Even when performing the same data analysis work, why do such positions offer an annual salary of nearly 200,000 US dollars?
Recently, a specialized job position has emerged in the global technology industry: Pricing Scientist / Applied Economist.
According to Amazon's official recruitment JD, as well as job salary statistics from BuiltIn and Levels.fyi, this type of position that focuses on causal inference to develop pricing strategies generally offers a base annual salary (excluding stocks and bonuses) of 120,000 to 180,000 USD.
This high-paying position is not limited to simple sales forecasting. Instead, in scenarios where large-scale price A/B tests cannot be carried out, it uses methods such as DID, instrumental variables, synthetic control, and CATE heterogeneous treatment effect to address price endogeneity confounding, estimate the real price elasticity from historical observational data, and support decision-making on price adjustment, discounts and subsidies.
Ordinary prediction models can only answer: If the price is adjusted to X, what will the sales volume probably be?
What pricing positions can answer is: If I actively raise or lower the price, what kind of real changes will be brought to sales volume, profit and user retention?
Domestic enterprises including JD, DiDi, Freshippo and Meituan have also applied this set of ideas to businesses such as commodity pricing, subsidy strategy and merchant commission evaluation, though the relevant positions are more scattered.
The core capability of this high-paying position is no longer traditional data forecasting, but the business decision-making capability based on causal econometrics. No matter for future job promotion or improving business data analysis thinking, causal inference and econometrics have become core competitive advantages in the fields of pricing, revenue strategy and business empirical research.
To build up this knowledge system, you can make gradual progress from causal intuition, econometric basics, software practice to advanced actual combat.
✅ Stage 1: Build causal intuition from zero foundation
When you first get in touch with data analysis, you may be confused by various reports and cannot distinguish causal relationships from data coincidences, and the formulas in thick textbooks are daunting.
This book avoids complex mathematical derivations, explains the core logic of causality with business and real-world cases, so that you can develop the judgment to identify analytical conclusions without delving into mathematics, which is very suitable for undergraduates and business analysts to develop causal thinking.
✅ Stage 2: Get started with econometrics, lay a solid foundation for regression and empirical research
Many people are eager to get started with trendy methods such as DID and instrumental variables, skipping the basics of econometrics directly. In the end, they only remember a bunch of terms, but do not know the prerequisites for using these methods, which makes it easy to make mistakes in actual analysis.
This is an indispensable book for empirical research and business strategy analysis. It does not focus on tedious derivations, but focuses on thoroughly explaining endogeneity, instrumental variables, and panel data, which are the most essential foundations for causal identification.
It is suitable for undergraduates who have been exposed to basic econometrics, scientific research students, and workplace practitioners engaged in industry research and business strategy. Many overseas applied economics related positions also take this book as the knowledge foundation.
If you want to learn econometrics from scratch, you can start with this book. It is a classic domestic textbook with friendly difficulty level. It explains clearly the core contents such as OLS regression, panel data and time series, and comes with EViews operation guidance. You can go through the underlying concepts steadily, which will make it much easier to learn causal inference later. Suitable for beginners to consolidate basic econometric skills.
If you mix up basic concepts such as regression and hypothesis testing in analysis, you can quickly review them by flipping through this book. It uses popular language with plenty of examples, which is very convenient for reference whenever you get stuck. You don't need to read every chapter from beginning to end, and it is more suitable to be placed at hand as a reference book.
✅ Stage 3: Software practice, implement and run econometric methods
You have understood all the theoretical principles, but when you get the real data, you don't know how to run the model practically.
You can learn it even if you have never used Stata before. It covers commonly used tools such as regression, instrumental variables, and panel models, and is equipped with public data sets, so that you can manually reproduce common analyses such as DID, filling the gap from understanding theories to hands-on practice. It is applicable for both students and workplace data practitioners.
✅ Stage 4: Modern causal inference, oriented to empirical research and business practice
It is difficult to implement complete A/B tests in actual work, and most of the evaluation of price adjustment and subsidy effects can only rely on historical data.
This book is closely related to business scenarios, covering a full set of methods such as DAG, DID, and synthetic control method, with ready-made Stata and R codes. It can handle price endogeneity interference, evaluate business interventions, and distinguish the effects on different groups of people, suitable for readers who already have a foundation in econometrics and need to implement business analysis.
Mastering econometrics and causal inference has long been an invisible threshold for high-end business strategy positions. Only by breaking out of the limitation of data analysis that only focuses on prediction can you truly have the core competitiveness to implement business decisions.
END
This article is from the WeChat official account "crup Economics", authorized for release by 36Kr.