Thursday, 19 April 2018

Pricing Part 3

Back in 2014, I shared some observations about pricing PDFs on OneBookShelf sites. I am told some publishers found it useful as a guideline for OBS marketplaces. Still, a number of questions arose as a result of that initial blog post, such as, how did some of the biggest, most prolific publishers skew the results? And how did those observations apply to print-on-demand pricing, as opposed to digital pricing?

I am not a statistician or a data scientist, and I won't pretend what you'll see here is any more sophisticated than my last blog on this topic. Moreover, as I noted last time, these data are correlative, not causative. Still, I hope you might find something that helps you in deciding how to price your titles more effectively.

Note: All numbers and data visualizations are based on gross sales revenue for both DriveThruRPG and RPGnow, combined, from all of 2016 and 2017. Dungeon Masters Guild data are not addressed here.

Revenue by Price Point

As in the previous pricing blog, let's take a look at what price points have done best on our marketplace in the past two years, in terms of gross revenue. After that, we can start looking at the same data sliced up a few different ways.

Fig. 1: Gross revenue by price point (in $1 increments)

In this regard, at least, things haven't changed much in the past three years or so. There are spikes at each of the $5 price increments, but those increments are also by far the most common price points for products on our site.

One improvement of the data visualizations here, over those I provided in 2014, is that now you can see digital sales separated from print sales: Print-on-demand book revenue is indicated in orange above, while digital revenue is blue.

Complication: For orders which include both digital and print titles, data are duplicated: When we sell digital + print product, for example, that order shows up as both a digital sale and a print sale in the metrics. That's why you'll notice, above, that a few print sales show up even at very low price points. Taken in aggregate, though, the visualization is still helpful for seeing trends, even if the data aren't perfectly clean.

Revenue per Title by Price Point

Now we'll take a look at the same data adjusted for the total number of titles available at each price point; looking at sales revenue relative to catalog size this way, we have a clearer sense of what have actually been the most "successful" (in the sense of lucrative) price points on our store.

Fig. 2: Adjusted revenue per available title, by price point ($1 increments)

As you can see above, the $19.01-$20 range remains the most lucrative price point on our site, just as we saw back in 2014. This is purely an observation of past buying behavior; it shouldn't be taken to mean you have to price things at $19.99 if you want to sell product. If publishers suddenly changed their pricing so that the vast majority of titles on our marketplace sold at, say, $6.95 or $22.99 (for whatever reason), you might see those specific price points gradually become more popular with customers over time.

Broader Price Bands

The histograms above only go up to prices of $60, while print prices can go somewhat higher. To take a look at print pricing specifically, we need to extend things a little further along the horizontal axis. To do that, I have grouped the values a little more tightly.

Here's the same information I've shown above, but extended upward to a $150 maximum, with prices aggregated into $5 increments, rather than $1.

Fig. 3: Gross revenue by price point (in $5 increments)
And here are those same $5 increments, again adjusted for catalog size within each increment.

Fig. 4: Adjusted revenue per available title, by price point ($5 increments)

You'll note that as the prices increase beyond $20, digital selling power drops drastically, while print power rises somewhat, up to about the $70 price range.

Looking at Figure 4, above, you might be tempted to think that digital titles selling for $10 or less aren't "worthwhile." However, in Figure 3, you can see plainly that the bulk of digital revenue actually comes from titles in those lower price ranges. There are vastly more titles priced at $10 or lower on our sites than there are more expensive ones.

Average Revenue per Print Title, by Price Point

Taking digital sales revenue out of the equation for a bit, let's focus on print-on-demand (POD) titles.

First, some basics: Just under 12% of the total revenue on our marketplace comes from print sales. However, fewer than half of the publishers on our site offer books in print, so that percentage, for those who do sell in POD, averages somewhat higher than 12%. On the flip side, print books do have a base printing cost that must be paid to the printer, so the margin on those sales is sometimes smaller than that for digital books (but not necessarily so, and publishers are always free to set their own margins!).

Offering Digital and Print: Past research indicates that publishers who launch a title in both print-on-demand and digital formats at the same time will see nearly 30% higher revenue for that title within in the first year. Print versions released within a year of the digital launch, but not at the same time, will see only about half that increase, or about 16%.

Now, how about some more graphs? First, let's compare black & white POD (print-on-demand) revenue to revenue from color books.

Here, you can see the gross revenue by price, again adjusted for the number of titles offered at each aggregate price point, for black & white print RPG books.

Fig. 5: Adjusted revenue per B&W title, by price point ($5 increments)

And here is the same view, but for color books in print rather than B&W:

Fig. 6: Adjusted revenue per color title, by price point ($5 increments)

Comparing the two, you'll see that customers tend to purchase black & white books more readily in the $15 to $40 range, whereas color books tend to sell strongly in a couple of ranges: from $20 to $50 (standard color), and then from $55 to $70 (premium color).

