By Carmel Dibner
Recently, Doug Clark of Douglas Dynamics, a premier manufacturer of snow removal equipment, co-presented a webinar with AMS. During the webinar, Doug shared his experience conducting his first machine learning study with our company.
One of Doug’s key points in the webinar was how pleased he was that the machine was able to find a mix of known and new insights. Doug stated that the combination of finding things he already knew and identifying new insights, “increases both [his] confidence in the validity but also the value” of machine learning. The known insights that the machine picked up on were reassuring, demonstrating to our client that the machine could correctly identify key themes related to snow removal. In turn, these confirmatory findings strengthened Doug’s confidence in the new insights he was able to collect, because the machine had demonstrated its accuracy.
While in any research study you may hear many of the insights you’ve heard before, hearing some of those insights again should make you feel confident that the new pearls of insights you gather are accurate and worthy of your consideration.
Also, keep in mind that the ratio of known insights to new insights is far less important than the potential impact of a new insight. Whether 10% of the insights are new to you or 50% of the insights are new to you, it’s important to remember that major innovations can be based on just one new customer insight that sparks disruptive innovation. The important thing is to be able to identify those new insights and to be confident that the new insights are correct. Our proven machine learning algorithm scans hundreds of thousands of pieces of data, ensuring all insights gathered are accurate and complete.
For everything you need to know about machine learning for insights, read our machine learning guide.