Intelligent Enterprise

Better Insight for Business Decisions

RSS
Webcasts
Whitepapers
Subscribe
Home

June 28, 2002

The Finance Function Knocks at Your Door

Are you ready to take on the analytic needs of your finance people?

by Ken Jones

Continued from Page 1

For the balance sheet, some enterprises will need to analyze the nature of their balance sheets in specific ways. For example, financial institutions will need to analyze how their balance sheets are sensitive to interest rate changes or currency exchange movements. Another big area of analysis for the balance sheet is in receivables and payables: Who owes the enterprise money and to whom does it owe money? Is the receivables cycle getting longer?

Four Common Pitfalls

Now that you have some idea of the types of analysis your customers will want to perform, what challenges will you face besides the common challenges of every data warehousing project? I think there are four worth mentioning. They may surprise you because they aren't technical.

1. It has to tie out. First and foremost are the people themselves. Finance people are often caricatured as very conservative, detail oriented, do-what-works kind of people. That's because they usually are. They have to be this way to do what they do, which is both a blessing and a curse in your dealings with them. It's a blessing because they're analytical people by nature, so you have a ready audience for your solution. These are people who surely need and want what you can provide for them. On the flip side, the numbers in the financial data mart will have to tie to the penny with the general ledger or forget it; your data mart is worthless. In fact, it's worse than worthless because now there's a new system to reconcile. My point here is that you have to understand how important it is to quality assure your ETL processes before you even show the results in the acceptance testing phase. And then, after you're in production, each new month of information must tie, or your credibility is gone.

4 Common Pitfalls

1. It has to tie out. You have to understand how important it is to quality assure your ETL processes before you even show the results in the acceptance testing phase.

2. Finance time crunch. You have to be sensitive to the nature of the accounting cycle as you plan your work and execute your plan.

3. Enterprise view. Users may have found their own inefficient way to gain somewhat of an enterprise view. They may not understand the delay associated with providing them the needed enterprise view from a data warehouse system. Education is key to acceptance.

4. Consider Excel. The solution has to be simple to use. Consider presenting the information in the financial data mart using Excel.

2. Finance time crunch. Second, the finance function is cyclical. Anyone who has participated in or even just observed the month-end close process understands the pain these people go through each month-end. The working life of a person in the finance group goes something like this. Gather information, put it in the system, wait for more information from that group that's always late, put it in the system, work like a dervish to reconcile all of the numbers at month-end, take a short breather for a day or two (if they're lucky), and then start the whole process all over again. Quarter-end is worse and year-end is the absolute worst.

What does this mean to you, the data warehouse manager? It means that you have to be sensitive to the nature of this cycle as you plan your work and execute your plan. Don't schedule key deliverables for your finanace customers around the month end or beginning. Don't bug them at this time; they have the close to attend to. Understand and respect what they go through, and they'll support you and the project.

3. Enterprise view. Third, much of the analysis they'll want to do requires information from all over the enterprise. This is precisely because they have the enterprise view. Some of this information will reside in the general ledger and subledgers in the accounting system or ERP system. However, much of it won't. Herein lies a significant challenge. You'll have to extract information from a large number of operational systems around the enterprise.

This will take more time than they expect. After all, they're doing some of this analysis now by pulling together various Excel spreadsheets they receive from different groups and probably extracts from their general ledger. Although this process is error prone and very limited in terms of time series analysis (each month is in a separate spreadsheet), it's being done now today. "What do you mean it will take six months to get the information we need from these other systems?" they'll say.

As a problem that follows from this need for external information, you'll probably find that some of the information that's key for their analysis isn't stored in a nice modern relational database. It's probably captured in Excel spreadsheets or an Access database somewhere on the corporate LAN fileserver. All you can do here is to communicate that you expect this delay will happen right up front. Keep explaining this at every planning meeting until they tell you to stop repeating yourself. In the case of the infamous Excel spreadsheet loaded with key information, you have a couple of options. The simplest and most expedient approach is to define a process where the customer provides the data file to a place where you can secure it and replicate it to your data staging area. The superior but more difficult approach is to examine why this key information isn't captured in a true system of record somewhere. You may need to take the first option and then adopt the second over the long term.

4. Consider Excel. Finally, as I've already mentioned, finance people are do-what-works kind of people. Precisely because they have this month-end time crunch, they would rather use a solution that they know will work even when it's suboptimal. Continuous improvement is in the vocabulary of finance people, but it must be gradual. Let's say you do manage to run the project in a considerate way. You do manage to pull in the required information from all over the enterprise. You do manage to get the numbers to tie to the penny. However, if your solution is more complex or more difficult to learn than the way things are done now, you'll be pushing a large rock uphill in your attempts to get people to use it.

Rate This Article

Comments:

Optional e-mail address:

The solution has to be simple to use. Consider presenting the information in the financial data mart using Excel. Pivot tables are wonderfully simple to use but powerful enough for most needs. If you do implement another presentation tool, they'll always ask, "So how do I get the data into Excel?" Don't get too sophisticated with the user interface; keep it familiar and simple to use.

Is That a Knock at the Door?

I hope I've convinced you that sooner rather than later you'll be getting that knock at the door. When it happens, don't despair. Instead, just relax, because you know how to avoid the four biggest pitfalls on the road ahead. Good luck.

Ken Jones [kjones@clariteegroup.com] is a partner with Claritee Group LLC, which develops business intelligence and performance measurement solutions for its clients' finance function. You can contact him at 215-534-1616.