This session zooms in on the flow of products. Ways to improve forecasts by quantifying lost sales due to stock ruptures are explored, as ignoring these "missing sales" leads to historical mistakes being repeated. A good grip on this can greatly improve purchasing decisions, as well as the allocation of products to stores. This includes the complex challenge of getting to good forecasts on the level of individual sizes and channels. Analytical approaches to identify and reduce problematic return behavior are also covered in depth.

Price is often one of the easier things to adjust, but many retailers have difficulties to get beyond traditional "cost-plus" approaches, or stick to old rules of thumb. This is ill-suited to an online and fast-paced playing field. During this session participants learn how elasticity can be estimated using historical data, how competitor prices can be employed, how end of season markdowns can be approached more intelligently and what to consider when adopting dynamic pricing.

Marketing departments are often early adopters of data-driven tools. In spite of this, tangible improvements and measurable results remain hard to achieve. This module covers experimentation and marketing automation in a broad sense. A hands-on exercise shows how setting clear goals and avoiding investments in experiments with only a slim chance of success can make a big difference.

Marketing experimentation is perhaps the most frequently discussed domain for retail, but it is important to get the basics right before investing in more advanced applications. This session focusses on the foundations of algorithmic marketing - what needs to be understood and what must be in place to understand and influence what happens.

Specific topics include:

• Good versus bad customer segmentation practices

• The use of specific models for the timing of outreach to (repeat) customers

• Best practices for setting up experiments, and ways in which we can trick ourselves

Most retailers must manage networks of physical and digital stores (channels). This means deciding where goods will be made available, and whether to relocate inventory over time. The goal of this session is to provide a clear perspective on the nature of the problem, how the various aspects of the problem can be quantified and finally how the problem can be solved.

Specific topics include:

• Detecting significant over- and under-performance of products

• Translating a store network into a formal diagram

• Calculating a threshold for moving products

At the core of retailing is the decision which products to offer. Especially in retail domains where the product collection undergoes major seasonal changes this remains challenging.

Specific topics include:

• Estimating lost demand: What could you have sold under other circumstances?

• Dealing with size distributions

• Making more robust sales forecasts

• Making purchasing decisions using probabilities

While most professionals and graduates have a basic understanding of price and elasticity, this often proves to be a hard concept to apply in practice. The first evening in this series discusses how retail pricing typically works in practice, and how this can be improved.

Specific topics include:

• Realistic demand models that can be applied to a product portfolio

• Setting the list price / sticker price of products

• Dynamic pricing - i.e. moving away from a single static price point

• Discounts and markdown pricing