In practice, data mining can be successfully applied in almost all economic and scientific fields.<\/p>\n
Example: A-B-C Customer Segmentation<\/p>\n
The probably most generic of all questions for companies is: how to best and most profitably classify customers<\/strong> into category A<\/strong> (most valuable), B and C<\/strong>?<\/p>\n In this context many companies often just consider the size of their customers (by their respective turnover or number of employees) or at least by its own turnover generated with this customer. Both is easy to determine \u2013 but may be quite inappropriate. If the company rather considers the margin \u2013 potentially even including the expected future margin \u2013 whilst considering many factors, this is less trivial, however, it is also much more suitable.<\/p>\n [Data mining]<\/a> can help a lot in this challenge. On the one hand a company may use such a data mining software to determine \u2013 on the basis of existing data \u2013 different customer patterns.<\/strong> These, in turn, may each have distinct driver attributes<\/strong>. This is much more revealing than the usual averaging analysis with conventional statistics programs. Companies are not really interested in knowing that the average age (B2C) or year of existence (B2B) of its customers is e.g. 30 years, but rather what exactly characterises the younger customers in cluster 1, in contrast to those in cluster 2; both may have entirely different key driver attributes \u2013 starting from the mostly bought products, via the sales channel (point of sale), to the most successful ad campaign. On the basis of this knowledge one can already trigger a lot of optimisations.<\/p>\n The calculated forecast quality<\/strong> may be even more important for the resource allocation within your sales department. Data mining can forecast on the basis of the analysis of past data whether a potential customer<\/strong> will become with great probability an A, B or C customer<\/strong>. This, in turn, is crucial if a company wants to make focused and efficient use of its sales force.<\/p>\n This way Data.Mining.Fox\u00ae<\/strong> can support all companies<\/strong> in optimising their sales activities and in forecasting on which lead customers the sales employees should concentrate on best.<\/p>\n Data Mining Example: Banking<\/p>\n Banks<\/strong> have the problem of predicting the credit-worthiness of new clients on the basis of historic data of past clients.<\/p>\n The creditworthiness<\/strong> also influences the interest rate of a credit. The sequel describes how [data mining]<\/a> can be applied to this problem:<\/p>\n A bank has data about clients to whom it gave credits in the past. The client data contain personal data, data describing the financial status and the financial behaviour before and at the time the client was given the credit. The clients are divided into four classes. The first class contains all those clients who payed back the credit without any problems; the second class those who payed back with little problems here and there; the thrid contains those who should only get a credit after detailed checks because substantial problems of payback occurred in the past; and the forth class consists of all those who did not pay back at all. Using this data table a prediction model is created in order to predict the probability for each class for new clients. By the way: the combinations of attributes which are responsible for clients to have a high probability of not paying back will be identified by the prediction model, too.<\/p>\n The example shows how Data.Mining.Fox\u00ae<\/strong> can help banks<\/strong> to better predict the creditworthyness of customers. This helps to reduce the number of loan defaults on the one hand. On the other it also allows to offer better conditions to other customers with lower risk. Both, the bank and its customers, are therefore benefiting from this advantage.<\/p>\n Download test data<\/strong> by Easy.Data.Mining™ for the financial sector: The above is just one of many possible examples how Easy.Data.Mining™<\/strong> can add value by being profitably applied to challenges in the banking<\/strong> sector. By the way, for credit institutions<\/strong> and professional investment advisors<\/strong> we have developed a special product variation. Upon request you will receive more information.<\/p>\n Data Mining Example: Beer & Nappies<\/p>\n The biggest value added can in most cases be derived from new surprising correlations. Would you know what the connection is between beer and nappies<\/strong>?<\/p>\n Here is the solution: A supermarket<\/strong> consequently put aside all its supposed assumptions, and it reassessed its sales strategy with respect to the positioning of goods in the market. Even the usual categories of products of the trade chain were ignored, i.e. foods were not just compared to foods, but also to everything else \u2013 a task which would simply be impossible for thousands of products if one tried with manual means and without the use of [data mining]<\/a>. Even so, the department store<\/strong> added intelligently other data for the analysis \u2013 e.g. the gender of the buyers, weekdays, and many more.<\/p>\n The interesting result: Men who have children and who (have to) do the shopping on Saturdays often tend to buy nappies<\/strong> for their little ones besides the beer<\/strong> for the weekend evenings in front of the television. Subsequently, the superstore decided to position the palettes of beer besides those of nappies on Saturdays \u2013 with the success of strongly risen sales figures.<\/p>\n Hence Data.Mining.Fox\u00ae<\/strong> may help a supermarket to optimise the positioning of<\/strong> its goods<\/strong> in the shop and raise the revenues.