For there to be a Pareto, the dimension (e.g. Products) must be sorted. After all, the whole point of a pareto is to understand how the top i% of one dimension comprises the top j% of some measure. So when we say "the top 20% selling products are 80% of all sales" then of course the Products have to be sorted descending by Sales in order to determine the "top."
So yes, having the Products not sorted would produce a different and unhelpful result. That does nothing to invalidate pareto analysis.