While New Relic is only 2 years old, I have been actually been focused on web application performance management for over 12 years. In January 1998, I founded a company called Wily Technology, which was the first to offer a production performance management tool for web applications implemented in Java. CA eventually acquired Wily in 2006 and Wily’s products continue to generate significant revenues for CA.
After leaving CA, I had the opportunity to relax a bit, spend some time with my family, and think about what had worked well and what could have been done better during my Wily days. A lot of the results from that analysis have gone into New Relic. I will take you through some of my observations in this and later posts.
Traditional Enterprise IT Management Software is Poorly Designed, Hard to Implement, and Expensive to Own.
Application support teams don’t have time, energy or money to manage the management system, but that’s what they end up spending a lot of time doing. Most IT management software comes with an inherent flaw – It requires lots of configuration and customization to make it function in a particular environment. Configuring and customizing the management tool for each application does not scale for a large number of applications, and it really doesn’t scale well if those applications are changing frequently or are moving all over a virtualized or cloud environment.
When I founded New Relic, I placed a very high priority on building software that just works. We focus relentlessly in reducing or eliminating the time investment to fully deploy our app management solution in a production environment. We want our customers to derive full value from RPM in minutes, not months.
Enterprise Software Companies are Very Inefficient, and the Customer Bears the Cost.
When I first discovered the average selling prices for enterprise software products, my naïve conclusion was that this cost was driven by rocket-science innovation that mere mortals were incapable of creating, and this groundbreaking technology justified prices in the millions of dollars. But the truth is that the high prices and the poor design of the products are directly related. At Wily we chose to sell using the traditional direct selling model and we ended up with more than half the company involved in selling and supporting the sales effort. We were not alone. Most software companies still use this model. The end result was that about 60% of the every dollar we spent went into sales and marketing related expenses, and less than 20% went to R&D. To grow our business, we had to add more sales people and more people to support the salespeople. What that also meant was that there were fewer and fewer dollars left to invest in the product. Customers ended up getting less value (for a lot more money.) In the New Relic model, about 60% of our operating expenses go to R&D. This means better product, faster support, shorter dev cycles and happier customers. When Wily was sold to CA we had about 500 customers, 260 employees, and 17 engineers assigned to write product code. At New Relic after two years we have 3500 customers, 24 employees, and over half our employees are engineers assigned to write product code! You get my point. I plan to dig deeper into this topic in a future post.
|Wily at Acquisition
|New Relic After 2 Years
|Age||8 Years||2 Years|
|Number of paying customers||Approx. 500||Over 1000|
|Number of production-deployed customers||Less than 500||More than 3500|
|Number of employees||260||24|
|Number of development engineers||17||13|
|Number of sales reps||30+ Direct Field, 15+ Inside||1 Inside|
|Number of sales engineers||30+ Direct Field, 10+ Inside||1 Inside|
|List price to manage 2 servers, 4 CPUs each||Negotiated but more than $50,000; plus 20% annual maintenance||$285/month (?)|
|TCO considerations||Servers, upgrades, software maintenance, patches, etc.||None. (SaaS model)|
I don’t claim that we are getting everything right at New Relic. But we are on a mission to create a completely new, more open, and customer friendly IT management software business.
I will discuss some of my other lessons learned in my next posting. I would love to hear your feedback below.