A Browse Through My Bookshelf: Leverage Moneyball Publishing to Design Your Own Success Algorithm

Michael Lewis’s Moneyball tells the story of the Oakland Athletics baseball team and their General Manager, Billy Beane. The book illustrates how the team used statistical analysis, rather than traditional scouting methods, to acquire new players. This novel approach allowed the Athletics, a team with a small budget, to remain competitive against teams with far more financial resources.

Most people know of the movie with Brad Pitt playing Billy Beane.

No baseball team operates now without the methodology. They essentially can’t compete without it. It is like the pricing algorithms used by airlines to fill airplanes. Without the work of pricing strategy, they would all go belly up.

Adopting a Moneyball approach in publishing

We have watched book sales continue to grow, with indies claiming more and more success while traditional publishers struggle and consolidate.

I believe the future of publishing lies in a Moneyball strategy. If imprints are to exist in the future, they will provide services that authors choose not to do for themselves and will use data to identify emerging authors to partner with.

Let me unpack this for you because the Moneyball approach will shape our industry even if you never want to use a publisher.

For too long, traditional publishing has used a portfolio model. Like venture capitalists, publishers spread money across authors, hoping that diversification would result in wins covering the losses for an above-average return.

Publishers, like movie studios, are in an unwinnable squeeze because they are subject to editors’ choices on who to publish and management’s need to increase profits.

By leaving decisions to humans, they make ill-informed choices. Over time, editors focus only on what meets management’s needs. If you wonder why, just reread the last few articles. Studios and publishers then choose reboots, sequels, or blockbuster authors because they are at the top of the Pareto curve and guarantee higher sales.

This concentration of choice results in paying a premium and finding it increasingly difficult for the winners to pay off, let alone have the profits to cover the losers. If you don’t believe me, look at the testimony from the Simon & Schuster anti-trust trial. The case is significant evidence supporting my hypothesis. The case was brought by the US government against two of the biggest publishers trying to merge to cope with the burdens of leveraged buyouts and a changing market where indies are consuming most of the growth.

The government filed an anti-trust case because the merger would harm the auction market for the top 0.1% of authors. I guess if presidents are going to continue to get multi-million dollar advances on books, that market needs to remain “fair.” The biggest irony for me was that the publishers didn’t defend themselves with the obvious. More top-selling books in the market were self-published than ever before when the trial took place in 2022. The companies involved didn’t have a monopoly because some authors were choosing to publish without them and doing better than those they published.

The real issue is that in a market like publishing that sorts itself out on a power law curve, the easy targets are the few authors at the top. All eyes are on the apex, and just like the sales concentrate, so does the attention of publishers. This is no different from what happens in sports markets like the NFL draft.

Did you know Tom Brady was the 199th pick in the 2000 NFL Draft? Even New England passed on him six times.

Thaler, the Nobel Prize-winning author of Nudge, whom I discussed a few articles back, wrote a paper on the irrationality and inefficiencies of the NFL draft. Baseball teams, football teams, and publishers will overpay for named talent.

Not all great players come from the top of the draft. The two best quarterbacks in recent years, Peyton Manning and Tom Brady, are cases in point. Manning was taken with the first pick in the draft, but Brady was taken 199th. Also, as we showed in the opportunity-cost analysis above, trading down to get more players does not reduce the chance of getting top players. ~ Massey & Thaler

The smart thing to do in the NFL draft is to trade your number one pick for more picks lower in the draft. You also need to have a system to identify emerging talent.

The problem in publishing is that finding upcoming talent has been nearly impossible. Go back to the trial testimony and read about the publisher’s admissions about how few advances pay out. This is where Moneyballpublishing will change the game.

The game is to get above-industry-average results with lower entry costs.

Rather than hiring acquisition editors, I would hire data scientists and programmers to create an algorithm designed to find upcoming authors and get more of them. Just like Billy Bean stacked his team with players with higher runs batted in ratings, I would stack the portfolio with the highest emerging author index in genres that I had identified with breakout potential.

Here are a few parts of the index I would want to include:

1. Sales Data:

• Initial book sales numbers.
• Sales growth rate over time.
• Pre-order numbers increasing over a series.

2. Reader Engagement:

• Number and quality of reviews on platforms like Amazon, Goodreads, and other book review sites.
• Average rating of books.

3. Market Trends:

• Genre popularity and trends.
• Alignment of the author’s work with current best-selling genres or themes.
• Demographic data of the author’s readership.

4. Competitor Analysis:

• Comparison with other emerging authors in the same genre.
• Benchmarking against novels that have turned into best-sellers.
• Performance of past books.
• Trends in genre.

5. Creative Output:

• Frequency and consistency of new releases.

Backtesting an algorithm designed to predict future best-selling authors involves simulating how the algorithm would have performed using historical data. This would take time to test across rolling windows and validate the system’s picks against historical outcomes, but when done, I would know the probability of picking winners.

By aggregating and analyzing these data points, the algorithm can identify emerging authors with the highest potential for future success. Machine learning models can be trained to predict best-selling potential by learning from historical data of past successful authors and applying these patterns to new, emerging talent. Even if I continued to use editors to acquire titles, only picking from those that ranked higher in my index should produce above-average results.

The first-mover advantage

Here’s the horrible part facing publishers: Just like pricing models for airlines or statistical models for player picks, the first-movers get tremendous value for a short time. Then, that advantage is lost because everyone uses the same methods. Sure, you can tweak your model, and maybe that gives you some temporary advantage, but the data arms races will continue to escalate.

Those of you who do get after this now will have a first-mover advantage for a significant period and likely do even better when the old guard acquires you to get your proprietary model.

Few readers of this are planning to start a publishing company, but most of you are looking to become successful authors who get big advances or rake in royalties as indies. Take what I’ve laid out here and use it as your blueprint for success. If you agree with me, then the better you are at doing what would be in the index, the easier it will be for you to either attract a Moneyball publisher or break out on your own.

How can you use these ideas to design your own success algorithm?

Are you looking at performance indicators that show a compounding of results and driving cumulative of advantage?

A little aside with regard to generative AI. For all of you focusing on its use in books and art, you’re missing the forest for the trees. The biggest impact will not be how authors use it to write books but how publishers and indie authors use it to market and identify sales opportunities. What I described in today’s article is possible now with AI and scraping tools to gather the necessary data.