Through machine learning, we can provide a level of automation that works for market researchers by sorting, categorizing and making insights accessible without the need for lengthy data processing.
If, like millions of others you watched the popular HBO series Westworld, you may be thinking about the implications of artificial intelligence (AI). While this fictional series takes AI to the extreme – where human-developed “hosts” take matters into their own hands – there are real-world applications within the AI field that are actually helping people do everyday things. Research firm MarketsandMarkets estimates that the overall AI market will be worth USD$16 billion by 2022, growing at a compound annual growth rate of 62.9% over the next six years.
One segment driving the growth of AI is machine learning – the ability of software to learn things. One typical example of machine learning is automatic photo tagging on Facebook. The algorithm learns from your manual tagging and remembers your face and the faces of your friends as you spend time inside the program. Taking this smart technology premise and applying it to the market research space promises substantial time (and frustration) savings.
Most researchers agree: the organization and processing of data is slow and tedious and can be quite complicated. Having a system that learns how you work with your data – your tweaks, rules, and exceptions – and then automatically applies your preferences each time can dramatically speed up processing. For example, after you import data and label your data points once, an intelligent system will automatically categorize data from other sources to fall into your predetermined categories.
When I show researchers how machine learning can speed up and simplify data processing, the consistent response I hear is “Whoa!!” Some have even said that the changes and updates I made in just a few seconds would take them up to three hours to code. What if you had a system that just came off the shelf knowing how to give you fast answers from your data? Drag a raw data file, such as an SPSS file, into the platform; let the system do its magic cleaning up labels and grouping common things; perform a few minor manual tweaks, and you have data that you can work with. When you add in a second data set, even from a different source, the system has learned from you and automatically categorizes the new data in the ways that make sense to you.
Applying the latest technologies to our jobs as market researchers is just good business. The faster we can start to get insights from our data, the more agile our company can become. Combining your knowledge and experience with a platform that is always learning can speed up every market research project exponentially.