Abstract:Studying competition and market structure at the product level instead of brand level can provide firms with insights on cannibalization and product line optimization. However, it is computationally challenging to analyze product-level competition for the millions of products available on e-commerce platforms. We introduce Product2Vec, a method based on the representation learning algorithm Word2Vec, to study product-level competition, when the number of products is large. The proposed model takes shopping baskets as inputs and, for every product, generates a low-dimensional embedding that preserves important product information. In order for the product embeddings to be useful for firm strategic decision making, we leverage economic theories and causal inference to propose two modifications to Word2Vec. First of all, we create two measures, complementarity and exchangeability, that allow us to determine whether product pairs are complements or substitutes. Second, we combine these vectors with random utility-based choice models to forecast demand. To accurately estimate price elasticities, i.e., how demand responds to changes in price, we modify Word2Vec by removing the influence of price from the product vectors. We show that, compared with state-of-the-art models, our approach is faster, and can produce more accurate demand forecasts and price elasticities.
What problem does this paper attempt to address?
### What problem does this paper attempt to solve?
This paper aims to solve the problem of studying market competition and structure at the product level rather than the brand level. Specifically, the paper focuses on how to handle the competitive relationships among millions of products on e - commerce platforms. Traditional analysis methods face huge computational challenges when dealing with such a large number of products. Therefore, the authors propose the **Product2Vec** method, a new method based on the representation - learning algorithm Word2Vec, for studying large - scale product - level competition.
#### Main problems and solutions:
1. **Computational challenges**:
- **Problem**: Traditional methods have difficulty dealing with the computational complexity of a large number of products.
- **Solution**: Use representation - learning techniques (such as Word2Vec), regard shopping baskets as sentences and products as words, and generate low - dimensional vector representations, thereby simplifying calculations and improving efficiency.
2. **Product relationship identification**:
- **Problem**: It is necessary to distinguish whether products are complementary or substitute products.
- **Solution**: Introduce two measurement indicators - complementarity and exchangeability. Through these indicators, the relationships between products can be more accurately identified.
3. **Demand forecasting and price elasticity estimation**:
- **Problem**: Accurately estimate the response of demand to price changes (i.e., price elasticity).
- **Solution**: Combine the random utility choice model and modify the Word2Vec model to remove the influence of price on product vectors, thereby more accurately estimating price elasticity.
4. **Market structure analysis**:
- **Problem**: Traditional methods have difficulty dealing with product competition across brands and categories.
- **Solution**: By using the Product2Vec model, the competitive relationships of thousands of products can be analyzed in a short time, revealing the competitive patterns across brands and categories.
5. **Automation and no prior assumptions**:
- **Problem**: Traditional methods usually require manual input or prior assumptions about market segmentation and structure.
- **Solution**: The Product2Vec method is fully automated, does not require any manual input or prior assumptions, and can automatically discover the market structure.
### Summary
By introducing the Product2Vec method, the paper solves the computational problems in conducting competition analysis in large - scale product data and provides effective methods to identify the complementarity and substitutability between products, thereby providing valuable insights for enterprises' strategic decisions (such as pricing, product design, and positioning). In addition, this method also performs well in demand forecasting and price elasticity estimation, outperforming existing advanced models.