What is crowdsourcing?

Crowdsourcing is the practice of getting work, ideas, content, votes, or money from a large group of people rather than from staff or paid suppliers, usually through an open call online, often run as a contest, a vote, or a structured submission process with a brief.

What is crowdsourcing?

Crowdsourcing is the move of taking a job that would normally sit with staff or a paid agency, and opening it up to a large group of people online instead. The job can be almost anything: write a slogan, design a T-shirt, name a product, suggest a flavour, build a Lego set, label an image, edit a Wikipedia entry, fund a new gadget on Kickstarter. The thread that runs through all of it is the same: a brief goes out, a crowd responds, and the organiser picks, aggregates, or funds the contributions that came back.

The word itself is fairly recent. The writer Jeff Howe coined it in a 2006 Wired article with his then-editor Mark Robinson, and the Wikipedia entry for crowdsourcing traces the path from there into product design, stock photography, software testing, scientific research, and the rest of the wider field. Merriam-Webster picked up the verb form not long after and defines crowdsource as the practice of obtaining needed services, ideas, or content by soliciting contributions from a large group of people, especially from the online community, rather than from traditional employees or suppliers.

For social media teams the working definition is narrower than the one in the dictionary. Most of the time, crowdsourcing on social means a campaign with a clear brief and a deadline: submit a photo of your dog with the product, vote on the next flavour, name the new colour, tell us your worst Monday in one sentence. The brand collects, picks, and rewards. The audience gets to be part of the thing rather than the target of it.

The main types of crowdsourcing

Most useful guides split crowdsourcing into four shapes, which cover the vast majority of what brands and platforms actually run. Sprout Social's glossary entry on crowdsourcing lands on a similar split, and the categories below match the ones marketers tend to recognise.

Most brands only ever use the first three. The fourth and fifth (crowdfunding and microtasking) tend to belong to a product team or an operations team rather than a social one, and they show up in the conversation when a launch is on Kickstarter or a model is being trained, not when a campaign is being planned for next quarter.

Crowdsourcing vs UGC vs crowdfunding

These three get used interchangeably in meetings and they should not be. The short version of the split:

The practical test in a planning meeting is the brief. If there is a clear brief and a way to submit, the project is crowdsourcing. If there is a hashtag and a hope that people will post, the project is a UGC campaign. If the entry mechanism is a payment page, the project is crowdfunding.

Famous crowdsourcing examples

A short tour of the campaigns most marketers cite when they explain what good crowdsourcing looks like. The pattern across all of them is a tight brief, a clear reward, and an editorial filter at the end so the worst submissions never make production.

The common factor in the campaigns that worked: the brand decided in advance what range of answers it would accept, and the audience chose inside that range. The factor in the campaigns that did not, covered in the next section, is usually the opposite.

When crowdsourcing works on social media

Crowdsourcing on social does its best work when the brand wants to do three things at once: produce something useful, find out what the audience actually thinks, and earn the engagement that comes from people having a stake in the result. The campaigns that hit all three share a few ingredients.

The wider lesson from twenty years of running these campaigns: the audience is happy to do the work, as long as the work feels like a contribution rather than a draft of an unpaid agency gig. Most of the brands that get crowdsourcing right end up running the same campaign annually because the audience explicitly asks for it back.

When crowdsourcing backfires

Two campaigns get cited every time the subject comes up, because both turned into textbook examples of the same mistake: putting an open-ended naming vote in front of the public with no veto path and no editorial filter, and being surprised when the internet did what the internet does.

Mountain Dew's 2012 Dub the Dew campaign asked the public to name a new green-apple flavour and put the submissions on a leaderboard before anyone graded them. The leader at the time the page was pulled was a name that cannot be repeated on the brand's own glossary site, and the campaign was killed inside a few days. The brand caught the problem before anything went to market, but the case study now sits in every "do not crowdsource a name" deck on the internet.

The 2016 NERC poll covered in the Boaty McBoatface Wikipedia article is the friendlier version of the same mistake. The UK's Natural Environment Research Council asked the public to suggest names for a £200 million polar research ship, the entry Boaty McBoatface won with 124,109 votes, and the council eventually overrode the vote and named the ship the RRS Sir David Attenborough, with the name Boaty McBoatface going to the autonomous submersible instead. The British Antarctic Survey's announcement is the official record of the decision. Nobody got hurt, the ship sails, and the case became a permanent lesson on what happens when a brand puts a creative vote on the internet without a way to step in.

