B2C: SKU optimisation

Background

The focus of this ad-hoc research is on the problem of optimisation of SKU assortment in order to maximise the margin of full category.

Hierarchy of questions:

  • What products can be excluded from portfolio with positive impact on total category IGM (Integral Gross Margin)?
  • Where to consumers will switch if a product disappears from the market? How they choose the close-by product?
  • Distances between products in the decision making feature space play the role. How to estimate these distances?
  • Based on the market share, to apply conjoint-alike approach to derive utilities (importance) of features. How to prepare and process required datasets?
  • Conclusions are based on the collected total attribute matrix, and market share data. What are the sources?

Our hypothesis on consumer behaviour

How consumers choose and react to various competing alternatives?

  • After collecting information, they make comparison and evaluation among them [3-rd step of customer journey]. The evaluation criteria are based on the various product attributes (brand, color, size)…
  • … and we want to understand the utility (importance) that consumers expect to obtain from those attributes (or, in math terms, weights of variables)

Data collection

20 years ago, a manufacturer would go to market research agency and they together conduct a comprehensive market research survey. Modern digital technology allows doing such study without surveying consumers. In principal, all the data is available. 

  • There are regular and detailed reports on sales. These reports available from major agencies include global coverage of all consumer product sales (counts) together with average sales price (ASP) - how much consumers pay.
  • All product features can be obtained using web scraping from major web-shops. They include product names and attribute vectors.
  • Integral Gross Margin is important for "our" products only, thus can be obtained from internal financial systems.

This chart below is just an example available on a web (here). It demonstrates the positioning of "our" products and competitor ones in the given space of product attributes.

Based on this map, it is easy to understand three important facts:

  • What areas of the map are too dense of our products. This actually means that such products cannibalise each other: it is too difficult for a consumer to make a decision. These are the first candidates for the exclusion from assortment.
  • What areas have competitive products and do not have ours? This means, we do not cover the whole potential universe of features.
  • Where we have a close-by competitors? If we exclude such product, even the one with low IGM, we loose market share, because our clients immediately switch to alternative offer.

IGM

Before we make a decision on product exclusion from assortment, it is useful to check what would be the impact on overall IGM for the full category. It is possible to estimate this impact, look at this chart:

It is easy to notice, that some of the products can be excluded and overall impact will be positive. However, we have to remember the previous map and keep in mind competition and strategic plans.

Importance (utility) of various features

One of the key outcomes of this project is in estimation of utility of various features of the products for consumers (of a given country). These utilities can be represented at a chart like this:

 

Implementation of the results helped the company to directly increase IGM and also had indirect impact: the reduce of stock of spare parts, slight reduction of marketing budgets, etc. 

It was decided to conduct this study every year to align the course.