Recommendation Systems - Part 1: Motivation and Basics

Recommendation systems are core elements of digital business models. This blog addresses two key aspects:

  1. For whom is a recommendation system relevant, and why?
  2. What are the basic variants, and how complicated is the implementation?

In this article, I’ll give you an overview, while my colleague, Josef Bauer, will delve into details in part two.

What does a recommendation system do?

A recommendation system addresses the interests of individual consumers. Conversely, it prevents customers from getting frustrated or overwhelmed to the point that they interrupt the customer journey. Hence, recommendation systems are the key for a big line of goods and services. Now, it is important to consider how customers use the services: whereas some online pages invite one to scroll and look around, others concentrate on helping the customer find the right product quickly.

Would my enterprise benefit from a recommendation system?

Recommendation systems are most useful for digital B2C business models, because that’s where they can directly influence the outcome of crucial KPIs and can strongly boost consumer experiences. Moreover, good recommendations are increasingly becoming the standard feature expected by consumers. Optimizing an existing recommendation system can also help utilize any unused potential.

In addition to their primary benefits, recommendation systems offer a host of ancillary benefits:

  • The teams involved can better understand the data, which forces a conscientious analysis of actual consumer behavior and preferences
  • Enterprises can use the observations to optimize both their offerings and user experiences
  • They can be used as a supplementary tool to try and control some of the traffic

Recommendation systems come in two types, as illustrated below, but the details define how well one can actually integrate this instrument.

Option 1: Item-based approach

Option 1 relates covers available products. The consumer sees products that are similar to those being sought. However, similarity can have many meanings in this context: e.g., product categories, price, or possibly the manufacturer. Naturally, a combination of various dimensions is also feasible.

Option 2: User-based collaborative filtering

Option 1 relates covers available products. The consumer sees products that are similar to those being sought. However, similarity can have many meanings in this context: e.g., product categories, price, or possibly the manufacturer. Naturally, a combination of various dimensions is also feasible.

Option 2: Consumer-based approach

Option 2 does not aim for similarities in products, but instead among consumers. The resulting recommendations may strongly reflect those under option 1, or not at all. This approach strives to model preferences and reasoning behind the consumer’s buying decisions, and less so the objective similarities between products.

Which option best suits my situation?

These days, legal restrictions on data access under the EU’s General Data Protection Regulation (GDPR) significantly affect one’s choices. Nevertheless, the item-based approach is an option, since one can set it up as a table and update it regularly. This makes integration in mobile websites and apps much easier. The consumer-based approach offers greater potential, depending markedly on whether you have a reliable streaming infrastructure.

What are the first steps?

To set up a recommendation system, it is necessary to assess your status quo on three fronts:

  • How closely is your business model connected with a recommendation system? Is it a key part or just an add-on? How large is the actual added value for the consumer?
  • How easily would it be to integrate a recommendation system into the existing infrastructure? Are the necessary data available, and could they be used for this purpose?
  • Does your enterprise have the required data science know-how? Do you have a team ready to take on this task, or would the needed resources have to be acquired?

Once you’ve answered these three questions thoroughly, the work can begin. For more details, please read what Josef Bauer has to say.

Views: 0
clear