New Product Development Using K-Means Clusting

I worked with a well-respected player and Gartner Cool Vendor in the Application Release and DevOps space. The company has a healthy revenue income, and an enterprise customer base in Banking and Retail, but faces competition from well-funded competition. Many of the company’s customers are also migrating non-core services to the cloud.

The company needed a new revenue stream and a new story to bring to market that would reinvigorate interest in the company, and drive interest in the core product. I led a hierarchical customer clustering exercise, where we asked our existing customer base to rank their existing business priorities. We used SPSS and K-means clustering.

From that work we were able to determine that there was a segment of The company’s existing customers that were willing to pay a higher unit cost for the ability to remove forecasting headaches for IBM WebSphere, faster test environments, and emergency failover. Currently they were paying a lot of money for peak capacity to IBM, but much of this capacity was going unused, for most of the time.

Following on from this exercise, we determined the engineering work that would be necessary for the company to provide IBM software on Azure and determined that we would be able to bring a new product to market within three months. We then worked with Amazon, Azure and IBM to bring this new product to market through the AWS and Azure Marketplaces, using Amazon Grid and Amazon Partner Network.

The new product, now has over 400 customers worldwide using the images on either Amazon Web Services and Azure. These include large Banks, global IT consultancy firms, software, Travel and Insurance companies. The customers range from independent companies to fortune 500 members. They have representations from the FTSE 100, Dow Jones Ind. Average, Nasdaq and many more.