Data Mining Automation

DynaMine – Data Mining Automation

Competitive edge through Advanced Analytics

In today’s competitive environment, most companies with comparable products or services compete with each other over the same customers. That’s why it’s imperative to quickly recognize and understand the individual needs of each and every customer, in order to start putting this essential information to appropriate yet cost-effective use in customer interactions. DynaMine feeds you this information automatically, thus providing you a decisive competitive edge.

Tags in terms of prognosis qualities

DynaMine’s prognoses also fulfill the highest quality standards without requiring a host of analysts to run it.


DynaMine supports the user during data analysis in that as many models as desired – whether automated or interactive – can be created and administered simply and efficiently via the front end, thus greatly reducing the time usually required for creating and maintaining data mining models. Furthermore, through regular re-evaluation, the prediction capacity of the models is notably improved. Especially in very complex environments (e.g. when used in next best offer or campaign optimization systems), DynaMine can show its real muscle: the major effort normally expended on the data resources required for such systems is cut to a minimum by DynaMine’s automation.

DynaMine automation

Fields of Application by Industry

DynaMine can be applied in any kind of enterprise that uses data mining models to evaluate data and make predictions for the future.

Overview of potential application areas
Marketing & Sales Next Best Offer
Generating customer-specific offers based on interests and purchase history, for example to present a list of the ‘Top 5 products’.
Generating complementary services that correspond to an already-purchased product.
Offering higher-valued alternative services or products for an already-purchased service or product.
Product affinities
Based on customer characteristics (e.g. gender, age), offering products that were deemed relevant by other customers with similar characteristics.
Optimization of potential through prognoses of propensity to churn
Recognizing early on which customers are at risk of churning, so you can take appropriate preventive action.
Forecasting (demand forecasts for call centers or sales forecasts for products)
The Advanced Forecasting module generates inbound call forecasts that can be used to manage the expected load for call centers.
Industry Predictive Maintenance
Automated evaluation of the output from a variety of sensors for early detection of a potential machine breakdown.
Accounting Fraud Prevention and Detection
Identification of fraudulent activities in accounting.


Selected application examples

Online/catalog retail – next best offer

Let’s say an online/catalog retailer wants to present the top three customer-specific products in the next edition of the newsletter. Such product rankings require a lot of analysis effort because they have to be individually calculated for each customer: for each product, a corresponding prediction model must be created, so that in the next step, the three products with the highest potential for being purchased can be calculated. DynaMine makes it efficient to set up and train these models and ensures that the calculated data remain current for each successive edition of the newsletter.

Mail order

Telecommunications – predicting propensity to churn

Recognizing customers at risk of churning is especially interesting for industries that operate on fixed contracts with minimum terms. With churn models calculated by DynaMine, you can identify these customer groups exactly, in order to contact them specifically with campaigns designed to keep them from canceling. Furthermore, using the innovative Uplift Modeling application, you can detect so-called ‘sleepers’: these are customers who are first made aware of the end of their contract when you contact them – and therefore wouldn’t otherwise be selected for a churn prevention campaign.


Manufacturing – predictive maintenance

A multinational manufacturer running heavy machinery non-stop to produce its goods: in order for it to perform maintenance and repairs, the machinery must be shut down for several days, leading to loss of output. But machine defects cause unpredictable secondary losses: besides costly repairs, a breakdown can result in even longer equipment downtimes. For the organization, it’s much more economical to detect possible defects as early as possible through irregularities in the machine data (RPM, temperature, vibrations) and then schedule repairs that fit with the production plan.

Real-time recommendations & campaign optimization

DynaMine Modules

In the area of data mining, you’re always going to be confronted with new challenges that weren’t on yesterday’s radar screen. To stay a step ahead of your competition, you need to continually adapt your data mining system to your needs.

We rise to these challenges by offering modular extensions for DynaMine that allow you to tailor the solution exactly to your momentary requirements.

Shopping Basket Analysis This module enables analysis of shopping baskets for identifying products that are often purchased in combination. Additional relevant products can be suggested to the customer based on the current contents of his shopping basket.
Text Mining Using the following elements, text mining enables the analysis and evaluation of unstructured information from e.g. Facebook posts, Twitter, Bazaarvoice:

  • German-language, semantic text mining
  • Text filtering and standardization
  • Stop words (the, a, and)
  • Stemming
  • Extraction of keywords
  • Sentiment recognition and classification as positive, neutral or negative
Forecasting Analysis of time series for predicting future developments. For this, two different approaches are used:

  • Predictions can be made based on past data: The higher the price of copper three months ago, the higher it will be in the coming months.
  • But it’s also possible to take other influencing factors into account: The higher the number dunning notices sent in the previous week, the higher the volume of calls in the week following.


Implementation and Amortization

DynaMine can be integrated easily into an existing data mart or data warehouse. As long as the necessary data foundation is available, the installation can be completed within a few days. Automation leads to savings of time, which pays off quickly when it comes to data preparation: in classic data mining, nearly 70% of the time is spent on data preparation. DynaMine greatly increases the optimization potential by automating repetitive tasks. The time saved can then be used for sophisticated analytical problems. Thus, DynaMine very quickly provides real added value.

When the training, scoring and evaluation processes are also automated, DynaMine can be amortized after just a couple of models.

Amortization of Data Mining