Crowdsourcing – Understanding the Pitfalls To Harness The Rewards
I love crowdsourcing. Actually, I love the ideal of crowdsourcing. What is not appealing about harnessing a collective intelligence, collaborating and drawing on a diverse multitude? It is one of the best uses is of social tools and our ability to connect from anywhere. But crowdsourcing is neither a magic bullet nor a stand-alone solution. There are some very limiting factors to crowdsourcing, which we should all take into account. Keeping these caveats at the front of our mind will help design a crowdsourcing project that minimizes the limitations, ensures we are monitoring and adjusts effectively during the campaign. More than that, it will contribute to the analysis and concluding after the campaign has ended.
This is an expansive and diverse topic. A single blog post cannot possibly do it justice because there are many other concerns, some of which relate to specific industries like design, translation and so on. So, instead of focusing on specific concerns, I’ve addressed some overarching elements that need consideration when using crowdsourcing. Please feel free to add to my thoughts. (What would a blog about crowdsourcing be without a request for people to participate?)
1. Preparing your Strategy –
The way you use crowdsourcing, and your preparation strategy, will depend on your objectives, resources and organizational philosophy. What might be an advantage of crowdsourcing to a development agency, may end up being a disadvantage to a firm needing to crowdsource the solution to a complex, highly specific problem.
2. Hitting the Right Audience –
There is a good reason why many of the successful crowdsourcing examples have been concentrated on events and disasters. Applying a keyword search across a multitude of free-to-access online communities and sources draws a critical mass of organic or unsolicited content and reporting. It also allows the data gatherers an opportunity to announce their intention to crowdsource within those communities and install an organized system of harnessing data such as SMS shortcodes, Twitter hashtags and specialist platforms such as Ushahidi. But what happens when you‘re crowdsourcing a specific task, or trying to solve a complex problem? How can you be sure you are hitting the right audience? How do you know people are applying themselves correctly? Does an informal setting produce informal results? What is the right level of insensitive?
Being transparent, offering clear and concise instruction and empowering guidelines will help. In the spirit of crowdsourcing, being open with what you are trying to achieve and sharing your existing data will improve effectiveness and the quality of responses.
3. Verification or Credibility –
Across all crowdsourcing – from citizen appeals, to breaking news reporting and business outsourcing – how do you know your answers are coming from a reliable source? Sure you can build a contextual picture around peoples’ public profiles, but at what stage does that become unsustainable? What about people who build anonymous profiles? On a similar subject, how do you get people to part with sensitive information if they feel they can be tracked and become a target? How do you guard against becoming the target of pressure groups whom are determined to force their own agenda?
We have seen with sentiment analysis, the relative unreliability of filter systems based on algorithms. Many people take a sentiment analysis as a rough guide, versus an absolute truth. At what stage does censoring responses, or excluding data from some audiences, skew the findings and corrupt the data?
4. Making Sense Of The Data –
Gathering enough data is one problem, but the other side of that is that once you have gathered all the data, how do you make sense of it? What are the costs involved? (Using software and filtering can help if your objective is to gain an insight into trends, get a take on mood, or harness types on incidents around a location). When you are after specific ideas that need to be derived from people with specific skillsets, sorting through the reams of data, you run the risk of taking more time and yielding inferior results than if you hired a qualified professional.
5. We Are Listening But We Don’t Like Your Answers –
What may start out as a PR exercise, or a marketing campaign, can run the risk of backfiring. What happens if you just got it wrong and you do not get the response(s) you were expecting? Is it feasible to not act on what is essentially a very public and visible show of opinion or thought?
6. Open Data –
Questions can arise over the ownership of the data. One can argue that participating on a public forum makes the data free to use, but that can work both ways. Don’t assume an organization’s competitors aren’t keeping a watchful eye over proceedings. Being clear from the outset, whether on a public or closed network would help but may not halt all challenges in this regard.
7. I’ve Participated, Now What? –
Many cite the reason they want to use crowdsourcing is to make their audiences feel heard, feel engaged and gain a sense of ownership to a brand, cause or organization. But that’s not the end of it. Without the appropriate follow up, reporting or visibility of a tangible report, “result” or related action, may feel used, their efforts unappreciated, or worse that, the whole exercise was just a stunt.
What are some of the challenges you have come across with Crowdsourcing? What are your Crowdsourcing best practices?
Photo by Left Hand