Here's a little inside baseball: In the world of machine learning, it is of paramount importance to maintain the accuracy and reliability of training data (in our case, our data set of over 25,000 photos). However, as data evolves and changes, this accuracy can suffer from a phenomenon known as data drift, where the input data shifts away from the data we had originally used to train. Basically, what FinishMatch is being asked to match doesn't match what it was trained with.
We see this with design trends and preferences, as well. What was popular yesterday may not be popular tomorrow.
To prevent data drift, we collect photos that users upload and match. This helps us to retrain our models as necessary to account for any specific data changing. This also allows to see exactly what design aesthetics homeowners across the market are searching for.
Stay Ahead Of Customer Needs
Imagine having insights into exactly what your potential customers are seeking, enabling you to anticipate and provide the perfect product even before they realize they need it. That is the exact information we put in our monthly FinishMatch Report.
Every month your organization will receive the top photos uploaded and matched on your website, as well as the top photos uploaded to websites across the market. From identifying new market opportunities to refining your marketing creative, this data can help you stay ahead of the curve.