For one thing, they need to be sure the business definitions being used are in line with company usage. Another is to ensure the organisation has all the data it needs to base sound decisions on. "Once you've got the business definitions and you understand the corresponding technical environment that's been put in place, work moves quite rapidly," Wu says. "Ideally you've got a team of internal resources that are going to help champion the solution.

"You also have to understand the corresponding business value of this data." Wu points out that most companies examine the sales cycle in terms of sales person and territory. This is fine, but it only represents a single dimension. Better to look at sales by process, from the issuing of a quote to the issuing of the invoice, shipment and product collection. If you understand the process and can implement measurements or analytics at each step, you get a much better understanding of what is really happening.

"For example, most businesses do not take a look at how many orders were actually cancelled in the quote stage," says Wu. "Why were they cancelled? What was the corresponding dollar amount of those cancellations? What is our corresponding pricing of our products compared to our competitors? Why are orders being cancelled once an invoice has been sent out, or once the product has been shipped? Do we know what's going on? That information is very easy to get a hold of, and it gives a different perspective on the way most organisations view their business."

Because it can be very difficult to work out which questions to ask across the business, it is important to gain consensus from all areas of the business and to have a product champion capable of resolving inevitable issues, Wu says. "There's a lot of effort that goes into specifying these requirements, conducting the interview, gathering consensus, defining it, prioritising what subject areas are going to be addressed, what is the perspective that one needs to take, what are the corresponding delivery mechanisms that one needs for the data itself, how often they need to see it. It raises a lot of issues in companies."

Having good enterprise analytics in place also raises significant change management issues. Making it possible for executives and others to instantly analyse company data can create a backlash as employees become more immediately accountable for results. And the company may need to put in place a new skill set.

"People thought that once you implement these types of applications head count would go down. Just the opposite happens," Wu says. "What happens is head count slightly goes up in some cases and in most cases remains flat. What occurs is you have a change in the composition of your work force. You have now got individuals that you want to hire that can analyse the data versus gathering the data, so it requires a different skill set.

"And now you've got more information than you've ever had before, so you want different perspectives on it. You can analyse more of it than you had before, because before it was spreadsheets, which [meant it] took a month to compile the data.

"Now you've got it on demand. So you can analyse a lot more of it," Wu says.