<\/p>\n Download profit matrix example<\/strong> by Easy.Data.Mining™: The above is of course just one our of many examples how Easy.Data.Mining™<\/strong> can be applied successfully in the area of\u00a0 supermarkt chains<\/strong>. Be creative \u2013 and successful!<\/p>\n Data Mining Example: CRM<\/p>\n In customer relationship management<\/strong> (CRM<\/strong>) it is well known that it is much more expensive (about seven times) to acquire a new customer than to take measures to keep an existing customer (churn prevention<\/strong>).<\/p>\n If an existing customer is lost to a competitor then the question arises whether in the past there were signs indicating the customer\u2019s dissatisfaction and whether it could have been predicted. The following example describes how [data mining]<\/a> can profitably be applied to such problems:<\/p>\n A company has a CRM-system for the administration of its customer data. It has plenty of data describing the customer behaviour and personal data about each customer. All the customers of the company of two years ago are considered. These are divided into two classes: The first class contains all those customers who are still customers today, the second containing those who are no longer customers. With this data a prediction model is created to predict the probability that a customer will be lost within the next two years.<\/p>\n With this prediction model the company predicts the probability of leaving within the next two years for each of its present customers. A pattern recognition on the basis of this prediction model will reveal the combinations of attributes describing customer segments with a high probability of leaving the company. The customers with the highest probability can be analysed in more detail, and special offers can help to keep the customer satisfied.<\/p>\n The example shows how Data.Mining.Fox\u00ae<\/strong> can help in CRM <\/strong>to reduce the churn rate. Thus the company can benefit from various advantages: a higher number of customers (reduced churn), less cost (lower acquisition cost per customer), and higher turnover per client (higher lifetime vlaue). In turn, the customers benefit from more customised offers.<\/p>\n The above is just one of many possible examples how Easy.Data.Mining™<\/strong> can add value by being profitably applied to CRM <\/strong>challenges.<\/p>\n Data Mining Example: Insurances<\/p>\n Insurance <\/strong>companies have to estimate the probability of a claim<\/strong> using historic data.<\/p>\n The potential of [data mining]<\/a> can be of immense importance to insurance companies. The following describes an example:<\/p>\n A car insurance company wants to create a prediction model to predict the probability of a car accident happening within a certain period of time on the basis of customer data which is available at the time of signing the insurance policy (e.g. personal data, attributes of the car to be insured, history of accidents.). Looking at the past it is known whether past customers had an accident within a certain period of time or not. Past customers are split into different classes with respect to the costs of their claims. Therefore, a data table is available with each data record representing the data of a past customer at the beginning of a year and the customer\u2019s claim class in that year. The prediction model is created using this data table. The prediction model also reveals interesting customer segments with a high risk of belonging to a bad claim class.<\/p>\n The example shows how Data.Mining.Fox\u00ae<\/strong> can help insurances <\/strong>to better predict claims of customers. Thus the insurance premiums can be attributed more accurately to clients. The insurance as well as most of its customers are equally benefiting from this.<\/p>\n The above is just one of many possible examples how Easy.Data.Mining™<\/strong> can add value by being profitably applied to challenges in the insurance <\/strong>sector (another good example is the detection of insurance fraud).<\/p>\n Example: IT & Computer Technology<\/p>\n Data minnig can generate substantial value added not only on the commercial side. Hereafter you can find an example from the technical area of information technology<\/strong>:<\/p>\n A company offers an online service for several countries from a central data centre. Based on customer feedback as well as so called last mile measurements (which can assess the service experience regarding website response time<\/strong> from a customer perspective) the company knows that its service suffers from an insufficient loading time of the homepage<\/strong> for some users. The employees of the IT department have already dived into the log files, but have not disclosed any significant driver for the problem.<\/p>\n The management knows that the response time of the site is correlated with the online sales. Therefore it has a high interest in solving the problem quickly. Subsequently, the log files are then analysed with a [data mining]<\/a> tool. Interestingly, it can be discovered this way that there is no such single type of typical problem customer (therefore the log files did in average not show any significant anomalies). In contrast, the data mining program discovers that there are a number of different problem clusters<\/strong>. One of these patterns could mainly be determined by a combination of country code (TLD), operating system, and the browser version of the customer; another\u00a0 different pattern may show the same bad performance in loading times of the website, yet, it may rather be determined by other factors: the speed of the Internet access (dial-up, DSL, etc.) in combination with the cookie settings of the customers\u2019 browser, as well as the number of parallel sessions on the central master database.