There are two questions about the impact of digital change on publishing that are just about impossible to answer.

One is: how much of the sale of ebooks is incremental business and how much of it is cannibalization of prior print sales?

The other is: what will be the fate of independent bookstores?

The two are connected.

As we watch the (long-term) inexorable but (short- and medium-term) unpredictable growth in ebook sales, it is really not possible to tell to what extent we’re just selling established customers the same purchases in a different form (certainly some of it and my personal guess would be the lion’s share of it) and to what extent we’re finding new customers (also certainly some of it and, to my way of thinking, more likely to the user of a multi-function device than a dedicated book reader like Kindle or Nook) or making incremental sales to established customers.

(We plan to address the whether the multi-function device users have a different consumption profile at the Digital Book World conference in January. It’s a knotty question but we think we have a way to get at it.)

The measurements of industry sales have been far too imprecise and muddied to address a sophisticated question like that. (The AAP and BISG are making a serious joint effort to remedy that situation; I have seen some of the great work in building a new data model that has been led by Tina Jordan of AAP and Scott Lubeck of BISG. More on that very promising initiative some other day.) The aggregate industry numbers that we’re used to probably won’t be sufficient to change any closely-held opinions any time soon.

Individual publishers might see data worth intepreting in the total unit sales of major authors that  have established clear sales patterns over time, if they can analyze their way past the fluctuations that must inevitably occur in the sales of each new major release by an established bestseller writer. One place one might expect to see an uptick is in the prior titles in a series (but, even then, you don’t know if the extra sales of four prior Carl Hiaasson titles weren’t instead of sales of four other books, do you?)

My own analysis has been simplistic, assuming pretty much flat sales into the digital future because that has been the case in our overwhelmingly non-digital recent past. When I do the calculations that lead me to think that the sales available to brick-and-mortar stores will decline drastically over the next five years, I’m assuming that the rise of digital sales results in a pretty much equivalent decline in print sales. I also assume that the increase in ebook sales and the reduction in retail shelf space allocated to books accelerates the movement of print book sales to online. If ebook sales aren’t largely cannibalizing, and they don’t themselves reduce the sales available to be made in stores as much as their growth would suggest, then shelf space might not disappear as fast.

My back-of-the-envelope calculations (which have been endorsed in a series of private conversations with publishers, booksellers, and analysts but also strongly resisted in a private conversation by at least one person whose judgment I really trust and also apparently contradicted by the expectations expressed by Random House CEO Markus Dohle in his recent interview) are that brick-and-mortar’s share of total trade book sales will reduce from around 80% today (some say it is higher) to about 30% five years from now. That would be a reduction of more than 60%. Let’s say the share is still 50% in five years (which I speculated might be the number in 2-1/2 years). That would still constitute a 35-40% reduction from where we are today. That’s drastic.

But it still doesn’t tell us “who fails?” Shelf space reductions can come in a variety of ways. Stores can be closed, chain and independent. Dedicated bookstores of all kinds can become less dedicated and turn over shelf space to other items. And mass merchants can decide to reduce the space they gave books or to eliminate them. All three things will happen to varying degrees.

This is a bit like trying to do a weather forecast based on one’s confident knowledge of climate change. The two are related but there are local factors in addition to global ones. Each time a store closes or reduces its shelf space (or, for that matter, in the rarer cases where a new store opens or one increases its shelf space), it affects the fate of the other stores in its vicinity.

On Tuesday night, I came home from a late meeting with a former Cabinet official who was thinking about buying an independent bookstore and seeking my advice, which, based on no specific knowledge, was “don’t.” I walked in to receive a call from a reporter who asked me for my comment on the Barnes & Noble “news.” “What was that?” I asked. “They’re putting themselves up for sale,” he said. “What has happened recently that would motivate that?”

Without having read the press release, which would have signaled to me that they weren’t actually putting themselves up for sale so much as beginning the process of taking themselves private, I strived to answer the question. I thought the acceleration of ebook uptake, some of it fueled by B&N’s Nook device, was recent news that didn’t bode well for physical bookstores. I thought the recent rescue of Borders, which could postpone their demise or shrinking, wasn’t happy news for Barnes & Noble. And I wondered whether the Ron Burkle lawsuit might make the Riggios less interested in owning the business.

Of course, all of those things are true but none of them apply because the premise was wrong. The Riggios are probably not trying to sell the business; they’re more likely trying to buy the business.

Then I checked with a commission rep friend of mine about the bookstore the former politician I met earlier that evening wanted to buy. It turns out to be an independent with a relatively solid future, with knowledgeable staff underneath its owners and a great reputation with the publishers which assures a continuing flow of traffic-building author appearances. In other words, “don’t” might not be the right advice in this particular case.

Whether the brick-and-mortar share of the business falls by 25%, 50%, or 75% over the next five years from what it is now (and all are possible), the reduction in shelf space depends on whether that reduction is against a rising base of total sales or a stable one. And how it affects any one particular store depends on what has happened to the shelf space allocations by others in that store’s immediate vicinity. That will be very hard for anybody to track.

I am still extremely skeptical of recent celebrations of the successes of independent stores, which we’ve seen coming out of New York City and Pittsburgh in the past couple of weeks. Anecdotal information is not projectable data; it is often misleading data. Nobody seems to be making the claim that bookstore shelf space is increasing in New York or Pittsburgh or anyplace else. Any one bookstore might still, for a while, be a reasonable bet. But this is a case where the usual laws of investment (diversify as much as you can) would likely not apply. It is hard to imagine bets on five or ten or twenty independent stores paying off in the aggregate in the years to come. Unless you were making those bets with knowledge about exactly where Barnes & Noble, Borders, Books-a-Million,Walmart, Target, and Costco were reducing their shelf space the odds will be against you, and I’m pretty sure there won’t be anybody who knows all those facts in a timely way.