There's no end to the things you can measure. Studies can correlate employee punctuality to manufacturing output, press release flamboyance to investor turnover, and frequency of atdesk massages (remember those?) to grammatical errors per email. So let's start simple and start with the Four R's: ROI, ROA, ROE, and ORI.

Return on Investment (ROI)

I first heard the words return on investment when I opened an account at Palo Alto Savings with a $25 birthday fortune received from my grandmother. I'd never seen so much money before and felt the bank was the best place to keep it. To my mind, it was just like my piggy bank at home, except it would be much tougher to sneak out a nickel for a candy bar without my parents knowing about it.

The woman behind the counter explained that I could get a better ROI if I was willing to let the bank hang onto my money for a longer time. Do you want your money anytime? You'll get a smaller return on investment. Let the bank play with it a while longer; they'll give you more back. That made sense, even to a 7-year-old.

Return on Assets (ROA)

Return on assets didn't make sense to me until I was 10 years old and set up a lemonade stand on the street corner. I grabbed an old card table and a couple of lawn chairs. I used the honestto- goodness Kool-Aid pitcher with the smiley face on it. I picked lemons from the backyard and bought (no freebies on this one) 1 pound of sugar. Total investment: 53 cents. So at the end of the day, I was very proud of the $1.22 I had in my hand. That would buy a delightfully unhealthy number of candy bars. I was delighted.

When Rich from across the street (he was 13 years of age) offered to buy my stand the next day for $2, I thought I'd struck it, well, rich. I'd have to work more than 2 days in a row for that! All on a meager investment of 53 cents. That seemed like a great return. Fortunately, Doug, my older brother, was on hand to explain that Rich would then be allowed to keep the table, the chairs, and our favorite Kool-Aid pitcher. The furniture and the pitcher were my assets. I didn't have to pay for them. I didn't have to buy them every day. But they would belong to Rich after I took his $2.

Not only that, my natural-marketing-whiz brother explained, I had built up some recognition in the neighborhood. People who had seen me the day before would remember me and come back the next day. My very first lesson in brand equity. By the time Doug had explained that I was likely to earn even more this day-being Saturday-Rich had lost interest and hustled up a whiffle ball game down the street. He was not the consummate entrepreneur.

Return on Equity (ROE)

The ROE lesson was short and sweet. I had borrowed the 53 cents from my father, making him a primary shareholder in my little enterprise. On Sunday evening, when he asked for payment in full plus the 5 cents interest we had agreed to, I had to admit that all of the equity had found its way down to the corner store with the large candy counter. Dad was not pleased that I had forgotten about him and explained that the nickel was his return on equity. He had not put any energy into the enterprise, only money. His pre-agreed ROE was now AWOL. Fortunately, Dad consented to payment in unconsumed Nestle Crunch bars.

Online Relationship Improvement (ORI)

Thirty-six years later, when I was sitting as a member of the board of directors of the Swiss consulting firm Omaco (www.omaco.com), the term online relationship management, or ORI, originated. The firm was working with a large publisher looking to implement a customer relationship management (CRM) system. The publisher wanted to know how much money the company was going to see returned from its investment. We realized that the answer was not to be clearly tabulated in Swiss francs.

As a consulting firm, we wouldn't be able to help until the publisher understood the value was to be realized in terms of goodwill, stronger brand affinity, and higher customer retention. There was also the value of ongoing direct market research to consider. So we created ORI and began to ponder ways to measure it for this particular client. As you can see, once you get past measuring money in and money out, things get esoteric quickly. So let's start with the simple metrics. The days of depreciating a computer to zero within months or expecting double-your-money-back within weeks are over. I'm also going to stick with the idea that if you get a return on your technology investment within a couple of years, you're doing well.

