Friday, September 05, 2003

Using Sarbanes-Oxley (SOX) to Improve Your Business Analytics, Competitive Position and Performance

Metagroup has suggested that Sarbanes-Oxley compliance spending can be thought of in two ways. You can either think of your spending as a compliance obligation. Or you can think of it as an opportunity to improve your Business Analytics - the more forward thinking way of keeping track of your business (vs. Business Intelligence or BI software).

Conventional wisdom suggests that SOX (as it lovingly referred) to will cause increased audit bills, more review work by software systems providers and increased spending on IT.

But CIOs are under pressure to reduce spending.

So, how can you resolve these conflicting demands.

The solution, I think, is to step away from conventional solutions. Conventional solutions - information warehousing, OLAP, etc. - are not going to solve the problem. And they are certainly not to going to build competitive advantage.

Let me briefly explain why.

1. Most companies using OLAP tools will find that they suffer from exponential growth. What that means is the more you are interested in connecting information in your enteprise reporting, the slower and the more expensive the solutions become. So, if SOX means an order of magnitude or more increase in complexity, your OLAP tool is either going to fail, or you are going to have to subdivide your problem into pieces. Lots of different cubes for looking at your business.

But isn't that going against the grain of what you were trying to achieve in the first place? You have a failing technology, massive and deteriorating technology performance and an increasingly high cost of maintenance. Not really a very good solution.

2. Information warehousing tends to be very expensive. Cleaning data is very expensive. And the best strategies in information warehousing mean starting small. but you have a compliance problem which means that you cannot really start small. You need to grow quickly and, even worse, change frequently as you figure more about your exposure. A very expensive problem.

When technologies start to break down, it's time to change the paradigm.

I have been working with a different paradigm - a object oriented enterprise peformance reporting, modelling and tracking paradigm for about seven years. For a number of years, I built strategic planning, reporting and modeling systems. What we found in our work with OLAP and database tools is that we were always building similar applications, but they varied just enough that each project required starting pretty much froms scratch. Which made consulting tough.

So we decided to ask the impossible question: "How could we deliver complicated views of businesses, their activities, their transactions and their performance measures using an approach that was incredibly fast at low cost?"

If you don't ask such a question, then you are pretty much guaranteed that you won't engineer that kind of exceptional value for money performance in your technology. Over three years, we figured out how to deliver complexity easily at 1/20th of the cost.

What we concluded was that the relational and OLAP framework did not match the strategic and performance reporting requirements that managers needed.

Essentially managers see the world as consisting of pieces of information that they want to combine in many different ways. Sometimes they want to look at information from a product perspective, sometimes from a process perspective, sometimes from a geographic perspective, sometimes they want to compare time periods, sometimes they want to compare versions (actuals, plans, scenarios, revised budgets, etc.)

But what remains constant is the manager's interest in looking at data in new and novel ways on short notice. And business is more complicated that most current software technologies can handle.

If you don't believe t his assessment, ask your top database designer how he is going to handle rapidly changing large scale many-to-many mapping relationships in a relational database. If he tells you the truth, he will admit that relational databases can't deal with the problem. He will tell you that you need to simplify the problem. But SOX and strategic performance improvement mean that you don't want to or can't simplify the problem.

And ask your OLAP vendor how they would model a 13 or so dimensional view of the business. Let's say you want to be able to track and relate the following:

1. Suppliers' suppliers.
2. Suppliers.
3. Administrative processes.
4. Manufacturing processes.
5. Products and product campaigns.
6. Distribution channels.
7. Customers.
8. Market research.
9. Time periods.
10. Actuals, budgets, plans, plan updates, scenarios.
11. Chart of accounts.
12. Currency.
13. Raw operational data.

Most OLAP vendors will be forced to explain that you need to decompose your problem. Which makes a lot of sense, but not if you don't have to. Unfortunaately decomposing your data means more work when you try to look across the decomposed data.

And processing times will start to grow very rapidly every time you add data and complexity. None will be able to handle the raw data. So you will be forced to manage data in at least two places, in relational databases and then in the OLAP environment.

The way out of this box is move to what the Convergent Engineering Institute has called an enterprise representation of the business. It's modern, it works and it is both fast and inexpensive. It's a new paradigm.

With this approach you can now start to look at stakeholders, processes, activity based models; be able to model products and marketing programs, market research and channels, you can focus on the issues that are important rather than the issues that have historically been easy to look at.

There is old story about the wise fool Nasrudin in the Sufi literature. He meets a drunk crawling on the ground beneath a lamp post. Nasrudin asks the drunk what he is doing. The drunk replies that he is looking for his keys. Nasrudin asks in turn: " But where did you lose your keys?" The drunk replies: "Well, I lost them over there by my front door, but there is more light over here."

Looking at SOX, strategic issues and improving performance means looking at what is important, not what is necessarily easy to look at in your legacy systems and legacy technology.

For more on this topic take a look at under Performance Reporting.

Alistair Davidson

No comments: