The preferred language of Math Types involves Xs and Ys, and symbols combined into formulae, which they structure into proofs or how-to cookbooks. English has a similar structure: letters, words, sentences with punctuation, paragraphs, and instruction manuals (which may be as unintelligible as math spreadsheets). Formulas communicate information efficiently—if you know how to interpret them. Otherwise, we should start with English—which is what I propose to do.
The only sane reason to run ads is eventually to sell stuff and make a profit. For authors, seeing their book in an ad might satisfy their ego, but ego doesn’t put food on the table. By profit, I mean making more money than you spend. You need to know both how much income you gained and how much money it cost to generate that income to calculate profit. If you spend more than the income you earn, you can’t make it up in volume.
If an author has only a single book, it’s easy for her to figure out if the book was profitable. Take all the money anyone paid her for the book and subtract everything she spent to make those sales. The difference is profit if it’s positive (and a loss if it’s negative).
Figuring out if an individual ad is profitable is not nearly as easy. What you paid for the ad is the cost part of the equation. Determining the extra income resulting from the ad is not as straightforward. Unless your book was not selling at all before you ran the ad, then it has some (unknown but estimable) run rate—the base level of income you expect without the ad. You need to choose a period to measure—say, a month. For our purposes, let’s say we expect to have $100 of profit for the coming month, assuming we do not run any ads.
Now, let’s run an ad for the e-book format of our spectacular novel. The ad costs $25. If the ad does not encourage any extra people to buy the book (and the ad isn’t so terrible it turns away those who were already going to buy the book), our income holds steady, our expenses grow by $25, and our profit declines to $75. Of course, we expect our ad to work—that’s why we’re running it! Let’s say for every book we sell, Amazon gives us $3.44 (my income on books priced @ $4.99).
To earn back our $25, we need to sell 7.27 books. (The cost of our ad divided by the income we receive from the ad. $25/$3.44 = 7.27). Since people don’t buy fractional books (we’ll leave out of the discussion that people do exactly that when they read books in a subscription service like Kindle Unlimited), it takes us eight books to earn a profit. That’s eight MORE books than we already expected to sell.
At the end of the month, we determine how many books we sold and how much extra money that brought us over what we had estimated would happen without the ad. That’s our extra income. We also determine the cost of our ads (unlike my example, in the real world the actual cost of ads may depend on how many people see them or click on them). The difference is our extra profit (or loss). Piece of cake.
Except we made a ton of assumptions, including these:
1. We assumed we knew exactly how many books we were going to sell for the month had we not run the ad.
2. We only ran one ad.
3. We only had one book.
4. Our ad didn’t affect paperback or audiobook sales.
5. No one who bought our book this month would have bought the book in some later month had they not seen the ad. (That is, we didn’t simply steal a future sale and bring it forward. If anyone did that, then we’ve overstated our profit.)
6. No one buys the book next month because of the ad they saw this month. (If they do, we’ve understated our profit.)
7. We assumed the ad only affected sales of the advertised book (because we said we only had one book).
The multiple ad problem
Many of us run many ads during a month. How can you determine which one led to which sales? Sometimes we can track the source of the sale to a particular ad through cookies and the like. However, research shows that buyers often need multiple exposures to a product before they buy it. So even if you “know” ad #3 triggered a sale, you can’t know whether the buyer also saw ad #1 and/or ad #2, and it was only the combination of the ads that produced the sale. Anyone who tells you different is blowing smoke to cover the lack of data and the messiness of human purchase decisions.
Which does not mean we shouldn’t try to figure out which ads were effective and which were not because we want to keep using variations of the effective ones and stop using versions of the duds. At the end of the month, the only thing you really know is whether you made more or less money than the month before.
The complication of having multiple books for sale.
To complicate matters exponentially, let’s release condition #7 and instead of a single novel, let’s assume we have four books published in a series. We could do all the same calculations but doing so tends to understate the effectiveness of our advertising. Why?
The Magical Beans of Sell-through.