<\/p>\n Download test data<\/strong> by Easy.Data.Mining™ for the field of information technology: Data.Mining.Fox\u00ae<\/strong> can help information technology<\/strong> companies or the IT employees of a company in general by searching for errors and problems. Thus the company frees up valuable resources for the actual IT development.<\/p>\n Data Mining Example: Marketing<\/p>\n In marketing<\/strong> in the area of advertising campaigns data mining can often increase the response and purchase rate<\/strong> by a factor of two to three.<\/p>\n The following describes a typical [data mining]<\/a> example:<\/p>\n A company wants to launch an advertising campaign for a product. Among its present customers the company wants to post product information to those with a high probability of purchasing the product. The company has data describing the past customer behaviour and personal data about each of its customers. There are also customers who have already bought the product, e.g. in a trial period. The customers of the trial period are divided into two classes: those who have bought the product and those who have not. With this data a prediction model is created to predict the probability of purchasing the product. After that the probability of purchasing the product is predicted for all other customers. Only those with a higher probability are addressed. As a side effect the company learns with this data mining analysis which are the relevant driver attributes of its customers buying a specific product (or at least being very interested in it).<\/p>\n The example shows how Data.Mining.Fox\u00ae<\/strong> can help in marketing<\/strong> to predict the purchase probability of customers for a specific product. This reduces cost, because sales activity can be focused much better (lower cost for mailings and flyers or for cost intensive sales agents\u2019 visits on the spot). The customers benefit at the same time because the average relevance of the company\u2019s offers increases (or the other way round: the \u201cspam\u201d quota of non-relevant offers is reduced).<\/p>\n The above is just one of many possible examples how Easy.Data.Mining™<\/strong> can add value by being profitably applied to marketing<\/strong> challenges.<\/p>\n Data Mining Example: Media Logistics<\/p>\n In many areas of logistics<\/strong> it is necessary to predict the number of goods consumed in different places.<\/p>\n At the same time a lot of data is accumulated in logistics since many deliveries take place regularly. The prerequisites for doing [data mining]<\/a> are most of the time fulfilled. An example in the area of media<\/strong> is the following:<\/p>\n A distributor delivers newspapers and magazines to many newsagents in a certain region. The actual number of sold newspapers and magazines at each newsagent varies a lot depending on various parameters like weekday, month, school holidays, sports events on the previous day, national holiday, weather\u2026 Using the number of sold items on days in the past and the corresponding parameters on those days one can create a prediction model for the number of sold newspapers \/ magazines in dependence of the parameters. Then for days in the future one can predict the number of sold items using this prediction model and the parameter values (note: the information which combinations of parameters drive high or low numbers of sold items is of course also highly interesting for publishers).<\/p>\n The example shows how Data.Mining.Fox\u00ae<\/strong> can help media<\/strong> logistic<\/strong> companies to better predict the necessary print run for various points of sale. Thus it can be avoided that on some days the newspapers and magazines are sold out whereas on others there are far too many. Subsequently a lot of money and natural resources can be saved \u2013 and the customers do get the favorite read even more regularly.<\/p>\n The above is just one of many possible examples how Easy.Data.Mining™<\/strong> can add value by being profitably applied to challenges in the area of logistics<\/strong>.<\/p>\n Example: Medicine & Pharmaceutics<\/p>\n In the medical and pharmaceutical<\/strong> area it is important to know whether a patient will react positively or negatively to a treatment or a drug and which are the determining factors for the reaction.<\/p>\n Also, questions in medicine and pharmacy<\/strong> regarding side effects of the treatment or drug can be solved with [data mining]<\/a>. The following describes an example:<\/p>\n In a medical test phase a new treatment is performed on test patients. A lot of personal medical attributes (e.g. weight, gender, medical history) is obtained and stored for each test patient. At the end of the test phase the test patients are split into different classes. The first class contains all test patients who reacted positively, the second class those who reacted neutrally, the third class those who reacted negatively to the treatment. Using these data a prediction model is created to predict the probability of each class for new patients. Moreover, the pattern recognition revealing the combinations of attributes responsible for a patient to react positively or negatively to the treatment is of great interest, too.<\/p>\n Until now, rather simple statistics have mostly been used in medicine\/ pharmacy<\/strong> for such problems. Data.Mining.Fox\u00ae offers the potential for much deeper analyses and predictions in this field. Even so, any medical attributes are non-numeric which further makes Data.Mining.Fox\u00ae<\/strong> a better choice in comparison to traditional statistics tools.