Can You Predict a New Book’s First‑Month Sales with Simple Math Modeling?

How can we predict the sales of the upcoming book Smart Use of ChatGPT for Mathematical Modeling ? Sales depend not only on the book’s content quality but also on marketing strategies, target‑group coverage, and details such as giveaways.

Mathematical modeling offers a systematic way to break down the problem, analyze it step by step, and arrive at a reasonable forecast.

In a previous article I introduced an initial idea; this time I build a framework to estimate the first‑month sales of my own book using variables, assumptions, and formulas.

Basic Sales Framework

Book sales can be split into two parts:

Natural sales – the volume generated without any special marketing measures, coming from platforms such as JD.com, Taobao, Pinduoduo, etc.

Incremental sales – the additional volume driven by marketing activities.

We need to estimate each part separately, combine realistic assumptions and data, and obtain a total sales range.

Estimating Natural Sales

Natural sales mainly stem from the target audience and the baseline market performance. For a mathematics‑modeling book, the core readers are students participating in university modeling contests, modeling coaches, and enthusiasts.

Assumptions:

The number of university students participating in national modeling contests in 2024 is roughly X (high‑school contests are much smaller, in the thousands).

The probability of reaching these target readers is Y, because the audience is niche and promotion requires specific channels.

The average purchase rate is Z, as motivation to buy modeling books is generally low unless the reader is a contestant or enthusiast.

Plugging the data into the model yields an initial natural‑sales estimate of 100 copies .

Predicting Marketing Incremental Sales

Marketing increment is crucial for a niche book. The model considers three common tactics: giveaways, advertising, and word‑of‑mouth.

1. Impact of Giveaways

Giveaways target high‑influence individuals (coaches, bloggers) to generate indirect promotion. The incremental sales from giveaways can be expressed as:

Increment = (Number of books given) × (Response rate of each target group) × (Influence size of each group) × (Conversion efficiency).

Assumed parameters:

100 books are distributed as giveaways, divided into three categories:

Substituting these parameters into the formula gives an estimated incremental sales of 638 copies from giveaways.

2. Advertising Impact

Advertising (e.g., platform recommendations, social‑media promotion) can increase exposure, but no concrete plan is provided here, so this component is omitted.

3. Word‑of‑Mouth Influence

Word‑of‑mouth provides a long‑term effect. Assuming each 10 purchasers generate 1 additional buyer, the extra sales are calculated as:

Assumed parameter: each 10 buyers bring 1 new buyer.

Using the natural sales and giveaway increment, the word‑of‑mouth contribution is estimated at 74 copies .

The total marketing increment (giveaway + word‑of‑mouth) therefore equals 638 + 74 = 712 copies.

Total Sales Prediction

The overall sales combine natural and marketing components:

Natural sales (100) + Marketing increment (712) = 812 copies predicted for the first month.

Sensitivity Analysis

To assess the robustness of the forecast, key parameters such as response rate, word‑of‑mouth coefficient, and other assumptions are varied. The analysis shows how the total sales range shifts with these fluctuations, providing a reasonable prediction interval.

Through this mathematical modeling exercise, we obtain an approximate first‑month sales figure of 812 copies , with contributions broken down as follows:

Natural sales: 100 copies

Giveaway‑driven increment: 638 copies

Word‑of‑mouth increment: 74 copies

This prediction offers a basic understanding of the book’s market performance and supplies quantitative guidance for future marketing strategies. The exercise also illustrates the practical value of mathematical modeling: extracting core variables from a complex problem, forming assumptions, iteratively refining the model, and arriving at actionable insights.

Note that the model is preliminary; many factors are omitted and parameters need careful calibration. It remains an interesting practice awaiting real‑world validation.

What do you think of this prediction model? Feel free to comment and share your own insights on mathematical modeling.