I think the first assertion needs a slight modification; a first version DW should produce one reporting subject area, not just one report. From there, it's a question of how many reports are needed, which may be only one but could be several.

So the question to the CxO becomes 'About what subject in our business do you need better information?' Typical answers include 'sales', 'inventory', 'customers'. Then we can ask what in particular about that subject is needed (sales across systems, inventory across systems, customers across time) as a genesis for some potential reports (plural) from that starter DW.

The single report question can set wild expectations and backfire into a monster project request that has to be scaled back, which is why I'm wary of it.

Asking up front for potentially anything sets a high first expectation that may need dramatic scaling down. Let's say the answer from CxO is "I really need a combined sales report from systems A, B, C, D, E and F". CxO smiles, having an instant vision of such a report. Good luck bashing that out in a few weeks, so right off the bat DW person has to tell CxO we can only do systems A and B to start. CxO frowns at the delay.

Now imagine asking about what subject first. CxO says, "I need better insight into total sales because I have to look at 5 reports in different formats." CxO frowns at thinking of the hassle. DW person: "I can make a version 1 DW for you in a month that has a unified sales report from systems A and B and we can add the other systems every month or two after that depending on their complexity." CxO smiles.

Meanwhile, a subject area DW should be able to immediately serve other needs. If the one report method is strictly followed then some small incremental data Y that is involved in the subject area in general but not needed for the report would be excluded. Then when VP1 hears about CxO's upcoming new report and says, "I need exactly that with Y added to it" DW person can either say 'we can put that in the report writers' queue because the data is already on the way' (VP and CxO smile) or 'I'll have to see about adding that in" (VP frowns at the delay).

The subject area method can take 10% to 20% more time but the ability to provide N reports instead of 1 provides N*100% more worth and visibility.