for them to organize their content.. As I noted in my book, this may be the most pervasive use of crowdsourcing of all. It also raises one problem we face in creating a current taxonomy: Crowd filtering is in fact so ubiquitous (try finding a web commenting system that doesn't allow you to "like" or "dislike" someone else's comment) that it has, in a sense, transcended the larger category of crowdsourcing. It's simply become, well, the Internet.

4) What the Crowd Funds: This refers to crowdfunding, a splinter term of crowdsourcing that I first saw pop up in reference to the Dutch music company Sellaband back in 2006. Of all my four categories, this is the most discrete and coherent. There are no blurry lines here: If crowdsourcing is Wikipedia with everything, crowdfunding is Kickstarter with everything.


Now we get to the challenge presented by the state of affairs facing us today. The purpose of this post is not to rehash old categories. It is partly to reconcile my old taxonomy with the taxonomy presented by Carl Esposti on this site back on 16 May 2011. More to the point, I'm proposing it as a working draft. (I'm open to the suggestion that we go so far as to post it as a wiki.) My hope is that it might lead to a consensus among practitioners, scholars, entrepreneurs, and of course, the poor hacks like me who make a living trafficking in these terms.

Here is my attempt at a refined taxonomy, something that's robust enough to bear scrutiny, yet flexible enough to contain all those examples of crowdsourcing we haven't even encountered. I'm going to move from the simple to the tricky. Let's start:

1) Crowdfunding: As we say in the news business, stet. That means: Don't touch a thing.

2) Tools: This category was proposed by Esposti, and I think it's a valuable addition to the way we think about crowdsourcing. Someone has to provide the engine for crowdsourcing, and we all need a category for the companies that aren't necessarily crowdsourcing themselves, but are providing the software, the experience, the wherewithal to do it. Ushahidi and Chaordix both provide archetypes for this classification.

3) Open Innovation: Most of the companies and crowdsourcing efforts that fall under this category would have originally been put inside my first category, "What the Crowd Knows." This area has grown dramatically over the past few years. InnoCentive has now been joined by new entrants like IdeaConnection, and large corporations like Nokia and Unilever have employed open innovation to augment their in-house R&D departments. 

I have a really simple argument for sticking with this term: Don't fix what's not broken. Open innovation has become part of the business lexicon, and there is fairly wide agreement over what it means. I would, however, guard against its misuse by restricting the label to companies who are, in Henry Chesbrough's words, wholly or in part relying on "external ideas as well as internal ideas" as a way to "advance their technology." 

4) Cloud Labor, or Distributed Labor: When I was writing my book I encountered a hell of a pickle: How would I classify Amazon's Mechanical Turk? As you can see above, MT doesn't fit neatly into any of the categories I created. The process by which MT--and the many third-party firms that used it as a platform--procured labor was similar to what services like CrowdSpring were doing: Matching clients with an always-on, scalable workforce. But the nature of the labor was hardly creative, right? In the end, I wrote a separate chapter on Mechanical Turk, and it wound up on the cutting room floor.

Esposti neatly resolved my error by creating a category he calls "cloud labor," which includes companies ranging from CrowdFlower to oDesk to Clickworker

I would go a step further. I think there's a strong case to be made for merging this category with one that encompasses "creative crowdsourcing." One man's "creativity" is another man's drudgery. How we classify the nature of the labor is subjective, and thus doing so creates a false distinction. It's no accident that oDesk (a "cloud labor" company, under Esposti's system) currently lists 149 jobs for illustrators. They've invaded Crowdspring's territory, and with good reason: They're really very similar services.

I'll go further with this, because it represents a wholesale rethinking of a classification system into which I put months of thought: I was never terribly happy to create a taxonomy around the type of work that was being crowdsourced. It's an inherently unstable designation. Let's take the term out of our taxonomy. As a final note, while I like Carl's term "Cloud Labor," which makes the obvious link to cloud computing explicit, I like "Distributed Labor" even better. Either one is preferable to any category that incorporates the slippery term "creativity."

