11. April 2018

Are you a mind reader?


Are millennial men the right target group for a Pampers marketing campaign? Of course - if they have children. Anticipating what the customer wants is like mind reading. This is now possible by analyzing customer behavior and situations.

More and more products are being developed at an ever-increasing pace thanks to digitalization. The associated disadvantage for the customer is that this makes it increasingly difficult to choose. Marketing therefore needs to use new strategies and tools to support customers in the decision-making process. In the ideal scenario, the customer is offered a customized and personalized product that is exactly what they want or need.

Successful sales and marketing departments work with business intelligence (BI) tools and technologies that allow them to transform customer data into useful information. This is used to acquire new customers and maintain existing ones, for example, when developing personalized marketing campaigns, determining personalized prices, identifying attempts at fraud or predicting churn rates. Creating meaningful customer profiles is the key technique. The more diverse the data source, the more meaningful the profile.

Gathering relevant customer data

Data sources can be roughly divided into three categories: customer channels, internal systems and external systems. Customer channels, such as company websites or online shops, are the primary points of contact that help us to understand customer preferences and interests. Internal systems are less customer focused. They reveal purchasing history or the nature of the relationship between the company and the customer, for example. External systems, such as social media, provide a broader view of the customer and give their actions a wider context.

Handling a variety of sources requires data to be integrated into analysis or BI platforms in an application-specific manner. Customer channels are usually connected in real time to trigger actions during the user session. Other data sources require on-demand integration or import via batch processing, for example. Whatever the integration method, there are two crucial points: Firstly, it needs to be possible to associate processed information directly with a customer. Secondly, data must be collected as soon as possible using an incremental approach. In this way, initial results can be obtained quickly and results can beconstantly improved using agile methods.

Customer profiles comprise all customer-related data. Groups (clusters) are formed from customers with corresponding characteristics, such as similar goals or the same marital status. These are described using virtual representatives, known as personas. They include names, photos and backstories. Example: To identify the relationships between different purchases, a BI specialist uses an affinity analysis to examine an online shop’s entire sales. He observes that a significant number of purchases include Pampers and beer. These sales correspond to the core characteristics of “Dave”, the persona of a young father who shops frequently. To simplify this persona’s user journey, Pampers are promoted automatically if beer is added to the shopping cart. But is Dave actually a father? It is necessary to incorporate several data sources, including social media, to get an unequivocal answer to this question.

From customer profiles to customer clusters

A customer profile is essentially nothing more than a multidimensional vector, with each data source adding a further dimension. A range of techniques and algorithms are needed to identify clusters from these vectors. Finding the right mix of algorithms and techniques, and setting their parameters, is an iterative process. The k-means algorithm is one of the most widely used. In this process, observations (an n-dimensional vector) are assigned to clusters on a continuous basis. These observations might be customer interests identified via customer channels, purchasing patterns or level of income, or the internally identified risk type. However, they also include parameters, such as travel habits or marital status, which are identified via external systems and are not directly related to the company.

Dave’s persona becomes more defined while experimenting with this data. It is possible to deduce that he is trying to save money from his middle-class income to secure his children's financial future. Therefore, a long-term risk-free investment product or life insurance would interest him.

From a clear picture to successful actions

Once significant clusters have been identified, products and services suitable for these customer clusters need to be determined. Marketing campaigns are designed using all the information about clusters and the corresponding customers, and these are tested and verified using A/B testing. Identifying test customers to validate different actions is at the heart of A/B testing. Group A receives an email newsletter with general information about life insurance. Group B receives a personalized email. Group C is shown an investment idea on their preferred channel. At the end of the test cycle, analysis will determine which action was the most successful (new life insurance policies taken out).

Are you ready for the next step?

Therefore, from a technical point of view, mind reading is already possible today. It is an iterative process that consists of collecting relevant user data, segmenting customers into clusters, developing and validating marketing actions, resulting in a global rollout. Modern technologies and exploratory approaches will support you to design the simplest process possible, allowing you to read your customers’ minds.

 

Need more reading material? This article was originally published in the 2017 edition of the ti&m special titled "Our digital identities - let's get personal". You can download the entire magazine for free here.


Pic
Dorian Tanase

Dr. Dorian Tanase has more than 15 years of IT experience. His current focus is on big data and analytics solutions. He graduated from ETH Zurich with a degree in complex systems.

Ähnliche Artikel

ManMachine: Building an Efficient Chat Tool Prototype

Artificial Intelligence (AI) has been a major theme in the last decade and numerous big companies have invested a lot of effort into the technology. Within the scope of our last ti&m garage project, we too developed a small but efficient chat tool prototype for a big company in Switzerland.

Mehr erfahren
Unsere Digitale Identität – Wege aus dem Einheitsbrei

Mehr erfahren
Big Data - Many Talk About It, We Do It!

We’ve been working on big data topics in our labs and with our clients for quite a while now. Over time, we built a framework of technologies and utilities we can build data driven projects on. We call it ti&m analytics.

Mehr erfahren
Es lebe der Spieltrieb!

Mit Elementen aus der Game-Industrie experimentieren nicht nur Wissenschaftler. Auch Marketingabteilungen versuchen, die Kundenbindung mit spielerischen Mitteln zu steigern.

Mehr erfahren
Impressionen von der ersten App Builders Konferenz der Schweiz

Die Schweiz hat mit der App Builders Konferenz einmal mehr bewiesen, dass sie ein iOS-Land ist. In diesem Artikel geht es um die Impressionen der „App Builders Switzerland 2016“, der ersten Schweizer Konferenz von Entwicklern für Entwickler in Europa.

Mehr erfahren