B2B analytics is data-based decision-making and a data science concept in the environment of corporate customers. It solves data consolidation and visualisation, income segmentation, predictive models, credit risk models, etc.
Statistical modelling and customer analytics have been traditionally associated primarily with the B2C world where there is sufficient data and the law of large numbers may work well. However, it no longer holds true that there are only few customers and data in the B2B world. Data-managed decision-making and the data science concept are penetrating B2B ever deeper, with concepts tried and tested in retail also working well in respect of companies. Deloitte has lately recorded an increase in the demand for B2B analytical solutions and has brought one of its own.
Our B2B analytical projects usually start with a discussion of the different analytical tasks on offer, the objective being to select a few with a sound business case. Data consolidation and visualisation, basic revenue portfolio segmentation, non-structured text mining from contractual documentation, propensity to buy (PtB) models, customer wallet models, credit limit optimisation (CLO) and credit risk models (PD models) are but some of the most frequent tasks. When tasks have been selected, the process enters the next stage, in which scoping takes place. Scoping entails the estimating of labour-intensiveness and the planning of resources, inputs and outputs. After this comes regular project management. A lengthy yet highly important phase of preparing, consolidating and cleansing data takes place, with algorithms subsequently applied to the cleansed data and predictive models created. The models are assessed in terms of performance and business logic. Based on the performance achieved, the business case is updated. After the resulting models have been accepted by the business, we implement them in production and integrate them in the process, and we develop reporting. Throughout the entire duration of the project, we hand over our knowledge to the client’s internal team – for more information refer to Building Data Science Capability. In the production phase, we perform what is referred as “baby sitting”, where we monitor the process as it unfolds, remove any imperfections and specify the business case based on actual results. When everything is running smoothly and no improvements can be made, we hand over the whole solution to the client.
A six times greater sales success rate and fifty times lower customer churn in the SUPERACTIVE segment as opposed to the NON-ACTIVE segment. Behavioural segmentation for the SME portfolio of a major Central European bank.