5) Distributed Knowledge: Again, I see the central division in crowdsourcing as laying between firms that tap people's knowledge and their productive capacity. I think it's a fairly obvious, and easy distinction to make (though I recognize that even here there is overlap.) Distributed Knowledge basically just puts my twist on what Esposti calls "Collective Knowledge" It includes a lot of what journalism outfits like the New York Times and CNN are doing in that they seek to exploit what someone knows (often by simple virtue of being at the right place at the right time). And its purest expression is probably companies like the social news aggregator, Reddit, or even Dutch navigation system TomTom. Reddit asks us to submit what we're reading, and what we think about what others are reading. TomTom simply asks us to update their navigational data based on what we discover as we drive.

There's a nifty trick here. I had thrown services like Reddit into my "What the Crowd Thinks" category. But what we think and what we know aren't readily distinguished. Esposti recognized this in creating the category "Collective Knowledge," and thus helped resolve another failing of my original scheme.

The Outtakes: One of my goals in this post was to simplify current crowdsourcing taxonomies. To that end, I'm going to advocate abolishing "civic engagement" and "community building." Like the notion of "creativity," I think we step down a slippery slope as soon as we try to evaluate whether a company or institution or non-profit has civic or philanthropic motives in mind. We should categorize them according to the type of crowdsourcing in which they're engaged. Finally, "community building" strikes me as one of the steps toward crowdsourcing, but it isn't in and of itself crowdsourcing. It loads the gun, in other words. It doesn't fire it. Thus, I would classify Get Satisfaction under Distributed Knowledge, not community building.

Am I wrong about any of this? Surely. As I said, this is meant to start the conversation, so give me your edits, your complaints, and your modifications. I look forward to hearing from you!


Flag This



  • Guest Dorothy Jul 21, 2011 11:02 am GMT

    And is there a new version of chicken and egg, as in 'what comes first, the crowd or the type of crowdsourcing'. Does the category define the crowd, or the crowd define the category?

  • Guest susan Jul 21, 2011 12:46 pm GMT

    I have been asked to write a chapter on Crowdsourcing in the Cultural Industries - and have just done so by trying to create my own taxonomy of the different ways this is happening in museums, libraries and archives. In spite of Jeff's seminal contribution to the concept of crowdsourcing - I agree that there is still a lot of work to be done. Thanks for this taxonomical approach - very timely!

  • Leo T Leo T Jul 21, 2011 01:08 pm GMT

    Jeff, you have a point, actually you have multiple points. When I think of crowdsourcing I focus on the "sourcing" part of it. We can absolutely source funds, knowledge and labor. With a little effort I can imagine sourcing innovation too (actually we are sourcing expertise, not innovation), but under no circumstances I see community as a way to source something. You're right, it's just a step toward achieving a great goal.
    I feel a little uneasy combining ALL types of workers and skills into Cloud Labor. There is a reason why traditional outsourcing folks make clear distinction between IT labor, process labor, etc., but we can probably accommodate it through subcategories within Cloud Labor.

    To summarize, my vote is:
    Crowdfunding: yes
    Tools: yes
    Open Innovation: yes, if you wanna be cute with name
    Cloud Labor: yes (gotta re-think subcategories)
    Distributed knowledge: yes (where else we would keep "Price of Weed" sites :) )
    Civic Engagement: no
    Community building: no

  • Guest Arie Goldshlager Jul 21, 2011 01:22 pm GMT


    Thank you for a very useful initial Crowdsourcing taxonomy. Just a couple of suggestions for your consideration:

    2) Reposition Tools. The tools facilitate all the other elements of the taxonomy. They are best presented as a separate class.

    3) In addition to Open Innovation consider Distributed or Collaborative Innovation.

    4) Change Cloud Labor/Distributed Labor to Collaborative Production or Collective Production.

