As researchers, we focus on asking the right questions, but we often forget why we’re asking them in the first place. We need to let the end goal drive the execution of our research so we can obtain accurate results. One way to do so is by automating market research. Put simply, automation is the replacement of human labor with machine labor, and it enhances all three corners of the project management triangle: speed, cost, and quality.
Automation has been increasing the speed of market research data collection and reducing human labor costs for decades. It all started when the first Computer-Assisted Telephone Interviewing surveys and auto dialers were used for telephone surveys in the 1980s. Now, modern research platforms have advanced light years by streamlining sampling, questionnaire design, data collection, and insights reporting.
The latest advancements in market research machine learning automate key elements that ensure quality insights, including:
- Identifying shared customer traits to target more precise sampling techniques
- Employing natural language processing techniques to conduct text analytics
- Determining problematic survey characteristics like unbalanced scales
The analysis and automated reporting of unstructured data in all forms has become critical to capturing high quality, actionable data and making informed, strategic decisions. As market researchers transition to automation, there are three essential lessons to consider to make a successful shift.
Find the Perfect Fit
While automation tools improve the speed and quality of market research data and insights (and thus, decrease the associated costs), it’s also important to avoid the pitfalls that may hinder those benefits — like purchasing the wrong automation tool.
Choosing the right product is essential in ensuring a successful transition to market research automation, and there are two important questions to consider when doing so:
- Does this tool have the right capabilities for my needs? Let’s say a team really only needs to capture open-ended text data, but the tool they buy is designed to capture all kinds of data, like videos and images. They’ve bought a product that doesn’t cater to their needs, and they may ultimately become frustrated with the cost of keeping up.
Researchers should begin the transition process to automation with a clear vision of what they need from a product. Vendors should also understand their customers’ needs, then design the tools they provide accordingly.
- Does this tool meet my team’s needs? Matching the right product to the right team (or product-team fit) is crucial. If an automation tool has too steep a learning curve, the cost benefit may be offset by additional training time needed for employees.
A team that is heavy on qualitative expertise may prioritize a tool that generates out-of-the-box statistical analyses. But a team of data scientists may want a tool that provides rigorous review of survey questions before they are fielded.
Democratize Insights for Transparency
Market research automation not only improves the speed, quality, and cost of project management, but it gives more people across an organization access to both data and insights — essentially democratizing information that may otherwise have been reserved for a select few. Platforms worth their salt not only collect and analyze data, but they transform it into visually-appealing dashboards that allow anyone in an organization to easily access and understand. When teams have organized data and insights available to them in real time, they can see the direct impact of certain actions and initiatives. Then they can make decisions that affect their individual teams based on complete data, rather than a slow and outdated report they receive weeks after the fact.
Automated insights also encourage teams to focus on specific key metrics. If everyone in an organization has access to a dashboard that shows real-time insights and recommended actions on customer experience, then customer experience will remain top-of-mind for the entire organization.
It’s important, however, to keep in mind certain risks of democratized information. Oftentimes, team members or leaders can draw incorrect conclusions based on data because they don’t understand inherent biases in the machination system. If there’s a mistake or nuance in the data, misleading information can move through an organization quickly.
The best automation tools are designed with mechanisms that allow humans to review and adjust for errors — before they ever make it out to the broader organization. For example, high quality data collection dashboards or statistical analysis tools will allow teams to monitor survey results and quickly adjust sample designs when respondent screening criteria are not yielding the right research participants. Ideally, those tools also provide visualizations that are easy to understand for all types of users in an organization.
Improve Qualitative Research
Recent research presented by RTI International’s Amanda Smith found that automated coding of open-ended data was successful in simplifying insights for use cases involving large amounts of data on one topic — like longitudinal tracking surveys or customer feedback data. While humans still have to define the qualitative codes, a machine learning tool excels at taking existing codes and applying them in cases where there are a lot of observations. The risks of using automation for qualitative research, however, lie squarely with the limitations of machine learning.
Most market researchers know, for example, that automated analysis of social media runs the risk of a machine misunderstanding sarcasm as genuine sentiment. But we will start to see similar issues with more complex technologies, too. As we witness more automation of biometric and neurological, non-conscious responses to stimuli, automated analysis of biometrics runs the risk of incorrectly interpreting emotions like fear as excitement, or vice versa.
As researchers keep these concepts and caveats in mind, they’ll learn how to best use automation to their advantage as they transition to machine learning. With the right mindset, automation will initiate an entirely new era of democratized, transparent, speedy, and qualitative information for companies looking to innovate across the globe.
