How to overcome biases in Crowdsourcing Research?
Online communities are now being conducted by many brands and categories in order to keep in touch with their customers.
- New mums are invited to branded mother’s clubs so they can receive parenting tips and product advice.
- Travellers subscribe to travel communities with cruise companies for special incentives.
- Bakers keep in touch with branded baking communities providing freebies, events and new product trials.
The most actively engaged communities tend to be those within high involvement categories. Like baby products, healthcare, alcoholic beverages and travel. It varies depending upon how the online community itself is run, community engagement strategies and the types of community members themselves.
Crowdsourcing has now become a routine way for customers and companies to engage.
Companies can get simple questions answered quickly. Test raw ideas and invite community members to share their own product and brand experiences quickly and cost effectively.
They can engage with their customers continuously with often a larger number of individuals who have opted in to do research.
There is also the ability to share and be transparent as a business. Asking customers to help co-create innovation and for the business to feel truly customer centric. For the customers, the feeling that the company is interested in them. Providing them with free product or benefits may alone be enough to engage.
However, there are many considerations to be made before jumping in.
Whilst companies are allocating more and more marketing and research budgets to crowdsourced market research and branded communities that they run themselves.
Crowdsourced communities alone are not enough to ensure a business has a clear on the customer, the competitor or even their own brand. The community itself can also be a hot pot of inherent biases and poor research design and analysis.
Take for example biases in the data. The kinds of biases you might experience could include ‘mateship’ bias and a tendency towards ‘over agreeability’.
- The mere fact someone wants your opinion might mean the crowd is over compensatory.
Social desirability bias is another – where the crowd responds in the way that puts them in the best light, and even sponsor bias.
- When the crowd recognises they are in a branded community and are more inclined to be overly positive toward the brand. Imagine a Harley Davidson community.
- Alternatively the crowd might want to ‘payback’ to a brand they have concerns about. Biases within the community data need to be understood and managed so that these do not provide data that can inevitably impair decision making.
There are also biases in the system of collecting data through crowdsourcing research in branded communities such as confirmation bias.
- Looking for answers to support the businesses viewpoints or hypothesis.
- Asking leading questions or priming question-responses through a question-order bias. In other words, leading the crowd down the garden path.
- The community could be bored and simply auto-pilot answers. Hinging results on a positive response or set of assumptions can also be problematic as can sample size, independence of the analysts and the nature of the sample itself.
Biases in crowdsourcing research can be managed. That is where an agency well versed in crowdsourcing research techniques and online communities can help.
An independent agency asking focused questions. Employing quality research practices. Whilst remaining aware and focused on sources of biases will be able to help a business run a more efficient and effective online research community.
Tips for Better Community Engagement Research and Crowdsourcing
- Invest in amazing technology that is simple for the business and for the customer
- Ensure your crowdsourcing or community team has access to a strategist and an independent community engagement expert
- Quality research design is critical. Beyond the platform provider, an independent market research agency can help, ensuring biases are accounted for and question type and flow is efficient and effective.
- Crowdsourced research data should have critical appraisal so that key shifts in customer needs; or any external forces impacting research data patterns are not missed.
- Ensure you have good participants. Only 9% will contribute or interact, an additional 1% will be obsessed with over interaction and the remainder will just lurk and read – just like Facebook! A good set of screener questions might help with this.
- Online communities need to be well engaged. They will want to participate in pop quizzes and product placements. They will want to feel like they are part of the solution, so ensure they are well engaged and rewarded.
- Remember that a branded community will have an inherent bias. Important to ensure you are also talking to your prospects, competitor buyers and non-buyers in order to stretch beyond internal naval gazing
Takeaways
Biases in crowdsourcing research are to be expected, however, they can be overcome. Engaging an independent agency with expertise can help you get more from your community engagement, data and analysis.