RIP Cesura

Cesura, one of the more promising startups in the applications and performance monitoirng space shut down the week of June 5th, 2006. Cesura had a lot going for it including:

1. It was lead by Bob Fabbio, who was the founder of Tivoli, and who was highly skilled at raising enough venture capital enable a startup to achieve some real traction
2. It was able to raise $15m in a restart after a previous positioning failed in order to target the applications performance problem with web and Citrix hosted applications
3. It sported an impressive set of functionality covering some true measurement of end user experience, instrumentation of a broad set of infrastructure metrics and a highly differentiated set of analytics to pinpoint the areas in the infrastructure responsible for the end user or applications issue.

However, despite lots of money and significantly differentiated technology, Cesura was done in by two big mistakes. The first mistake was that the company did not maintain a focus upon a target market long enough to take market feedback and build something that truely met a market need. Cesura relaunched with a focus upon Citrix in the fall of 2005. By the middle of Q1/2006 the focus was upon health care applications and partners. From April of this year on, the focus shifted to value added hosting vendors and MSP's. Three target markets in three calendar quarters is enough thrashing to confuse customers, prospects, the press, analysts and even upper and middle management. A CEO with a pedigree, and $15m in the bank was apparently not enough to overcome the confusion caused by a strategy du jour operating plan.

The second big mistake was that things that the company throught were features of the product, were in fact viewed by prospects as impediments to adoption. For example:

1. The analysis engine and database were housed in an applicance, that had to be LAN connected to each switch that was a part of each unique subnet that hosted portions of the software that comprised the target applicaitons system. The feature here was the idea that the mangement product had its own network, and would not die if the production network died. The reality is that caused the sales team to have to get the network folks to approve the installation of something that was supposed to solve and applications and business level problem.
2. The product consisted of one and only one appliance. Each appliance could only support around 100 servers, and there were no cross appliance analytics or reporting. So, a product with a high end message and value proposition, could not scale up to high end enterprise applications systems.
3. The Citrix offering was unique in its ablity to measure true end user experience at the ICA client, but this only worked for a subset of the applications and transactions published through Citrix, and only worked on a subset of the Citrix client environments. Again, something that was truely unique in certain cases, did not support the breadth of environments to make cusotmers comfortable that this was an enterprise solution.

In short, the rules for startups targeting the user or applications performance space remain the same. Make sure that what you have to sell adds value to the monitoring products that customers have already invested in, and focus like a laser upon a set of prospects (otherwise known as a market segment), that have a common problem, and that can be accessed as a community or a market. It is amazing that after all of the money that has been invested in monitoirng startups, that companies still get funded and fail for these simple reasons.