Hi,
Try considering it in this way: first, what must you exclude from the whole dataset? Then, what must we ensure is included? (Processing actually evaluates Include before Exclude, but this thought process helps me to translate my ideas properly into segment criteria).
You stated this is the requirement: the goal is to only include desktop visitors that did not visit 2 particular pages
So, we would deal with visitor containers, but I'll assume that you mean visit containers (since you used visit in your samples). Also, for now, let's just talk a single page, rather than two, because the approach is extendable to multiple pages and far easier to explain.
First, you need to exclude data where the visit contained page X. I can interpret your statement only include...[visits] that did not visit [page X], but we must be more strict in the criteria definition. If we use the Include canvas to specify visits where page does not equal X, we get visits that had at least one web beacon where page does not equal X. However, this would then include visits that touched other pages, consisting even of those that also touched page X! We need the Exclude canvas to strictly cut out visits where page equals X, so that the entire visit is removed if page X was touched at least once.
After that, the limitation of include desktop [visits] is simply added as you suggested, with a rule based on Mobile Device Type. You do not have to use the same approach as above, because of the nature of this criterion. Mobile Device Type does not change, across the entire visit. So, by including visits based on this criteria, you've inherently excluded all those that were mobile devices.
Andy
