So, there's some bits of this story that hint at the problem here. The big headline is that as they removed Breitbart, their spend was reduced. If you're really doing "performance marketing" in the sense that people think of it, that doesn't make a lot of sense. Of course, it makes total sense in the industry.
If you are properly doing performance marketing, that means you're going after the best opportunities to get customers, and you're paying as little as you can for those opportunities. If you cut out a whole set of opportunities from consideration, that should not change your spend. If those were really some of your best opportunities, that means now you'll have to show ads where you have worst opportunities. Either you'll have a lower chance of getting a conversion or you'll have to spend more for each ad... either way, your opportunity cost should go up. Basically you've constrained your supply without constraining your demand, so the price might go up, but demand still gets fulfilled.
Except, this isn't what really happens in the industry. Instead, you're given an ad budget and a campaign window, and you're told to burn through the budget within the campaign window. Most ad networks will try to find the best inventory to give you out of that campaign window... except often the "performance" they're chasing isn't "performance". It's "impressions" or "unique users" or "clicks" or "conversions" (where the goal post on what a conversion means is moved to "goes to the landing page" instead of "buys the product"). So what they end up doing is finding the cheapest (i.e. crummiest and least likely to actually perform) of whatever your "performance" goal is. So you'll get impressions that are least likely to click, or clicks that are least likely to convert, or conversions that are least likely to actually buy your product.
This happens, because there is actually a fixed constraint on the quality side of the supply.
You're trying to spend $X in a tight time window, and the dirty industry secret is the need to spend that money so quickly is the real challenge. Ad networks don't get paid the part of the budget they don't drain. The ad spend is a big deal. More important than performance. Ad execs' comp are tied to that spend.
But when you cut a big chunk of the supply, the expectation is you're not going to hit the same spend, so... you cut your budget too. Well now that changes everything. Now you don't have to spend so much money over so much time. Imagine if you cut your budget in half, but because you are performance focused, you want to lose the worst performing half. You should expect better performance. Way better. Because now some other idiot "performance" marketer is going to buy up all those crummy bits of inventory, and you're going to pay the big money to outbid them for the bits of inventory they were going to get that would have performed well.