Conform facts or not?

Hi,

I'm building a Data Warehouse for a media company. They are selling products, but in different ways:
They are selling Apps, they have a Webshop where you can buy different things like smartphones, t-shirts etc. and they also sell printed newspapers.

I'm not sure if I should build a conformed fact table or multiple fact tables and just use conformed dimensions.

The sales processes differ from each other,e.g. the sales from the webshop could be represented in an accumulated fact table: The customer orders an item, when he paid for it the item will be delivered and an invoice created. And when the customer doesn't like what he gets, he can send it back.
When a customer buys an app...he just buys it. Nothing more.
Both sales share some dimensions, but not all. The webshop has a few dimensions more than the app sales.

The easier way would be to build separate data marts with conformed dimensions, but do I have to conform the facts, too?

What is a conformed fact? I've only heard of conformed dimensions. You should end up with lots of fact tables based around ordering, inventory, sales, etc.

I don't think there is such a thing as 'conformed fact table'.

My understanding is ... if same fact (same definition) can be asserted at different level of granularities (in different fact tables, e.g. daily/monthly etc), then the key is to have the same fact name in all tables to avoid confusion; this may be called 'conformed fact'.

Conformed dimesions will let you drill across to:
a) different fact table(s) of the same business process with different granularities, or
b) different fact table(s) of a different business process

Generally facts are broken down by what they represent (type of activity, granularity, etc). They may also be separated by sources if there are operational considerations.

It is also important you get clear definition from the business as to what is a sale. In almost all cases, an order is not a sale, so orders would be a separate fact table. Fulfillment usually constitutes a sale.

You may also consider aggregations to consolidate information across channels and products.

Kimball wrote about conformed facts on page 87 (DWH Toolkit).

My colleague want to put all together in one fact table but I think it's better to build separate tables. Thanks for the hint that usually orders are not sales.