Product Network Analysis

PNA. Self-organizing, customer-driven product bundles and categories

Number of different products sold in a supermarket, an e-commerce site or any retailer can easily go up to tens or hundreds of thousands. While each of these products have a buyer, not all of them are relevant to every single buyer. ‘Buyers of this also bought’ type recommendation engines work well for tactical marketing actions, but mostly fall short in helping improve category management or more structured bundle management.

With our Firestarter "PNA", we detect natural groups of products and relations between each product based on customer preferences. Using these networks, we provide unique insights for improved next best offer development, category management, bundle creation and even inventory management.

Increase cross-sales and basket value

identifying cross-category product sets for improved next best offer

Optimize product assortment

of every physical store based on popular product networks of people living in the vicinity

Improve product bundles

based on relations between individual products and brands

Increase campaign performances

predicting items customers are most likely to consider even if they never bought from the category before

PNA is an analytics suite with algorithms based on social network analysis for product and category management domains.
Typical deployment time is 1 week.
  • Product Network Maps: Micro grouping of all products and relations between each, based on purchase behavior of customers
  • Product Network Dashboards: Automated dashboards presenting characteristics of each network and product
  • Bundled Offer Engine: Bundle recommendation engine, presenting bundled offer ideas based on a specific product or a customer segment
  • Portfolio Roles: Roles of significant products - such as “Hook Product” or “Add-On Product” - identified for each network
Sneak Peek

"Product Network #145 consists of around 200 products frequently purchased by single young professionals living in a central location downtown"

"Once a customer buys the hook product in network #56, his likelihood to buy the rest of the products increases by 12%"

"Based on historical customer preferences, our store in the school district should keep higher levels of stock for product networks #116 and #168"

"There are 3 potential bundle alternatives to introduce the New Parents segment to their least likely purchase categories"