In Figure 6, you'll also note tall bars for color books around $100 and again around $120; these last two price points correspond directly with big hardcover color books on major game lines from top publishers, such as core World of Darkness and Chronicles of Darkness books from White Wolf and Onyx Path Publishing, or something like the massive Ptolus hardbound book from Malhavoc Press.

Softcover vs. Hardcover

Another important distinction among print books is softcover versus hardcover, so let's look also at those numbers.

Fig. 7: Adjusted revenue per softcover title ($5 increments)

Fig. 8: Adjusted revenue per hardcover title ($5 increments)

You may have noticed that Figures 6 and 8 look very similar, with larger bars at consistent price points: Again, as you'd expect, big premium color hardcover books can sell well at much higher prices than other types of books, and those are the more popular formats on our sites.

Pricing print books is a deep rabbit hole. Let's move away from this, for a moment, and consider a new complication while we let things sink in.

Publisher Revenue by Tier

One big question people had after the last blog had to do with publisher size and popularity. If major game publishers sell PDF books at $19.99, does that mean a small, one-person enterprise should price books at lower prices?

Again, let's start with some basic information. Back in 2014, I stated that the top 17 publishers on OBS sites accounted for about 50% of revenue. That fact is still more or less true, although the numbers have changed slightly since then. In the past two years, the 25 most active publishers on our sites generate 50% of the revenue (which is a good thing, indicating that several new publishers have pushed their way up into the bestselling range since 2014). The remaining 4,100 or so active publishers on our RPG marketplaces ("active," in this sense, meaning those who have at least one public title) account for the other 50% of revenue.

One final point of interest: The least active (i.e., lowest-selling) 70% of publishers account for just 1% of the total revenue generated on our sites. Framed the other way, the top 30% of publishers generate 99% of the revenue.

Publisher "Tiers" 

For the purpose of this analysis, I have split all RPG publishers into four "tiers," each of which accounts for roughly 25% of revenue on OBS, as indicated below.

Publisher Tiers and gross revenue per tier

Gross Revenue by Tier

Now, how do those publisher tiers break down in terms of revenue by price point? The visualization below shows the gross revenue for all digital product orders in 2016-17, broken out into the four publisher tiers described above.

Fig. 9: Gross digital revenue by price point, per publisher tier

Here is a similar view, but this one focusing on just print revenue as opposed to digital.

Fig. 10: Gross print revenue by price point, per publisher tier

In Figure 10, note that Tier B (comprising the lower 20 of the top 25 publishers on our marketplace) earns considerably less from print sales than any other tier of publishers, even though they generate 25% of site revenue overall. What does this signify? A number of those publishers don't offer their books in POD at all.

Title Performance by Price Point, per Publisher Tier

As we've done before, let's now look at revenue relative to the number of titles offered within a given price range, rather than just gross revenue by price point.

Fig. 11: Adjusted revenue per digital title by publisher tier

And here is the same analysis for print revenue.

Fig. 12: Adjusted revenue per POD title by publisher tier

The data might be a bit hard to see with all four publisher tiers together in two stacked bar graphs, so below, I've also split them out into four separate bar graphs, one for each tier. In these graphs, as in those earlier, the orange section represents print revenue, while blue represents digital revenue.

Fig. 13: Adjusted revenue per title (Tier A)

Fig. 14: Adjusted revenue per title (Tier B)

Fig. 15: Adjusted revenue per title (Tier C)

Fig. 16: Adjusted revenue per title (Tier D)

Revenue over Time - Area Graphs 

One type of visualization I've found particularly useful in helping various individual publishers understand how to price their titles is to use an area graph to show their past revenue over time, with colored areas representing sales within designated price ranges.

Area Graphs (Digital Revenue)

Here, I've created area graphs for each of the four publisher tiers, with digital product prices broken down in $10 increments and color-coded as follows:

Fig. 17: Gross digital revenue by price over time (Tier A)
Fig. 18: Gross digital revenue by price over time (Tier B)
Fig. 19: Gross digital revenue by price over time (Tier C)
Fig. 20: Gross digital revenue by price over time (Tier D)

Area Graphs (Print Revenue)

This time, for print products, the prices are arranged in $15 color bands. I've also grouped Tiers A and B together, comprising the 25 most active publishers who account for 50% of revenue. Then, Tiers C and D, those remaining 4,100 or so publishers who generate the other 50% of revenue, are also grouped together.

Fig. 21: POD revenue by price over time (Tiers A+B)
Fig. 22: POD revenue by price over time (Tiers C+D)

Comparing Figures 23 and 24, you can see that Tiers A and B tend to sell comparatively more books in the $45+ range, while Tier C and D publishers sell nearly all of their print books at $30 or below.

So How Does All This Help Me?

Again, I must stress, the data here are simply showing trends for 2016-17. They don't necessarily tell us what we should do in the future. I find the information interesting, and I expect some of you will too. But how does it help us?

I think it's fairly reasonable to assume that customers in the next year will buy more or less the same kinds of things they did last year, at more or less the same prices. If that holds true, then we can use the data to inform us about what kinds of prices might work well again.

