Prelytic Software was a seed stage start-up which had developed a Predictive Analytic application targeting IT and Operations for Fortune 500 companies. The system was based on splines, specifically Multivariate Adaptive Regressive Splines (commonly referred to as a MARS algorithm) and was designed to predict and adapt IT resources towards avoiding performance degradation and optimizing cost efficiencies.
I was responsible for Marketing Communications, Product Management/Marketing, and Strategic Planning.
Given that we were seed stage and had an alpha product for the first year, our MarCom targeted the analyst community, specifically those focusing on IT Operations and Predictive Analytics. The traction with this group was fairly strong, up until this point no company had applied predictive analytics towards the overall IT performance of a Fortune 500 company (lots of vendors providing real-time analytics, none were offering a predictive application). We were doing something new, innovative, and interesting enough that we were tracked by a number of analysts firms.
Given the stage of the company and our tactical requirements, most of the web site, collateral and associated support material was oriented towards potential investors. The two companies we were working with (General Motors and The Gap) were too early in the cycle to be press references.
Product Management and Marketing
The product, called ProfilePro, integrated Topology Mapping, a Scenario Based Planning Engine, Tuned-Cube View Analysis, and a High-Dimension Curve Fit, front-ended by an executive dashboard. On an operational level, the product integrated in three areas, 1) IT Asset Management Systems, 2) Network, System and Service Management, and 3) IT Portfolio and Project Inventory Systems. This was a complex and sophisticated application best suited toward large IT-centric companies.
Product management and marketing focused on building out functional and design specifications, as well as building out the web site and associated collateral materials. We also spent a significant amount of time and effort working with potential end-users to understand information requirements for an executive-level dashboard. For IT and Operations, each element in the network had it’s own predictive agent, which delivered a constant stream of updates into a dashboard that monitored a wide range of KPIs (Key Performance Indictors) for out-of-variance events, then projected forward using the MARS-based predictive algorithm. Prediction timeframes could be projected out to twelve months, with reasonably high rates of accuracy.
Towards the latter part of the product’s development cycle, we began working on the opposite end of applying predictive analytics; using the algorithm to predict system failure at the individual processor level, a requirement driven by semiconductor companies.
Strategic planning focused on fundraising and providing alignment from product integration and positioning in an ecosystem that included Business Intelligence/ Analytics, ERP/BPM, and IT Systems Management vendors.
We contacted nearly four hundred VCs over a period of three months, and eventually scheduled face-to-face meetings with over seventy in the bay area, averaging two to three meetings per day. We eventually received a term sheet from ICCP Ventures.