<\/p>\n The example shows how Data.Mining.Fox\u00ae<\/strong> can help in the area of medicine and pharmaceuticals<\/strong> to better determine which patients benefit from a given treatment or drug and which do not. This helps the medical and pharmaceutical industry, but first and foremost of course the patients.<\/p>\n The above is just displays just one of many possible examples how Easy.Data.Mining™<\/strong> can add value by being profitably applied to challenges in the medicine and pharmacy<\/strong> sector.<\/p>\n Data Mining Example: Online Services<\/p>\n In the online business<\/strong> data mining can be applied in many ways since lots of data is usually collected in an automatized way.<\/p>\n The following describes a representative [data mining]<\/a> example of an online service<\/strong>:<\/p>\n An online provider offers both a free service and a chargeable premium service on the Internet. The users of the website leave quite a number of data points \u2013 e.g. the click behaviour, the number of visits on the site, or the time spent on the site. For the company it is important to forecast on the basis of this data whether the visitor is likely to buy the premium service or not, and which characteristics are primarily driving the buy decision. Knowing this, further measures can be taken more efficiently to satisfy the customer \u2013 and hence the online provider.<\/p>\n What the online service<\/strong> company actually does is to use available data from the past with the help of the data mining tool in order to get to a forecast model. From the past visitors one does not only know their data traces, but also whether in the end of the day they bought the premium service or not. With such a model built on the past data, the company is able to forecast the probability of whether or not a new user will buy the premium service.<\/p>\n The pattern recognition allows the identification of characteristic, non-trivial sub groups of website users. For each of these clusters Data.Mining.Fox\u00ae calculates on the one hand the average probability for buying the premium service, but on the other hand also the relevance of each data attribute within each of the (different) user clusters. Subsequently, the online portal can offer each user an individually customised product or service. Thus a company can offer its customers always the next best offer (NBO) at any given point of time during the customer lifetime.<\/p>\n In the online industry<\/strong> it is probably more important than in any other sector to re-evaluate the data mining models based on historic data very regularly. The context and factors in the online world<\/strong>change so rapidly that redetermining the model is the only way to maintain a high validity of the forecasts.<\/p>\n The example shows how Data.Mining.Fox<\/strong>\u00ae can help online business<\/strong> to classify its online users on the fly into clusters, which in turn allows a more customised approach. Thus the customer benefits from a more individual online offer, whereas the online portal<\/strong> can raise the conversion rate (from prospects to paying premium members).<\/p>\n Yet, the above is just one of many possible examples how Easy.Data.Mining<\/strong>™ can add value by being profitably applied in the online world<\/strong>.<\/p>\n Example: Quality Management<\/p>\n The potential for profit levers in the area of quality assurance\/ qa management<\/strong> is often still underestimated.<\/p>\n In industrial production processes a lot of data is accumulated today. It is possible to extract much more useful knowledge from these data with [data mining]<\/a> than with simple SQL-queries or reporting. The following describes a typical example:<\/p>\n A company produces a certain product in a fully automatized production process. The production process consists of several production steps and several prefabricated components are incorporated in the final product. For each production step and each component there are several describing parameters. For instance, the age of the component, the origin of the component, the temperature of a production step\u2026. The final products are classified into several quality classes after inspection. It is a crucial advantage to be able to find out which combinations of production parameters have an important influence on the final product belonging to a certain quality class. Subsequently, the application of data minnig software in the field of quality management is obviously particularly suitable for achievement regarding six sigma objectives.<\/p>\n The example shows how Data.Mining.Fox<\/strong>\u00ae can help companies in the area of quality assurance\/ qa management<\/strong> to reduce defective goods and increase overall quality, to avoid callbacks, or to pro-actively trigger maintenance just before a damage even occurs. The result is not only a higher profit, but often a significant increase in customer satisfaction (NPS\/ net promoter score) \u2013 simply because nothing is more annoying from a customer perspective than being faced with a damage or complaint.<\/p>\n The above is just one of many possible examples how Easy.Data.Mining<\/strong>™ can add value by being profitably applied to challenges in the area of quality assurance\/ qa management.<\/strong><\/p>\n Example: Telecommunication & Mobile<\/p>\n In hardly any other sector so many data points are generated in such a short period of time like in the mobile<\/strong> and telecommunication<\/strong> industry.