So Nov. 30 has come and gone, the day that President Barack Obama promised would be up and working. And his administration says that the site is working, dramatically better than when it first went up. On the other hand, anecdata suggest slow enrollment. So what does it all mean? Is the website working? Is Obamacare saved?

To answer that, let's break down the details a bit. First, the good news: Compared with the chaos of October, the consumer experience is much better. It could hardly have gotten worse; for the first eight days, to a first approximation, no one could get coverage, and things were only marginally better by the end of the month. But the administration says that error rates have fallen dramatically, and the site can now handle 50,000 simultaneous users, where previously as few as 500 or 1,000 users would completely crash the site. In early November, the site was down more than half the time; now it has greater than 90 percent uptime. A source in the know tells Bloomberg that 100,000 people signed up in November, four times the pace of October enrollment.

At the very least, we've moved from "is obviously making it impossible for people to sign up" to "people can get through the process of selecting a plan if they're willing to make a reasonable try." That's huge progress.

And "here's what's indisputable," writes Ezra Klein. " is improving, and fast. Or, to put it differently, will be fixed. In fact, for most people, it is probably fixed now, or will be fixed quite soon":

The repair job is likely proceeding quickly enough to protect Obamacare from the most severe threat to its launch: Democrat-backed legislation unwinding the individual mandate or other crucial portions of the law. So long as people can actually purchase insurance through the federal exchanges, congressional Democrats are likely to support the basic architecture of the legislation they passed in 2010.

Now for the bad news: Although the administration says its repair team is working with "private sector velocity," a private-sector website would still count this as a very bad performance. Any business person will tell you that every time a customer has a bad experience, you risk losing them. And the customers you lose are often the ones you most want to keep.

Further bad news: The administration has chosen to count not how many people have actually enrolled in a plan, but how many people have put a plan in their electronic shopping cart, whether they checked out or not. Some of those people will never actually enroll in any plan. Some further number will never pay, or they will stop paying after a short period of time.

Will they pay? We couldn't tell even if we wanted to, because the part of the website that pays insurers still hasn't been built yet. In fact, while the part that serves consumers is working much better, the part that sends the information to insurers is still having a lot of problems. So even if we knew how many consumers intended to actually enroll in, and pay for, a policy, we wouldn't know how many of them would actually have an insurance policy come Jan. 1.

Which sums up all of our information about the site, broadly: We don't know, because the administration doesn't really have the information we want or any way to get it. It's clearly choosing the most optimistic metrics possible while ignoring more obvious ones. But even the more obvious ones wouldn't tell us what we really want to know, which is: Come Jan. 1, or April 1, how many people in the U.S. will have insurance?

The administration undoubtedly has a good count of how many error messages it has gotten. But how many people broke off their visit to the site because it was frustrating them? How many people didn't visit, and therefore didn't buy, because they'd heard the site was buggy and insecure? On the other hand, some other number of people haven't even touched the site yet but will eventually log on, choose a policy and pay for it. How many of each? There's no way to tell; we're in uncharted territory.

As I've remarked before, ultimately the metric that matters is how many people sign up (and whether enough of them are young and healthy). The Congressional Budget Office projected that 7 million people would get policies through the exchanges in the first year -- 2 million switching over from other policies, and 5 million new purchasers of individual policies. Obviously, 100,000 is not very rapid progress toward that goal. But how close should we expect to be?

Sorry to sound like a broken record, but there's no way to tell. The law's supporters have been pointing out that Massachusetts got very few people to sign up in the first few months, but it gave people longer to sign up before the mandate kicked in. And I'd have expected a post-Thanksgiving surge, which is apparently what we saw with Medicare Part D, but that obviously didn't happen on the federal exchange. I'd guess that a minimum of a couple of million people need to buy insurance by Dec. 23, just to replace the policies that were canceled because of new Obamacare rules. can probably handle that load, if it's evenly distributed. But of course, it may not be evenly distributed, and it may not show up at all.

Here are a few things I think we can say: It's working better now than it was two months ago; we don't know whether that's good enough; and the fact that the administration is choosing odd metrics, such as comparing the performance of the site to October rather than how you'd actually like it to perform, is probably not a happy omen. (As I believe Scott Adams once noted, irrelevant comparisons are a great favorite of salesmen with mediocre products. "Sure, 37 mph isn't great for a sports car, but you have to compare that to hopping!")

And I do think this tells us something else important: The administration has given up on success, as it might once have defined it. The object is no longer 7 million people signed up through the exchanges, with 2.7 million of them young and healthy, and the health-care cost curve bending back toward the earth. It is to keep the program alive until 2015. The administration's priorities are, first, to keep Democrats from undoing the individual mandate or otherwise crippling the law; second, to keep insurers from raising premiums or exiting the marketplace; third, to tamp down loose talk about the failures on the exchanges; and, only fourth, to get to the place where it used to think it would be this year, with lots of people signed up for affordable insurance. It is now measuring the program's success not by whether it meets its goals, but by whether it survives at all. And all of its choices are oriented toward this new priority.

You can see this in the decisions the administration made about fixing It focused on the part that voters can see, even though the part that accurately transmits data to insurers is arguably more important -- is it better, or actually worse, to "sign up" a bunch of people when you can't get that information to the insurers who need to write the actual policies? You can see it in the strategic delays, particularly the delay of the open enrollment deadline until after the 2014 elections. And you can see it in how the administration is treating insurers. As plan cancellations became a big political problem, the administration looked like it was preparing to blame insurers, which has been a very successful political tactic for them in the past. But it quickly walked that back, because with the program's survival on the line, it needed insurers on board. That's why the administration is looking to get extra money to the insurers; it's the sweetener it needs to forbear little things such as possibly not getting accurate enrollment data, or payments, for months.

From the administration's perspective, this makes a lot of sense. A program that survives until 2015 can hopefully be fixed. A program that is fatally damaged by Democratic or insurer defections definitely can't. The question is whether the public will embrace this new way of measuring the program's success ... and what happens if it doesn't.

This column does not necessarily reflect the opinion of Bloomberg View's editorial board or Bloomberg LP, its owners and investors.

To contact the author on this story:
Megan McArdle at