The majority of manufacturers sell their products the same way: online through third party retail partners such as The Home Depot, Lowes, Walmart and many more. While these and other leading retailers typically invest heavily in website analytics and artificial intelligence (AI) driven recommendation engines, the manufacturers that produce and provide the products being sold often have little to no visibility into those analytics. This lack of competitive intelligence hampers manufacturers’ ability to effectively price and promote their products.
Not having the capability to glean meaningful insights from online retailers is problematic for manufacturers selling primarily online. While manufacturers are not completely in the dark—many online retailers provide weekly aggregate sell-through data that provides information about product sales—they are not privy to data about their competitive position within each product category. Without those insights, it’s difficult for manufacturers to understand the factors that drive customer behavior, including price, product features, and promotions. Manufacturers are left to guess which products are similar in the eyes of consumers, which product features are driving sales performance, whether or not pricing is negatively affecting sales, and more.
While manufacturers are left in the dark about these important points that so greatly affect profitability, retailers are often not. Their sophisticated analytics provide answers to the many questions manufacturers may be having; the problem is they aren’t sharing those insights for competitive reasons.
One major U.S. appliance manufacturer that has products sold online through The Home Depot and Lowes wanted to find a way to mine valuable data that would allow them to make more intelligent decisions regarding their product features, price-points and more.
The manufacturer reached out to Vertex Intelligence, a data science company, to see if network analysis and machine learning could be used to provide valuable insights directly to them without help from the online retailers that sell their products. Vertex Intelligence leveraged its expertise building complex web crawlers, or bots that can effectively index content on websites, to create an automated platform for harvesting data from millions of product recommendations from product pages across multiple e-retailers.
Vertex Intelligence leaders knew there were valuable insights hidden in the publicly-facing product recommendations shown on individual product pages (such as products listed under headings like “related products” or “people who viewed this also viewed this product”). On their own, individual product recommendations don’t tell manufacturers very much, but the aggregation of similar products from across pages and websites presented an invaluable source of competitive intelligence. This is because those seemingly-simple product recommendations are often created based on complex consumer browsing and sales behavior, social network analysis, neural networks and more.
With a large dataset of retailer product recommendations to work with, Vertex Intelligence was able to analyze relationships between all products and brands in a category simultaneously using network analysis and machine learning to effectively reverse-engineer the retailer’s recommendations and reveal a remarkable amount of detail about consumer behavior and preferences. The models created allowed Vertex Intelligence to make data-based predictions about the client’s competitors’ sales volume, top product features and more. To ensure accuracy, a regression analysis was performed essentially comparing the Vertex Intelligence model’s estimated sales performance data for the manufacturing client and the client’s actual data. The analysis showed the Vertex Intelligence model’s prediction was very accurate, confirming the model’s reliability.
Now that the system is in place, continual data collection allows the manufacturer to receive regular reports with up-to-date competitive intelligence data. The system—which was created specifically for The Home Depot and Lowes websites—could now be used for other manufacturers selling on the same retail websites, or could be replicated to analyze other retail websites.
In this case, the manufacturer is able to glean a number of data-points, including identifying:
- Underperforming product categories
- Opportunities for cross-promotion
- Which features drive competitor pricing
- Competing products (the model is also able to rank them based on likely sales volumes)
The manufacturer continues to use these reports to inform marketing and product development efforts.
Delivering an AI-driven Data Warehouse Migration Strategy
How to leverage AI for planning enterprise data warehouse migrations
Competitive Intelligence and Performance Optimization Models Provide Smarter Insights for Manufacturer Selling Products Online
How manufacturers can improve online retail sales performance with better competitive intelligence
Forecasting cash flow with predictive modeling
Accurate cash flow forecasting drives profitability in construction industry