<\/p>\n And the development and change of supply and demand is probably one of the highest, too. This is a big challenge for the providers. Here is an example how [data mining] <\/a>software can help to tackle these defiances successfully:<\/p>\n It is a wellknown secret that the competition in the telecommunication industry is fierce. The acquisition of new customers is difficult and often very expensive. Subsequently customer retention has become more and more important. Data.Mining.Fox\u00ae can determine characteristic customer clusters on the basis of collected historic data points from customers \u2013 such as for instance the frequency and timely distribution of customers\u2019 usage of services (calls, text messages, MMS, navigation, mail exchange,\u2026). For each of these customer patterns the company can then offer tailored customer-life-cycle messages and offers.<\/p>\n The example shows how Data.Mining.Fox\u00ae<\/strong> can help a telecommunication service<\/strong> provider<\/strong> to customise their offers. This leads to higher customer satisfaction as well as to an increase in turnover and profit by risen sales over the whole customer life cycle (lifetime value).<\/p>\n The above is just one of many possible examples how Easy.Data.Mining™<\/strong> can add value by being profitably applied to challenges in the telecommunication<\/strong> sector.<\/p>\n Data Mining Example: Travel & Tourism<\/p>\n In the tourism<\/strong> sector it is vital to understand customers\u2019 needs very quickly and respond to them with adequate offers \u2013 whether in the online or the offline business.<\/p>\n At the same time companies in the tourism sector dispose of large amounts of historic data collected from their clients \u2013 a perfect field of application for [data mining]<\/a>. Please find an example hereafter:<\/p>\n An online tourism portal offers a number of worldwide holidays. With the help of the data mining software the company finds two major groups of customers: on the one hand the primarily price driven customers who are relatively flexible regarding their travel destination; and on the other those customers who know where they want to go and who simply look for the cheapest offer.<\/p>\n In addition, the data mining software can analyse the historic web analytics data, allowing to differentiate the customer clusters by their actions and click behaviour. This in turn enables the tourism portal provider to offer a customised design and content of the homepage already after the first moves of a customer. In this way the company can provide the users of the portal with exactly those services they are looking for.<\/p>\n The example shows how Data.Mining.Fox\u00ae<\/strong> can help companies in the area of tourism<\/strong> by online customisation to raise customer satisfaction as well as profits.<\/p>\n The above is just one of many possible examples how Easy.Data.Mining™<\/strong> can add value by being profitably applied to challenges in the tourism & travel<\/strong> industry.<\/p>\n Data Mining Example: University & Research<\/p>\n In the field of education and research<\/strong> \u2013 whether at a university<\/strong> or another institution \u2013 data mining software can be applied very suitably.<\/p>\n Below a (non comprehensive) list of various examples for applied [data mining]<\/a>:<\/p>\n Forecast model geneation; psychological analysis; profit maximisation models; complex statistics; questionnaire evaluation; efficient reduction of combination choices when investigating cause and effect chains; mathematical multi-variate challenges; etc.<\/p>\n The examples above show how Data.Mining.Fox<\/strong>\u00ae can help a university<\/strong> or research institution<\/strong> to enrich lectures with a simply and well comprehensible tool, to support research, or to provide students with a powerful and free tool.<\/p>\n These are for sure just a few of many possible examples how Easy.Data.Mining<\/strong>™ can add value by being applied at universities and research institutions<\/strong>.<\/p>\n Data Mining: Wine & more Examples<\/p>\n The application <\/strong>of our data mining software Data.Mining.Fox\u00ae<\/strong> is basically unlimited<\/strong>.<\/p>\n With a little creativity and entrepreneurial spirit you will for sure find many possible [data mining]<\/a> applications in your company or field of expertise. Besides our special [data mining goodies]<\/a> you can find below a couple of representative ideas (some of the examples are taken from Ian Ayres\u2019 \u201cSuper Crunchers\u201d):<\/p>\n Football Result Predictions using Data MiningIn practice, data mining can be successfully applied in almost all economic and scientific fields. Example: A-B-C Customer Segmentation The probably most generic of all questions for companies is: how to best and most profitably classify customers into category A (most valuable), B and C? In this context many companies … Continue reading Data Mining in Practice<\/span>
\nCan you reproduce the test results with our data mining test data for banks? Get our exemplary test data via this [free Data.Mining.Fox\u00ae test data download link for banks]<\/a> and just try it! For this purpose please choose \u201ccredit rating\u201d as the attribute to be forecasted (last column in our example file).<\/p>\n
\nWhen clicking the [free Easy.Data.Mining™ profit matrix example link for beer and nappies]<\/a> you will be able to follow a step by step (visual) explanation for the DMF profit matrix, based on the beer and nappies example.<\/p>\n
\nWhich test results do you get with our exemplary IT data? Would you be able to tell the management what the problems are? Get our exemplary test data via this [free Data.Mining.Fox\u00ae test data download link for IT]<\/a> and just try it! For this purpose please choose \u201cresulting performance\u201d as the attribute to be forecasted (last column in our example file).<\/p>\n\n