    5) Consider adding another element: Narrowsourcing for Crowdsourcing focused on selected “crowds” of lead-users, creative consumers, or emergent consumers.

    Thanks again,


  • Guest Michael Jul 21, 2011 01:49 pm GMT

    Johan Oomen and Lora Aroyo have published a paper outlining a taxonomy of crowdsourcing tasks in the cultural heritage fields. The paper can be viewed here:

  • Jeff Howe Jeff Howe Jul 21, 2011 04:26 pm GMT

    Thank you all. I'm generally out of pocket today on a rough deadline, but will respond in detail to these tonight or tomorrow morning. I'm really hoping to keep the conversation going and bring in as many voices as possible. I actually think the taxonomy via wiki might be a good bet.

  • Daniel Daniel Jul 21, 2011 10:13 pm GMT

    I agree overall with eliminating the "Civic Engagment" and "Community Building" categories. I do believe however that the "type of work" that is being crowdsourced is relevant. Would you suggest to completely ignore this aspect, or would you capture it in some other ways, such as by using sub-categories (i.e. the current taxonomy on contains sub-categories that can be seen in the "Direvtory of Sites)?

  • Carl Esposti Carl Esposti Jul 22, 2011 06:50 am GMT

    6) The problem I see with Arie’s recommendation to change Cloud Labor/Distributed Labor to Collaborative Production or Collective Production is that the efforts are not always collective or collaborative. For this reason I am also okay with Jeff’s preferred term “Distributed Knowledge” rather than Collective Knowledge.

    7) Re Arie’s comment – “consider adding another element: Narrowsourcing for Crowdsourcing focused on selected “crowds” of lead-users, creative consumers, or emergent consumers” – I see this as a practice within crowdsourcing and not a separate category.

  • Carl Esposti Carl Esposti Jul 22, 2011 06:51 am GMT

    Thanks to all for the fruitful discussion.

    1) I’m easy with either Cloud Labor or Distributed Labor – we went for the association with Cloud Computing but I could live with either.

    2) We felt that there was so much activity that was specific to the creative fields of art, film, design etc. that Collective Creativity – at least at the time – was worthy of a stand alone category – as some work is the result of collective contributions that come together to form an individual artifact and some the result of independent work maybe Distributed Creativity is a better term.

    3) I’m still an advocate though of Community Building – I have to disagree with Jeff, Daniele and Leo on this one. Re Leo T’s comment “but under no circumstances I see community as a way to source something”, there are examples of where the directive – what you are “sourcing” is in fact the community in its own right with the benefits being the results of either the animation of the community for purposes such as customer retention through greater brand engagement or to develop brand advocates who create other customers. I don’t see this a Cloud/Distributed Labor as you are not specifying a defined task for “workers” to perform. Good examples are CrowdEngineering (customer driven support and customer-to-customer selling) and CrowdTogether (driving customer/fan engagement with brands/events etc.)

    4) I agree with Arie Goldshlager’s comment that Tools is a separate class – we didn’t show it as such on the CS Industry Landscape document but you will see on the site that it is a separate link and not one of the category menu items.

    5) Another of Arie’s considerations was to view Distributed or Collaborative Innovation in addition to Open Innovation. I would see these as sub-categories – forms of OI.

    (continued below)

  • Guest Maxine Horn Jul 22, 2011 09:25 am GMT

    Thank you Jeff - good overview. I agree with Arie and Carl regards inclusion of collaborative innovation and would add, co-creation as specific methodologies/ approaches to Open Innovation.
    An important distinction that has been mentionod marginally in the comments, is the need for differentiation between crowd-sourcing from the everyman/consumers and that of the professional creative industries.
    The danger of not doing so will lead to a further commoditisation of problem-solving professional services and thereby a de-valuing of their entire economic model and their professions. Hence they would all end up being lumped under crowd-labour - who will do it faster, cheaper - not necessarily better.
    Problem-solvers need to sit in collaborative innovation / co- creation in a professional branch of open-innovation.

