Account opening and closing. 1 or 2 business processes

Hi,

Opening an account is a separate process to closing and account. Would this be 2 separate processes with 2 fact tables?

Also opening an account generates no metrics. Is that a factless fact?

Thanks
M

It could just be dimensional attributes of the account. Is there a use case?

The accounts are like bank accounts, they are opened and are of a specific type. We need to analyse new accounts, when accounts are funded, fees applied, change type etc.

Would I have a one fact which is account and have the open date, account type, close date as dimensions?

Thank you for your help

Anyone?

I think the answer is the typical "it depends".

If you analytic requirements require analysis of the account activity (How many accounts opened, how many closed, how many stale etc) - then perhaps a fact table is warranted to track these events/states.

If Your analytics only require the account state to provide context (How much growth on accounts opened in the past 6 months), then opening date, closing date etc can just be attributes of the account dimension.

I personally would regard an account opening as a fact. It's a business process that you can map on your bus matrix.

If you're talking about retail banking, then you'd presumably have dimensions associated along the lines of:

Time - Account Opened Date
Location - Which branch did the opening take place at?
Account Type - Which type of account was opened?
Customer - Who opened it?
Sales Person/Employee - Who gets the credit for doing it?

Other possible degenerate dimensions:
Person age at opening (if you wanted to do age profiling of new accounts)

The fact would be: Accounts Opened, nearly always populated as 1

Then you'd have similar for closing, but could include Closure Reason to try whether it is instigated by the bank or customer and why it was closed, again with a fact nearly always populated as 1.

Hopefully that's a decent starter for ten for you.

You can always build aggregate tables on these to exclude customer level transactions - accounts opened by branch and type per day would provide a decent set of metrics, allowing you to see what day is most popular for account openings, etc.