Jargon Buster
At Data Matters UK, we pride ourselves on passing on the secrets of your data in plain English.
We particularly focus on demystifying all the analytical processes, particularly those which are involved in the modelling we’ll use to build your segmentation.
We want you to feel that you clearly understand those segments, their attributes, and how they should be used in real terms for your marketing.
To further the demystification process, here’s our Jargon-buster.
It aims to explain any unfamiliar terms you might run into on our website:
Analytics
Studying data in order to identify patterns which can help us to understand customers’ behaviour.
Churn
A general term which is applied to any kind of customer loss.
Clustering
A specific way of segmenting customers who share common characteristics. Clustering can group customers in many different ways but care needs to be taken to ensure that segments generated are useful for driving marketing opportunities.
Predictive Modelling
Analysis carried out with the prime purpose of predicting how individuals will act in the future with respect to any given criteria. For example, how valuable a customer are the likely to be, how likely are they to respond to a campaign, how likely are they to cease being a customer. These models are built by combining all information available about individuals, and how they have interacted with the company to make the best possible predictions.
Segmentation
A relevant and practical way to group customers or prospects in order to allow you to market to them appropriately. Segmentation may be based upon shared historical behaviour or predicted future behaviour. Segmentations may be fixed and reviewed after a fixed period of time, or they may be dynamic and customers may move segment in real time.
Profiling
A simple analysis that provides basic information, often demographic, about customers or a group of customers in order to describe the characteristics of the group. Often a profile will compare two groups of customers against each other or make comparisons to national figures.
Reactivation campaign
A marketing campaign designed specifically to engage lapsed or churned customers and encourage them to repurchase.
Transactional data
Information that comes directly from an order processing system. Usually transactional data will hold information about how much was paid, on what day for what products etc. The information may come from online retail sites, from tills in store or from purchase capturing systems.
Data Pools
A repository of similar information from different sources. For example. a combination of lists of prospects, or transactional data from a range of companies.
Warm data cells
Groups of prospects who whilst not being customers have shown some degree of interest in your company or products. This might have been through asking for information or have been recommended by an existing customer etc.
Customer journey
Recognition of the fact that all customers will have a different relationship with your organisation dependent on their own previous interactions with you. Some companies will have very common customer journeys for all customers in that they follow a similar path from acquisition through similar product purchases or donations, other companies will have many different journeys that customers may follow.
Responder modelling
A particular type of look-a-like modelling that finds other people in your database who “look like” the people who have responded to a particular campaign. Usually so that those people can be selected for the next campaign.
Look-A-Like modelling
Using advanced analytical and statistical techniques to find customers or prospects who share similar critical traits with a particular type of customers that you would like to have more of, or similarly reduce the number of. In this case looking alike need not mean demographically alike, but could relate to showing similar behaviour or actions which mean that they are likely to behave in the same way as the chosen target group.
RFV
A simple historical segmentation that refers to R – Recency, how recently someone has engaged with you, F –Frequency, how frequently someone has engaged with you, V-Valuable, how much value is associated with those interactions. This is a common type of segmentation but is restrictive in that it doesn’t necessarily predict future behaviour and so can reduce the effectiveness of future campaigns.
SPSS
A common statistical software tool used to carry out many of the analysis projects by Data Matters UK
SAS
A common statistical software tool, also usable as a database and data processing tool
Data scoring
Applying the results of models built or segmentations derived onto the database so that all individuals have their own “score” against a number of variables.
Linear regression
A statistical method that attempts to assign potential values for one variable based upon values on one or more other different variables. E.g. children’s height can be predicted by their age.
Decision Tree Analysis
A statistical method that is used to build predictive models and classifications by combining many variables in different combinations.
Selection models
A particular type of predictive model built to rank individuals as to their appropriateness for selection for a particular communication, campaign or product.
Acquisition data purchase
The ability to buy names and addresses of individuals or households that are not currently your customers for inclusion in prospect communications.