Every now and then I see someone make mention that hand-in-hand with the decline in post-free sales, free books are not being downloaded in the same numbers as they were earlier in the year.
Is it true? Or does it just feel that way? It's a statistical claim, so what data is the observation based on?
Authors who aren't getting as many downloads per book as they once were might be tempted to extrapolate that data point across the entire freebook model. But just because we're tempted to do something, doesn't mean it's the right thing to do.
I am, however, going to fall into that same temptation. The only differences are that 1) I have a LOT of free runs to compare, 2) many of those free runs resulted in comparable ranks so I'm comparing like to like, and 3) I understand how to look at the captured data in a way that (hopefully) returns meaningful results.
For instance, on its first free run in January, SECTOR C hit #25 and ended up with 13,000 downloads in the US. That's how most authors would report their free run result, right? Did that mean it took 13,000 downloads to hit #25? Not at all. It took about 10,000 to hit that rank. In fact, 8 hours later it went off free at a rank of #35 with those 13,000 downloads -- and it took about 7000 downloads to hit #35 the first time. But the way indies often report results complicates and skews those results -- which makes it difficult to work with other people's data.
So, working with the pool of data I do have -- 10 books in Select through Steel Magnolia Press, which is admittedly quite a small pool but is perhaps better than what many indies are armed with -- I charted how many downloads it took to reach a given range of Top 200 ranks during each month in 2012.
Here are the caveats:
- Within a given month, the free runs logged may have taken place on different dates.
- There is always server lag in reporting downloads vs rank. I've used actual rank figures and backed up an hour to two hours to get the download numbers associated.
- In the Top 50, it's very competitive. While I've segmented the ranks by 10s, the difference between a rank in the low end of the segment and one at the high end can easily be 1000 downloads.
- Each month we had more books in Select, so data for January and February is not as robust as that available for March, April and May.
Still, there are some trends we can plausibly infer from the charted results below.
- January appears to be the high point for freeloads across the store, followed by February.
- April saw a bit of a depression of about 20% across the board.
- May freeloads appear to be at about the same level as February.
- Readers still appear to be enthusiastic about freeloads.
And here's the kicker to what this data appears to indicate (and please note the frequency of the word "appears" throughout this post!):
There are now 153,000 books that are Prime Eligible compared to about 65,000 in January. Not all of those are in the KDP Select program, but for simplicity sake let's say the number of books eligible for a Select free run has doubled since the first of the year and posit that there are now twice as many books being offered free as there were then. If all the books being made free on a given day were being downloaded equally, then it should take roughly half the number of downloads now than it did in January to reach a given rank. But that's not what we're seeing. We're seeing it take the same number of downloads to hit a given rank even though there are almost twice as many books available.
That means either the overall number of free books being downloaded has increased exponentially because there are more customers downloading them or the same number are downloading more books, or else the overall percentage of books rising to the top has gotten considerably smaller, meaning there are a LOT more books not getting many downloads at all.
It is, of course, likely to be some complicated version of all three scenarios.
What this means is that using free as a tool for discoverability is still a great option assuming your book gets into the Top 200 or so, but it rapidly loses its effectiveness otherwise. Note I said "for discoverability" and not "for gaining post-free sales." With the new algorithms and their accompanying new price bias, books likely need to get into the Top 150 Free to have any noticeable post-free sales bounce, and into the Top 80 to see an appreciable bounce.
More and more, you have to be in this game to win it.
The bar keeps being raised. Are YOU up for the challenge?