The pattern is consistent. Crowdsourcing fails when the brand surrendered the editorial seat before the campaign started, gave the audience an open-ended creative job with no shortlist, and trusted the result to be on-brand because the people voting were a self-selected mix of customers, critics, and the wider internet. The audience is rarely the problem; the brief almost always is.

How to run a crowdsourcing campaign on social

A practical workflow that fits inside one quarter for a small team, and scales up to a major campaign without changing shape.

1. Decide what the crowd actually owns. Content, idea, vote, or money. The four shapes have different mechanics and the rest of the plan changes once this is settled. Most teams that start with "let's do something crowdsourced" without picking one end up running a UGC campaign with extra steps.
2. Write the brief in one sentence. If the brief cannot fit in a caption, the audience will not read it. "Submit your worst Monday in one sentence and we will pick five to feature next week" is a brief. "Share your stories about how our brand helps you live your best life" is not.
3. Pick the submission mechanism the audience already uses. Replies, tagged posts with a hashtag, a duet or stitch on TikTok, an Instagram Stories sticker, a DM to the brand account. Anything off-platform halves the entry rate at best.
4. Set the editorial filter in writing. A shortlist, a final pick, a moderation policy, and the right to remove entries that break the rules. State the filter in the rules at the start. The audience will accept a filter they were warned about; the same audience will burn the brand down if the filter shows up after the vote.
5. Decide the reward before the campaign goes live. Cash, royalties, a feature post, a free product, public credit, a place in the next campaign. The reward sets the tone of the whole brief and tells the audience what kind of campaign this actually is.
6. Plan the publish calendar around the submissions. The campaign is not the call for entries; it is the trail of content the entries make possible. The shortlist post, the finalist explainers, the announcement, the winner interview. Roughly three to six posts of follow-on content come out of one well-run campaign.
7. Credit everyone who contributed in public. Tag every featured entrant, name the winner, link the designer's account. The campaign that follows next year will run twice as well because the audience saw the last one treat its contributors right.

Common crowdsourcing mistakes

The mistakes below show up in almost every crowdsourcing post-mortem. Most of them can be fixed by adding one paragraph to the brief.

1. Asking the public to name a product with no shortlist. Dub the Dew, Boaty McBoatface, every copycat poll that came after. Naming votes only work when the audience picks from a shortlist the team has already screened.
2. No moderation budget. A submission page without a person clearing entries turns into a leaderboard of the rudest possible answers inside a weekend. Build moderation into the campaign as a line item from day one.
3. Counting on free labour from an audience the brand has not earned yet. Crowdsourcing campaigns work best from brands the audience already likes. A small brand asking strangers for free creative work tends to get tumbleweed; the same brand asking its existing customers for a quick contribution tends to get a flood.
4. Skipping the rights and licensing terms. Submissions that end up in production need clear rights assignment, and contributors need to know up front that their entry could become a product, a campaign asset, or a paid ad. Leaving the licence ambiguous is how crowdsourcing campaigns end up in court a year later.
5. Treating the campaign as the goal instead of the start. A good crowdsourcing run produces an audience asset, a product asset, and a content asset. The brands that stop at the campaign itself miss most of the value. The brands that build the next quarter's content pillars on top of the submissions get a return measured in months rather than days.
6. Forgetting to credit contributors in public. The fastest way to make sure the next campaign flops. A public tag, a named credit, a link back to the contributor's account. Costs nothing, pays back the rest of the year.
7. Ignoring the long tail of submissions that did not win. Most crowdsourcing campaigns get 90% of their value from the long tail rather than the winner: the patterns in what people submitted, the themes that came up repeatedly, the unprompted comments under the campaign posts. Reading the long tail is the cheapest qualitative research the brand will run all year.

For the wider planning context this sits inside, the campaign entry covers how a crowdsourcing run fits next to the rest of the quarter, and the brand awareness entry covers the longer-term effect campaigns like Lay's Do Us a Flavor have on how easily the brand is recognised in the first place.