Product Matching: Explained | eCommerce Matters Ep. 051

Listen Now

Episode 51 of the eCommerce Matter's Podcast is now available, entitled ‘Product Matching: Explained’.

At BlackCurve, we often get asked: "how do you automatically find matches so quickly?". Well, we’re going to tell you how!

We won’t give the crown jewels away, but a bit of a tease. Demystify the process. We will cover:

What is product matching?
Why getting it right is so important.
Different product matching methods.
The future of product matching.

Hosted by Philip Huthwaite, CEO & Founder of BlackCurve, and Rob Horton, Chief Product Officer of BlackCurve.

Where to listen

🎧 Available to listen now on all other major podcasting platforms.

🎧 Available to listen now on Apple.

🎧 Available to listen now on Spotify.

Podcast Summary

Introduction

This podcast episode delves into the significance of accurate product matching in e-commerce, particularly for pricing decisions. It explores the challenges associated with manual matching, the importance of reliable data, and the advantages of precise matches in a competitive market.

Challenges of Competitor-Sourced Data

Incorrect GTN Matches

One challenge arises when competitors surface incorrect Global Trade Numbers (GTN), leading to inaccurate product matches. This problem is further exacerbated when both competitors and Google categorise products incorrectly on their websites.

Ensuring Data Accuracy

To address these challenges, it is recommended to employ pricing technology that can collect data at scale. This reduces reliance on potentially flawed competitor data. Additionally, using data points such as GTIN, brand, and product codes helps ensure accurate matches based on the available information.

Adding Another Layer

To enhance product matching accuracy, Blackcurve recommends using their tool, which adds an extra layer of support. This includes excluding outliers, such as competitor pricing, to ensure more precise decision-making. By disregarding extreme data points, pricing decisions can be based on the most accurate information available.

Pros and Cons of Manual Product Matching

Hand Product Matching Process

The manual process involves copying and pasting GTIN, product names, and prices from Google into spreadsheets. Initially, there may be a high degree of accuracy due to the e-commerce manager's familiarity with the products and their visual validation.

Potential Errors and Limitations

However, the manual matching process is prone to human errors. Distractions, fatigue, or oversight can lead to mistakes, particularly when dealing with numerous product variants. Over time, the accuracy of manual matching declines, impacting productivity and decision-making quality.

Methods of Product Matching

Rule-Based Approach:

The rule-based approach utilises heuristics and predefined rules to determine matches based on specific conditions. This method provides the advantage of being interrogable, enabling better understanding and transparency in the matching process.

Machine Learning (ML) Approach

ML approaches involve training neural networks using known matches and non-matches to classify new instances. However, ML models can sometimes lack interpretability, making it difficult to understand why certain matching decisions are made. Occasionally unreliable results can also occur due to edge cases in the training data.

Blending Rules and ML

To overcome the limitations of individual approaches, many businesses combine rule-based methods with ML models. This hybrid approach allows for better matching accuracy while retaining the transparency and control provided by rules. ML models handle complex patterns, while rules ensure interpretability and guide the matching process.

Benefits of Accurate Product Matching for E-commerce Retailers

Precise Pricing Decisions

Accurate product matching facilitates informed pricing decisions in a competitive market. Identifying underpriced or overpriced products compared to competitors helps optimise pricing strategies, maximising profitability while remaining competitive.

Improved Conversion Rates and Market Share

Precise pricing based on accurate data enhances conversion rates and market share. By adjusting prices to align with market standards, retailers can increase visibility on Google and attract more customers. The ability to swiftly adapt pricing based on reliable product matching provides a competitive edge.

The Future of Product Matching in E-commerce

Increasing Simplification

Google's investment in aggregating and improving data within its ecosystem simplifies the product matching process. Access to accurate data from Google Merchant Center reduces the need for direct scraping from competitor websites, streamlining the matching process.

Market Forces and Dominance of Google

The COVID-19 pandemic accelerated the shift to online sales, making accurate product matching even more critical. Google's dominance in search and advertising, particularly through Google Shopping, reinforces the need for reliable data to gain a competitive advantage.

Customization and Expanded Applications

The future of product matching lies in customisation for specific use cases and matching products from various sources beyond competitor websites. ML models will continue to evolve, enabling better classification and matching of diverse product information, such as data from marketplaces or store images.

Conclusion

Accurate product matching plays a crucial role in making informed pricing decisions in e-commerce. Overcoming challenges associated with competitor-sourced data and manual matching requires employing professionals, utilising relevant data points, and leveraging tools like Blackcurve. The combination of rule-based approaches and ML models, along with Google's dominance, contributes to the future growth and simplification of product matching in the industry.