Well, Mad Men cost an estimated $2.85M per episode (which costs were estimated after Season 4, when it was pulling in an average rating of 2.27M viewers each episode). At $2.99 an episode a la carte, that breaks even at 0.95M. The only season it averaged ratings lower than break-even was the first season, where it averaged 0.90M viewers per episode. This doesn't count the additional profits from streaming licenses and DVD sales, promotional marketing, etc, which are estimated to add well over $1M more per episode.
Maybe that sports columnist chose a bad example, but Mad Men is exactly the type of show which would thrive in an all-a-la-carte world.
I should add that sports in general does get the short end of the a la carte stick, because over 50% of all money going to pay for content in the current bundling scheme goes to funding sports, and sports constitute far less than 50% of the entertainment viewing hours each week. A sports columnist thus has a very significant vested interest in the status quo, because without it the actual costs of the sports he covers would be apparent and likely fewer people would consume as much sports as they do today. In fact, at the link above he clearly states that sports fans are getting massive subsidies from the rest of us.
From a general economics perspective, his argument fails. I'll avoid drawing unflattering parallels between what he is proposing and what his politics likely deem to be evil political systems of the world. Essentially, though, capitalism works best when there is a clear "vote with your dollars" ability on the part of the populace. That is, if someone doesn't like something, they don't buy it. That makes it very clear which products are succeeding today and which are not, so that development dollars tend to go more to those which are successful rather than being wasted on things no one finds valuable (late-night filler or instance). Now, projects can still be funded if there is not an audience for them; that is how the entire production side of the entertainment industry already works. And, if there are projects with socially redeeming value despite low viewer interest those too can be funded via social funding means (NEA grants, etc).
In any case, bundling of services is almost universally seen as consumer-hostile in the world of economics. The only place where bundling is beneficial to the customer when that bundling is itself a service - a way to reduce complexity from an untenable level to something the customer can easily understand and intelligently evaluate. A la carte content purchases are nowhere near untenably confusing, so there is no value in that aspect to bundling. Which leaves us just with the costs of bundling. Which means that a bundling solution is always, on average, worse for the consumer than an a la carte solution.
Of the 11 shows nominated for Golden Globes in the Drama and Musical/Comedy areas this year, 3 were on HBO ($15/mo), 2 on Netflix ($9/month), 2 on Amazon Prime ($8.33/month), and 1 on Hulu ($8/month). With just over $40 in (generally redundant) subscription costs, you have 72% of the best TV out there fully funded (or for just $32.50 you have 64%). And this is assuming that the rest of the stuff on those services (the redundant movie libraries) take none of the subscription revenues to fund themselves, which is of course ludicrous. While this is still bundling, it shows just how inefficient the multi-level bundling operation (shows -> channels, channels -> multi-channel cohorts, multi-channel cohorts -> packages) is at providing what most people consider "good TV".
I don't think the Netflix model is quite as horrible for quality television as you seem to believe it is.
That said, I'd much rather buy each of those shows separately ($40/season on HBO for instance, which is just under 3 months' of the service) to avoid paying for the redundant movie and TV show back-catalogs on multiple services. Since doing so for some of them means I have to wait a year after airing, instead we just binge-watch shows in 1-2 month subscription "bursts" to keep our overall subscription costs per month under $20 reliably.