    And on that note, without wishing to distract the story - there are two key subject ommissions. Both inter-related

    The first is Intellectual Property and the second remuneration systems.

    With the new business models need to come new reward systems & economic infrastructures, fir for a digital and open-innovation age.

    In that regard, you may wish to take a look at - an open-protection model that protects and respects the value of creativity whilst making it more accessible to collaborative innovation with corporates and co-creators.

    The consumer crowd may be satisfied with small rewards and PR recognition for their input to brands product and service development processes - but the professional crowd are not.

    Thereby open-innovation will never reach its full potential if arguably the most knowledgeable and capable parties are disenfranchised from the process or de-motivated due to inappropriate remuneration processes.

    Intellectual property is of vital importance to corporations and the professional creative community alike.

    A taxonomy cannot brush IP and economics under the carpet

    Hope that helps -

  • Guest SGiV Jul 24, 2011 11:29 pm GMT

    Hi Jeff:

    I agree with Carl that community should stay as a stand-alone group. It’s helpful to remove any sense of altruism from the concept and, instead, depend the definition on curating, structuring, and managing what the crowd is thinking, as opposed to knowing or doing. Carl’s examples of brand loyalty and advocacy are good examples of businesses sourcing the community. As Carl mentions, it is notably different from the other categories in that it doesn’t necessarily rely on the crowd to perform or complete any task. If effective, community will lead from crowd thinking to crowd knowing or doing. The aim of community or consensus building is discreet and powerful enough to be its own group.

    I would include all creative endeavors as a subset of distributed labor, and I would keep the term ‘distributed labor’ instead of ‘cloud labor’ because it is less ambiguous. Collective contributions are a subset of all labor contributions.

    I agree tools is a valuable addition and I would keep it as part of the taxonomy. As the matter surrounding crowdsourcing becomes more complex (granular, specific, what have you), the tools used will change to meet the challenge and provide for new ways to source; they may take very unexpected or novel turns.


  • Guest shaun abrahamson Jul 28, 2011 02:36 am GMT

    thanks jeff + carl.

    is there a benefit to thinking about what the crowd is doing? this might look something like use cases or user stories.

    + voting
    + reviewing
    + evaluating
    + translating
    + funding
    + qa'ing
    + editing
    + etc.

    main issue I see is the wide variety of actions which brings us back to how to group these actions. on grouping, might be worth looking at key org functions from like research, service, support, sales, marketing, etc.

    main motivation is to keep concepts accessible without leaning on another layer of concepts like open innovation or crowdfunding.

  • Richard Spiegel Richard Spiegel Aug 11, 2011 08:15 pm GMT

    On the debate over community building being included in the crowdsroucing taxonomy—while I’m happy Carl though of CrowdTogether as a example of a company providing services in this space, I’m going to have to disagree with Carl that community building should be included in the taxonomy. I believe a company like CrowdTogether is clearly in the tools category as we simply provide the platform, all be it one where we emphasize it’s abilities in community building due to it’s ability to facilitate activation, engagement and connections. The trouble with classifying community building as a form of crowdsourcing is that it’s more a byproduct then the primary output. So I’ll give an example, brand x wants to run a social marketing campaign built around a video contest. They set the rules—create a video about our product and why you love using it—if your video get the most votes you win. The creating of the video and the voting, those are crowdsourcing tasks, clearly and undeniably. So the key separation here is the level at which you’re viewing the system, looking at the tasks you see crowdsourcing. But look at the collection of tasks in the context of the campaign and you see social sharing, participation, comments, engagement, activation, etc = community building.

  • Guest Küstenpatent, Jan 17, 2012 07:01 am GMT

    I agree with Carl that community should stay as a stand-alone group. It’s helpful to remove any sense of altruism from the concept and, instead, depend the definition on curating, structuring, and managing what the crowd is thinking, as opposed to knowing or doing. Carl’s examples of brand loyalty and advocacy are good examples of businesses sourcing the community.

 Join or Login