Some people who enjoyed reading your first book will buy another of your books. And if they like that as well, they might buy the third. And if they like all three, they’re even more likely to buy the fourth. Might – no guarantees, even if they rave about how good your book is. Like Jack’s beanstalk, once you plant the seed of a good book, the stalk can sprout to produce purchases of your other books. Authors with stand-alone novels experience the same effect, but it’s stronger with series. No surprise, then, that publishing companies like series.
But like Jack’s beanstalk, the farther you move away from the sale of the first book (the seed in the ground), the weaker the stalk linking future sales becomes. What IS he talking about?
Readers loved our book—at least some of them did—and that encourages them to read another. For discussion purposes, let’s say we have a 50% conversion rate from book #1 to book #2; that is, for every 100 people that read book #1, fifty will buy book #2.
If they like both books #1 & #2, they are even more likely to buy book #3. Let’s say 80% (80 of 100). If, however, book #2 suffered from a sophomore slump, the sell-through could drop to say 20%. Assuming they read the first three books, the probability that they buy the fourth book should be higher (again assuming book three was not a dud)—we’ll use 90%.
Why do we need all these numbers and percentages and all that math stuff? Because if an ad gets a reader to buy book #1, we can also expect some portion of them to buy the other books. To understand the effectiveness of our ad, we need to reflect these later sales because they only happened because our ad encouraged the reader to buy the first book (and our terrific writing encouraged them to read the next ones).
Let’s see how this all works with our four-book example.
The reader bought book #1 – we earned $3.44.
Of those, 50% (50/100 = .50) bought book #2. Let’s assume it sells for the same $4.99 and earns $3.44 (BUT only for the 50% who bought it). We also expect to earn $1.77 ($3.44 x .50 = $1.77) on book #2 for every book #1 we sell.
Now, of those who bought books #1 and #2, we’re assuming 80% buy book #3. All three things have to happen: the book #1 sale (100%) the book #2 sale (50%) and the book #3 sale (80%). When multiple things all must happen, we multiply the probabilities (the percentages) together. So, 1.00 x .50 x .80 = .40 (or 40%). That means we expect 40% of the people who buy book #1 also to buy book #3. Let’s say book #3 sells for $5.99 and yields us earnings of $4.14. If we sell book #1, we expect eventually to earn from that person a book #3 sale worth $1.656 ($4.14 x .40 = $1.656). Don’t round these numbers because it leads to errors when you use them. Only round at the very end.
Continuing the process to book #4 (say it also sells for $5.99 and earns $4.14), we are up to 90% of those who bought book #3 will buy book #4 (remembering they had to buy books #1 and #2 along the way.) The chances of that happening are 100% (book 1) x 50% (book 2) x 80% (book 3) x 90% (book 4) or 1.00 x .50 x .80 x .90 = .36 (or 36%). Again, our book #1 sale eventually leads that person to give us income on book #4 of $4.14 x .36 = $1.4904!
Add them together
Book 1 $3.44
Book 2 $1.77
Book 3 $1.656
Book 4 $1.4904
Total = $8.3564!!
Voila! A sale of book #1 that gives us $3.44 of income ultimately generates $8.36 of income. See what I mean about the magic beans of sell-though?
Remember our break-even analysis where we needed to sell 7.27 (rounded up to 8) books to earn back our $25 ad spend. Given the assumed sell-through rates, we only need 3 (actually 2.99, but again we’re ignoring the complexity of subscription service sales) book #1 sales eventually to break even because of the follow-on sales.
These add-on sales may not occur for some time. Sure, there are binge readers who will buy your first book, blast through it in a day or two, buy the second, third, and fourth, all before the month ends. More people will buy #1, stick it in their TBR pile, get to it eventually, and then maybe buy #2 right after they finish #1—or maybe later when they remember. You get the idea.
Great! How do I determine my read-through rates?
