If you are anything like me, you live in market research because it gives those intangible insights about human experiences – the kernels, the ‘aha!’ moments that can make a difference in how a product is developed, a brand perceived, a connection with the consumer that alters and sharpens our solutions to pain points in everyday experiences that move it from solved to delighted. But in many cases for large enterprises and agencies, that means using tools to discern signal from noise in hearing the voice of the customer. You can use artificial intelligence, specifically text analytics to pivot product and marketing strategies to the needs of consumers especially when you get large volumes of feedback that comes with text analytics, machine learning, and sentiment analysis.
So, let’s start with text analytics. If you are newer to text analytics solutions for market research, it is a data analysis tool to automate and bring insight via artificial intelligence technologies into trends, sentiment, and tactical areas of focus and improvement for open-ends in surveys and market research. Text Analytics can do in moments what has often taken weeks and months to do manually – making sense of thousands of entries of open-ended comments from consumers, customers, and prospects into something meaningful and actionable. Beyond just general sentiment (NPS – would or wouldn’t recommend our company …so 2018) into who, where, why of what to fix in the consumer, product, or brand experience. Text analytics solutions are getting more sophisticated at not only detecting where experience is lacking (billing, customer care, product mobile experience for example) but also in catching customer issues early on before they become a trending issue on Twitter. With the consumers’ always-on connection to the brand, using technologies that keep us in lockstep with our customers is critical!
So when choosing a text analytics partner, here’s five key questions to ask your text analytics partner to ensure the solution they offer fits your needs specifically in making sense of market research
What is your experience working with open-ended consumer statements in market research?
If you are working with a partner with a text analytics solution, it’s critical that the text analytics solution is created from and for market research. This means that the text analytics solution detects customer pain points (product, billing, onboarding, etc) and can separate sentiment about specific areas of the business that impact the consumer. Just like a race car, the algorithms under the hood of the solution need to be tuned to making sense of open-ended feedback – otherwise, you will not get the insights you need.
How does your text analytics solution interpret sentiment for international brands?
If you are working with or for global brands – context is everything. The text analytics solution should understand local languages, culture, and nuance so the results speak to a global footprint. The results globally should reflect these nuances. Ask the vendor to demo in languages and regions to ensure they are accounting for language and cultural differences.
How do you tune your engine? Can I tune the insights for my own research?
Because so much of interpreting research means a solid understanding of the brand or product audience, there should be some basic tuning of words and phrases critical to your audience, brand, and product. So there should be the ability to tune the algorithms to what makes sense in the context of your research – and – they should be continually tuning the text analytics solution to ensure context, language, nuance, and natural language is analyzed correctly as a product in and of itself.
What about reporting?
Make sure your solution can scale beyond NPS or CSAT open ends. You should be able to get reports on multiple open-ends across languages, longitudinal measurements, and then be able to filter into insights by audience profile, over time, or in the subject area. Make sure you can compare and contrast different brands, products, and business units. Anticipate what your client or business stakeholder would ask to ensure it’s available. You should not only be able to have dashboards for measurement, but also drill-down capabilities to better understand insights across the experience to make recommendations easier to see and act on.
Can I test drive with my own data?
This is perhaps the most critical. If the on-boarding of your own data takes weeks or months – that’s a sure sign the text analytics solution is probably not as robust as it should be or tuned for a different use case than what you need as a market researcher. If you have a good solution, you should be able to immediately see time savings to what you do every day – insights across experiences, languages, and sentiment to allow you to recommend solutions to your business stakeholders. By applying text analytics to open ends – you should not only be able to preserve individual insights but save days of work by analyzing themes, sentiment, and recommended action to allow your research to be one step ahead of the consumer. So ask for a test drive. If you don’t see time savings, move on.
Choosing a text analytics partner isn’t easy and there are certainly no silver bullets. But whomever you choose ensure your recommended insights aren’t lost in the solution. While text analytics solutions will help streamline analysis of open-ends, they don’t quite yet have the fine-tuning to what researchers do every day – find the nuance and emotive connection that move insights from just a problem solved to human delighted in the new experience.