Measuring the Value of Technology

If you run your own servers, you're going to have to buy them, house them, feed them, and keep caretakers within reach 24 hours a day. If you outsource the care and feeding of your Web server, you've got expenses. Either way, you're going to be playing with the total cost of ownership. Sure, there's hardware, software, personnel, and electricity to buy. But there's more: lost productivity when the server goes down; converting to the new systems when that server doesn't come up again; training your people to care and feed the new systems. In his white paper entitled "Calculating ROI for Business Intelligence Projects" (Base Consulting Group, December 12, 2000, www.baseconsulting.com), Jonathan Wu spelled out how to estimate the cost of implementing a business intelligence project. He felt that cost could be classified into one of three categories: hardware, software, and labor. They are defined as follows:

• Hardware costs relate to the information system devices that are required by the project such as server system(s), client systems, and network communication.

• Software costs include the software applications required by the project such as the business intelligence (BI) application and the Relational Database Management System (RDBMS) license fees.

• Labor costs pertain to the use of internal and external resources dedicated to the project. Individuals working on the BI project fill the following roles: project manager, business analyst(s), BI application specialist(s), database administrator, systems administrator, and trainer(s).

The total cost of implementing a BI project can be divided into two categories-initial and recurring:

• Initial costs are those costs an organization incurs for the BI project once. These costs include hardware costs for the project, software license fees, and labor costs to configure and implement the system as well as train users.

• Recurring costs are those costs an organization incurs for the BI project that will continue to be incurred after the project has been completed. Understanding the total cost of implementation is imperative to stay within the organization's budget.

Understanding costs can also help determine the phases of implementing the BI application. An estimate of the cost components of a BI project such as software costs, maintenance or support fees, and average configuration cost can be obtained from external resources such as software vendors/resellers and professional services firms. The estimated cost of implementing a BI application will be used in the ROI calculation. Be prepared to compute the cost of having a network connection that's too slow once a week. Accept the fact that your revenue projections are going to falter periodically due to completely unforeseeable problems like the weather, the competition, or your whole Web team ordering the same entree at a local, substandard seafood restaurant and being out of commission for 2 days.

You're never going to be able to report back to upper management on the exact ROI of your entire Web site. But if you're good, you can manage a portfolio of Web initiatives over time, something suggested by Robert Scheier in an article in the May/June issue of Computerworld (www.computerworld.com/ROI).

Quick riddle: What's the difference between a 250,000-square-foot warehouse in Singapore, an upgrade to your firm's Web site, and 10,000 shares in Genetech?

The answer: Nothing. They're all investments you make after weighing their risks and returns. As conditions change, you buy, sell, or hold the overall rate of return on your risk.

The question is always the same: How are you going to measure success? If those shares of Genetech go up 9 percent, is that enough? If your Web initiative saves $25,000 per week for 6 weeks, is that considered "successful"? So pick a target and pick a date. Watch how things run and be ready (here's the hard part) to drop the project like a hot rock when it falls below a predetermined threshold of acceptable progress and profit.

In the end, it all boils down to revenues and costs. Yes, improvements in customer satisfaction can increase sales per order, order per customer, and overall income. And increased employee satisfaction can lower the cost of doing business by reducing the costs of recruiting and training. But they both point back to the triple-threat question: How much did you make? What did it cost? What's the difference?

The big difference is that we've been accustomed to following the money since cuneiform was scratched into rocks. We understand, on a very instinctual level, that if I have five apples and you take away two, something very important has happened and we want to record that event for posterity.

On the Web, our knowledge of metrics began with server log Look at all that data! Is there any information in there?

Web servers were created by programmers, not business analysts. A Web server spits out massive amounts of data, and we business types have been trying to make sense of it ever since. Instead, we should have told the programmers what we needed, so they could build the proper information-gathering mechanisms into the servers. Back when servers were being brought into existence, we didn't have a clue how important our Web sites were going to be. As it stands, businesspeople are tugging on the technicians' sleeves, begging for scraps of relevant knowledge.

With more investment, we can take advantage of more intelligent technology. In the meantime, we can impact the knowledge we glean by carefully constructing our approach to measurement.