There are a couple more ways of looking at the data that might help you in setting your own print prices going forward.

Print Margin

First, people often ask us what is a reasonable margin on print books? The print cost is always deducted from each sale before the remainder (the margin) is divided up into royalties. What is the optimal size of margin compared to a book's base print cost?

I can't tell you what "optimal" might be, but what I can do is show you what the relative margin size is of bestselling books from publishers in each of the four tiers I've used above. That way, you can judge for yourself what might be a good margin for your books.

Let's take a look at what sort of margins other publishers have applied to their own bestselling print books, by tier. In the graphs below, each bar represents one specific print title; the red area indicates that book's base printing cost, while the blue area is the margin. The size of the bar indicates the total revenue of the book in 2017.

Fig. 23: Print cost and margin (Tier A)

As you can see, print books show the same kind of Pareto, or "long tail," distribution as do digital titles (and most other things, really). Here, the relative amount of margin to print cost is pretty consistent down the curve. Taken as a whole, bestselling Tier A print books have a margin that comprises 55.2% of the book's final price. Tier C and Tier D top sellers are quite similar, with average margins making up 58.6% and 57.4% of their final prices, respectively.

Here, Tier B publishers are the outliers, with their margin accounting for 68.7% of a bestselling book's final price on average. However, recall that Tier B publishers are also the ones with the fewest POD books for sale as a group. In fact, so sparse are these titles that their graph does not resemble that of the Tier A group above; only thirteen titles from this tier qualify as print "bestsellers."

Fig. 24: Print cost and margin (Tier B)

Pricing by Format and Page Count

All of this is helpful, but again, it doesn't really get at how much a book should sell for. Well, let's not say "should." Instead, let's look at how much the most successful books in each print format have sold for historically, and then we can let the data lead us to our own conclusions.

Here, I have limited my analysis to those books that earned the greatest margin relative to their page count and number of orders, omitting the bottom 50%. Taking these "most lucrative by page count" books and then plotting them on a scatter graph, with selling price on one margin and page count on the other, we can see the distribution of those books.

Fig. 25: Price by pagecount, top 50% bestselling hardcover books (all formats)

Now, by slicing the data away for each book format and type (soft and hardcover, and black-and-white vs. standard color vs. premium color), we can start to see what prices have sold best on our marketplaces.

For example, here is a look at the selling prices for the top 50% of softcover books in black-and-white format during 2017.

Fig. 26: Price by pagecount, bestselling black-and-white softcover

Now, using the median line or "line of best fit" visible on the graph above, we can find the equation that produces that line, giving a sort of "formula" that represents the median price for such books. Again, I cannot stress enough that this approach is not predictive, but simply provides a mathematical abstraction that represents what prices, in aggregate, sold best for books of various sizes and types last year.

However, if you are in need of some sort of guideline, in the absence of any other, you could do worse than to use this data to provide a starting point, or at least a discussion point, when considering how to price your titles in print.

Here, as another example, is the plot for premium (full-color) hardcover bestsellers.

Fig. 27: Price by pagecount, bestselling premium color hardcover

In case you find any of this useful, or if you'd just like to see what the relative prices of bestsellers have been in 2017, I put together an interactive Excel spreadsheet I'm happy to share. It contains equations for each book type, along with a low-, mid-, and high-range value, all calculated using the kinds of data I've shown above. With it, you can enter a page count for a book of any format, and the sheet will spit out what the historically best-earning (by page count) low, medium, and high prices have been, typically, for books of that size.

You can download the spreadsheet by clicking here:

(And remember, the worksheet is simply showing you typical prices that have sold well in the past, not telling you how you *should* price your book!)


The relationship of pricing to revenue is not necessarily causative. We can't tell from the data, for instance, whether Tier A and B publishers might generate more revenue simply because their prices are higher, or if they are able to price things higher (and therefore make more money) because they are well-regarded.

However, there are a few fairly safe observations to be made, even if we can't say explicitly "this is how it should be" or "here's what you should do":

  1. Selling digital titles at $20 or less remains wise, unless you have a large, premium, full-color title with lots of art. Even in that case, the upper limit should be $30 at most. (See Figures 1-4.) Customer just don't normally spend more than that on digital titles. 
  2. Print-on-demand book prices can vary greatly (but if you want a hint as to what kind of prices other publishers might have used for top-earning books similar to yours in the past, see the embedded spreadsheet above). 
As I stated in my previous pricing blog a few years ago, price ranges are very likely affected by the (perceived) quality or value of a given publisher or product line. As seen pretty consistently in Figures 9 through 22, higher-selling publishers tend to sell more products at higher price ranges.
Further, we do know that regular release cadence and catalog size are very important factors in establishing higher revenue. Publishers who release more titles more often draw more attention to themselves, thus driving more sales, and they also have more titles on the "shelves" for casual browsers to find, increasing the odds of an impulse buy.

So, maybe in another two or three years (but sooner than that, I hope), we'll get around to taking a close look at some of those other factors that can affect pricing and sales.