EuroDirect is part of the Callcredit Information Group alongside GMAP Consulting, Callcredit, Legatio, DecisionMetrics & Broadsystem

Registered Office Details

Marketing Analysis & Modelling

A New Data Science That Delivers Exceptional Results  

The EuroDirect Marketing Analytics team of professional analysts has over 25 years experience of statistical modelling and geographical analysis across most industry sectors. Many of the team have been trained on the specific use of statistics for marketing analysis and have experience of a wide range of multivariate modelling techniques, including:

• Decision Trees - CHAID and CART
• Regression Analysis – both linear and logistic
• Cluster Analysis – k-means, Wards method
• Neural Networks – supervised and unsupervised Kohonen networks

The team are all degree qualified in a numerical discipline such as maths, geography or statistics. This core group is focused on the delivery of client solutions, but can also draw on the expertise of PhD qualified statisticians who work on product development initiatives. There is additional analytical expertise within the sister companies of GMAP Consulting and Callcredit, which can provide extra resource and domain knowledge for projects involving retail network planning and credit risk analysis. The Skipton Information Group encourages collaboration between analytical teams and this is easy to achieve as all three companies are based in the same building in Leeds. 

A key strength of the EuroDirect Marketing Analytics team is their expertise in SQL programming. This skill enables the team to manage model deployment across relational databases that are held in SQL Server. This means there is no reliance on IT or Production departments for the implementation of scorecards and classification models. Instead the whole process of model development and deployment can all be controlled by the marketing analysts.

Analytical Methodologies and Value Propositions

EuroDirect provide the following analytical services under the Forensic Analytics brand. These range from simple profiling through to complex behavioural modelling projects to predict purchase propensity and churn. 

• Profiling Analysis - Comparing customer groups or profiling customers against a national or regional base across a range of geo-demographic and lifestyle characteristics 
  
  The starting point for most customer analysis projects is a profiling exercise to identify the key characteristics of a client’s customer file. The international range of CAMEO classifications provide a quick and easy way of gaining customer insight from profile analysis. EuroDirect analysts are skilled in the geo-coding of these classification products to maximise the coverage and accuracy of CAMEO data across a client file.

In the UK, more in-depth individual level profiles can be produced by comparing a customer file against a national or regional base to identify the most discriminating characteristics from more than 350 variables held on Data Exchange – EuroDirect’s UK consumer universe file of 42 million consumer records. 

• Prospect Scoring - Bespoke targeting models to optimise customer acquisition and cross-sell
  
  This is the natural next step after profiling to combine predictive variables into a multivariate model using regression or decision tree techniques. Models are optimised by layering geographic, household and individual level data and separate models are developed for the same outcome using natural sub-population splits. This avoids the limitations of building a single ‘one size fits all’ model for the non-existent ‘average’ customer. Modelling within discrete sub-populations caters for the fact that independent variables predict the same outcome differently for separate groups of similar customers. 
  
  
• Campaign Evaluation - Response analysis by list source and profiling and modelling of response and/or conversion
  
  EuroDirect has also developed a fully integrated end-to-end methodology to optimise campaign evaluation. This process harnesses EuroDirect’s legendary data processing capability, with intelligent modelling and extensive consumer data assets to help clients achieve a 25% increase in response rate. The analytical methodology involves sub-population splits and a layered modelling approach to develop an integrated set of brand affinity, response and conversion models. These can be deployed back onto the clients own customer database or applied to 3 rd party bought-in lists to select the best prospects.
  
  A core component of the Forensic Analytics Campaign Evaluation service is the ability to apply models against millions of records in a few hours. This means that robust statistical models can be built and deployed in a few days to condense the traditional time taken for campaign analysis. As a result our clients can achieve true closed-loop campaign management and apply learnings from previous campaigns to future marketing activity. 
  
  
• Customer Segmentation - RFV, rule-based decision tree models or cluster analysis, using client and/or 3rd party data
  
  EuroDirect has extensive experience of building segmentation systems for a variety of clients incorporating product-holding and transactional information with customer demographic and lifestyle data. Various multi-dimensional segmentation approaches are supported, to create solutions that enable clients to implement niche marketing campaigns to a number of distinct and homogeneous customer groups. A segmentation analysis project is essential for unlocking customer insight, to reveal the natural clusters within a customer base. It can highlight crucial differences in the value, behaviour and attitude of customers, which can trigger new product development opportunities and creative communication strategies. 
  