Unless your books have been out for some time and have had sales in the thousands or at least hundreds, you are probably going to have to guess. And even if you do have those higher levels of sale, you might still need to do some guesswork. The classic advice is to pick a “typical” period without significant variations from the norm because of promotions or new releases.
Sell-through from book #1 to book #2 equals the sales of book #2 divided by the sales of book #1. Sell-through from book #2 to book #3 works the same way, dividing book #3 sales for the period by book #2 sales for the period.
If you don’t have lots of sales, random fluctuations can skew your results. Let’s say you pick a month and discover you sold 10 copies of book #1, 3 copies of book #2, 4 copies of book #3, and 2 copies of book #4. Your sell-through rates are 30%, 133%, and 50%. You can’t use those: No way do some people buying book #2 suddenly go out and buy multiple copies of book #3. You just don’t have enough data to make reasonable estimates. Use a longer time period and add in common sense.
Other issues with sell-through rates
To butcher the old Master Card ad, reading my novels is a “priceless” experience. Despite that, readers are price sensitive. Give away book #1, it costs them seconds to download it. They may or may never read it, but let’s assume they do and like it. When they see you priced the next book at (in our example) $4.99, some will buy and others will turn to the next free book in their TBR pile. Someone who purchases book #1 is more likely to buy book #2 than someone who got book #1 for free. [Which is not to say you shouldn’t give book #1 away just to increase your sell-through rate. Assume books 1 & 2 are the same price (4.99 in our example); it turns out you earn the same money from giving away 1,000 books #1 with a 2% sell-through rate as you do selling 10 books #1 with a 100% sell-through rate!]
Your normal sell-through rate also decreases when you promote a book by temporarily reducing the price to $0.99.
Not everyone starts reading a series with book #1. Some people discover your series through your most recent book or one in the middle. If they like it, they may purchase book #1 to see how the series started. Authors often see increased sells of the early books in their series when they release a new series book. Not only will some people discover the series, but others who have read the series realize they forgot to get book #3 only when book #4 comes out; they see an ad for book #4 and buy book #3.
Okay – I give up
Holy mackers, I hear you say, this is soooooooooooo complicated, I want to give up.
Yep, me, too, and here’s a super secret (well, at least until I publish this blog). I don’t sell hundreds of books a month, so I spend little time trying to determine sell-through rates and how they affect profitability.
I know when I don’t advertise for a long period, my sales look like Wile E. Coyote’s trajectory after running off a cliff. Therefore, ads work. I track sales blips from distinct, short-term sales promotion campaigns (Free Booksy, Bargain Booksy, Bookbub, Fussy Librarian, etc.). Successful price promotions earn out almost immediately, and I know they will be even more profitable in the long run as the magic bean stalk grows.
For book giveaways, I pay careful attention to the cost per free book delivered, because it’s the only thing I can measure accurately. I assume I’ll get the same sell-through rates no matter how I get a free book in a reader’s hands. (That isn’t an accurate assumption because a bunch of cozy mystery enthusiasts are less likely to enjoy my books than are those who prefer a grittier read. I minimize the problem by carefully choosing the audiences for freebies.)
For other advertising venues (Amazon ads, Facebook ads, etc.) I do not measure individual profitability. I often have many versions of these ads running at the same time for multiple books. Because of the interrelationship in that environment between ads seen and purchases is impossible to know, I decide based on the metrics I can trust: number of impressions an ad generates, number of clicks, total ad dollars spent. I axe any ads that are clear losers; keep running ads that are clear winners until they become losers and let those in the middle go until I change advertising focus. I’ll keep doing that as long as I continue to make an overall profit from my books.
The real takeaway from the magic bean of sell-through: write and publish another outstanding book. I look forward to your questions and comments.
* * * * *
James M. Jackson authors the Seamus McCree series. Full of mystery and suspense, these thrillers explore financial crimes, family relationships, and what happens when they mix. Furthermore, a novella is the most recent addition to the series. You can sign up for his newsletter and find more information about Jim and his books at https://jamesmjackson.com.