Measuring Performance

I recently had an interesting email exchange that I decided to blog after being asked how I measure the performance of a company. I've x'ed out the names, but fundamentally believe that these measurements apply to any enterprise software company.

--- snip ----
Subject: company objectives

Here's some refinement of the earlier email I sent with suggestions for measuring the performance of the company:

- Customer sat can be made a portion of commission for the sales person, but would be hard now as we do not have a benchmark. Having said that, we can make 25% of a commission payable when the customer goes into production. This may make for unhappy sales people, so we could make it a positive motivator by making a spif for go-lives. This also makes a sales person sensitive to selling what can actually get done. We can also track the number of customers going live and the number of days that it takes them to get live (obviously accounting for different size customers).

- The DSO discussion has been ongoing, here's some feedback I got from a CFO of an enterprise software company currently doing $11m a year in revenue with moderate growth. They are selling to the same customers that xxx is:

"I would say that DSOs should run somewhere between 40 and 70 days, unless the company agrees to some pretty bad payment terms. We are targeting DSOs of no more than 55 days and achieved 50 days in Q1 this year. One word of caution. The DSOs could be skewed to look better by simply increasing the bad debt reserve since this reduces the accounts receivable balance on the face of the balance sheet, so the DSO calculation should be reviewed in concert with the bad debt expense for the period under review. Typically, the bad debt expense is buried in the G&A expense line."

- Additional financial performance measures could be Days Payable Outstanding with a goal of 45-60 days.

- I'm pretty keen on putting some revenue-per-employee targets in place. I spoke with another CFO at a mid-sized enterprise software company about their targets, here's his feedback:

"Without knowing specifics, I can't suggest an absolute target for revenue per employee, but assuming a 12% target operating margin, and an assumed truly fully burdened cost per employee of $300,000 per year, then targeted revenue per employee per quarter needs to be at least $85k."

We are currently at $130k per employee per YEAR, so it's evident on it's face that this needs to go up significantly.

- Bugs. If we accept that we have a history of excessive product issues, then we should be in agreement that we need to monitor our progress in this area. A couple of measures could be the number of bugs reported to customer support each month, together with a caculation of the average number of days to resolution of a bug, with a goal of reducing these amounts over time. Measuring the cumulative number of active customers, (ie those renewing their annual maintenance), and setting targets for the increase in this number would also be a measure of customer satisfaction. Conversely, a fall off in customers renewing their maintenance would indicate potential problems that need to be addressed. Through closer monitoring of the bugs we would also be better able to spot trends in application, environment, and performance issues, and then adjust our development resources accordingly, as well as set expectations in the sales cycle.

- customer acquisition targets are still a concern. Not only should we continue to push for pick-a-number of referenceable customers, we should consider a target of x number of overall new customers signed. CEO should have a target of reference customers that he is measured on. In addition to reference targets, there are a number of industry initiatives that could be strategic to us, e.g. RFID, and it's imperative to get a foothold in these markets. While we do not have strong emphasis in pharma or CPG (the bulk of the RFID pilots) we do know that it's somewhat mature in automotive where we do have a good customer base.

--- end snip ----