Obamacare's late enrollment surge was extremely impressive - an amazingly powerful testament to the American powers of procrastination. At the end of February, the number of people who had signed up for health-care plans stood at 4.2 million; as of last Monday, it seemed to be about 7 million. More people selected plans in March than in the first three and a half months of the open enrollment period.

There's a lot that can be said about this, and almost everyone is saying it. Here's one thing I haven't seen mentioned, however: All the data we had about enrollment are now blown to hell.

I don't mean the data don't exist, of course. We still have surveys and exchange data from the first five months of operation, and when the history of the Affordable Care Act is written, they will provide a very interesting narrative. But they are pretty much useless for answering the important questions:

• What percentage of people who selected a plan were previously uninsured?

• What is the demographic profile of enrollees?

• What percentage of people are going to actually pay their first month's premium and then keep paying every month?

Forty percent of total signups came in the last month of operation. If those signups were systematically different from the earlier ones, then predictions based on data from earlier months will be wildly off base.

It's impossible to say for sure, of course, but I think that the people who signed up in the last few days will be systematically different from the earlier groups. Because our data have, as mentioned above, been blown to hell, allow me to speculate a bit about the ways that they probably differ.

Start with what we know about these people: They waited until the last minute to sign up. This tells you that a lot of them will fall into one of a few groups:

• People who are incredibly disorganized.

• People who are so financially pinched that it was important to wait until the last minute so that they could pay eight months' worth of premiums instead of 12 (many people who bought last night will not have coverage until May).

• People who are young and healthy enough to make acquiring insurance less than urgent.

So we are probably looking at a pool that:

• Is somewhat younger than we'd have expected in previous months - probably at least 30 percent young adults, maybe 35 percent, maybe even the original 40 percent that the administration was projecting, though that's still not the most likely outcome. In order to hit that original projection, 60 percent of the newly covered people would have to be ages 18 to 35. That's a pretty heavy lift, because people in that age group often come with minor children or older spouses.

• Contains more uninsured people than previous surveys would suggest.

• Will suffer more attrition from nonpayment, as the irresponsible fall behind and the financially strapped find they can't pay the new bill every month.

So almost everything you read in the next few weeks about the characteristics of the insurance pool is likely to be wrong.

This puts pundits in a tough spot - what the heck are we supposed to fill our column inches with? But the folks in the toughest spot are the insurers. Depending on the state, they may have to file their preliminary 2015 premium information as early as May - when they won't even know how many of their new customers are actually paying.

And if they don't know, the rest of us certainly don't. So beware any pronouncements on what exchange enrollment looks like. Other than the raw number of people who selected plans, we just don't have any idea.

Megan McArdle writes about economics, business and public policy for Bloomberg View.