  
• Attribution Modelling - Applying internal or research based segmentation and propensity models onto customer databases through extrapolation and data fusion
  
  Many customer insight projects begin with a piece of market research that delivers a segmentation typology. This is of limited value when it is only available on the sample of customers interviewed for the research study. EuroDirect have experience of using predictive modelling and data imputation techniques to map survey data onto a customer database. This means that the resulting propensity scores or segment codes can be used for direct marketing. Making use of external data such as CAMEO with universal coverage provides a bridge to extrapolate variables from a research sample to a full database. This is a proven methodology that has been employed in the UK to model variables from Market Research panels onto the Data Exchange database. 
  
  
• Behavioural Modelling - Predicting key customer events e.g. churn, cross-sell, segment migration
  
  Behavioural Modelling employs time series analysis to predict future events such as product purchase, account closure or migration from a low to a high value segment. It is useful to help marketers anticipate likely customer behaviour and allocate development and retention budgets accordingly.
  
  A key requirement for a behavioural modelling project is historical data that allows a sample window to be defined. This timeframe for analysis needs to identify an observation point in the past (e.g. a specific date 6 months ago) and an outcome period (e.g. the last 6 months) when the anticipated event could have happened. This outcome is then predicted using data that was known at the observation point. This ‘longitudinal analysis’ often involves appending retrospective data that applied at some time in the past. For UK financial services companies, the Marketing Analytics team makes use of a newly developed retro process to append credit data pertaining to an individual at a historic observation point (e.g. at time of mailing, enquiry or purchase date. The retro process is engineered to locate data from different time periods for individual records. This means that all customers within a file can be “retro’d” to the same point in their lifecycle irrespective of when that actual date occurred. 

Recent Projects

Using the Forensic Analytics Campaign Evaluation methodology (see above), EuroDirect helped Lands End Direct Merchants achieve a 25% increase in their response rates. Other recent case studies include: 

• Response and conversion models for numerous lenders, credit card companies and brokers
• Integrated suite of targeting models and business intelligence support for development for large scale international DM property investment company
• Customer value, share of wallet and delivery channel segmentation plus market research attribution modelling for UK Top 10 financial services company
• Development of bespoke postcode level response and acceptance targeting models for a doorstep lender
• Customer segmentation development for a leading travel company
• Behavioural modelling to predict mortgage redemption for the UK’s 2 nd largest building society
• Modelling energy consumption and attributing segmentation codes for a national utility provider 

 Working Practices

The EuroDirect Marketing Analytics team adopts a collaborative approach to project delivery. This means that tasks are shared between analysts to optimise the approach and ensure that any errors in analysis or interpretation are trapped at the earliest opportunity. All output is signed of by a senior member of the analytics team for quality assurance purposes. EuroDirect believe that working in partnership with client companies achieves the best results. To this end, we encourage regular dialogue between analysts and the client team to ensure a good understanding of the business requirements. 

Tools and Techniques

The Marketing Analytics team use the following software tools for modelling and analysis:

• SAS
• SPSS
• MICROVISION
• Angoss Knowledge Studio – specialist data mining software
• SQL and FoxPro programming languages
• Intelligence Creation

This enables the team to use a wide range of modelling techniques including :

• Decision Trees (CHAID)
• Logistic and Linear Regression
• Factor and Cluster Analysis
• Neural Networks and non-parametric methods 

Agency Support

EuroDirect will develop models on behalf of agencies and end-user clients where the associated IPR can be vested with the client company. Analysis output can delivered in raw format for agencies to present to their clients with their own branding.

All models are provided with supporting documentation – the focus is on data visualisation, backed up with a written commentary where necessary. A number of the team members are qualified geographers enabling EuroDirect to ensure that the i mpact of geography on any profile or targeting model is measured and understood. As proprietary mapping software (MICROVISION) is one of the tools used by the team this helps to ensure that any geographical bias in customer data is not overlooked.

For customer segmentation projects, the Product Management team can be involved to develop pen portraits to characterise different segments. This team’s experience in launching more than 40 UK and international geodemographic classification systems is invaluable when it comes to interpreting cluster analysis results and creating lifelike segment descriptions.

For more information on how we can help you with your data analysis and modelling requirements please contact us on +44 113 242 4747 or email us at info@eurodirect.co.uk