NEWS FROM MXBEN - WORK IN PROGRESS…
2022/04 – Our 'swarm' takes off - from technical dream to customer benefit
- Isn't it possible to orchestrate hundreds of computational steps, which are in a complex conditional relationship, by machine and to complete them by parallel processing within a fraction of time?
- Can a 'swarm' of similar machines be interrelated in a way that the computers self-organize and use collective power to accomplish tasks that would take each member days to solve alone?
- Thinking ahead, can the power of the collective be scaled at will - to the performance limits of the hardware?
In early April, our 'statistics fabric' went live with a federation of eight machines. It enables an enormous leap in processing speed - and the hardware offers plenty of room for improvement. We opted for a robust in-house development: This gives us the security of being able to control the processes and react quickly to new requirements.
We thus offer our customers significantly shorter delivery times and the security of being able to meet increasing requirements with the appropriate resources.
2022/02/26 – Support for peace in Europe
The war of aggression of the Russian Federation under the leadership of its president on the territory of Ukraine, which is contrary to international law and obviously prepared for a long time, saddens us deeply and marks a caesura in the development of peace policy in Europe as well as in neighboring states. Our behavior as only apparent outsiders to the conflict region is always a statement, even if we remain passive. We have decided to use the possibilities immediately available to us to respond to Ukraine's international call for help.
Donations to support the urgently needed emergency humanitarian aid are accepted by the relevant well-known national and international aid organizations, e.g. UNICEF or Aktion Deutschland hilft can be found on Google under 'Emergency aid Ukraine'.
2021/08 - Interface to Wilken NTS.suite/MS Dynamics NAV
We are pleased to announce that we're going to develop a bidirectional interface between Wilken NTS.suite/Microsoft Dynamics Navision and the Customer Analytics Service in one of this year's new projects.
In addition to SAP/IS-U, kVASy (SIV) and lima (rhenag), we are thus building a bridge to the fourth ERP and billing system widely used in the industry. We are also demonstrating that the diversity of source and target systems used by our customers is not an obstacle to the further digitization of customer-related processes.
2021/05 - NEW sales consulting proposal (next best offer) is live
For "Customer Analytics - bottom up by mxben" we have developed the module for the production of individualized sales consulting proposals (next best offer) from scratch.
From the intersection of the individual interests of the supplier energy customer and the sales goals of the supplier, those offers in the portfolio are selected which are characterized by the highest possible attractiveness and purchase probability.
To support customer service, the information is put into language. Integration into automated systems (e.g. online service) is also possible. The approach is made intelligent by integrated feedback loops: changes in the customer relationship and feedback from the consulting dialog flow into the organization's experience memory and serve to optimize future proposals and offers.
An Excel form puts the management of consulting proposals in the hands of our customers. This reduces the effort for communication and administration.
We guarantee the adaptability of the module by mapping a large part of the recommendation logic in flexible data structures. Compared to the previous implementation, the amount of static program text was reduced by about 70-80%.
2021/04 - Energy in our own business
Our services address the energy industry, but without energy nothing works for us either: For example, the cost of green electricity has accounted for half of the operating costs in the data center since the beginning of the year. Our consumption is now well above the needs of a large family. Alongside good ideas, data and IT, electricity is the most important operating resource for mxben as an innovative information service provider.
2021/03 - Changing suppliers in times of Corona
The existential uncertainty caused by the surprising and unexpectedly persistent crisis also affects the willingness of energy customers to switch suppliers: Since last fall, we have been observing a divergence between predicted and actual supplier changes for the upcoming 12 months. Whether the seemingly postponed, but not necessarily cancelled, interest in switching by the end of the year will lead to an increase in the seasonal number of customers giving notice is currently an open question. This is likely to depend to a large extent not only on epidemiological developments, but above all on economic developments, which are currently influencing many people in their consumer behavior and economic planning.
Valid forecasts of switching probabilities enable energy suppliers to address business partners at risk of switching preventively and to reduce customer losses.
2021/02 - Customer analytics integrates 3rd segmentation dimension
This has nothing to do with virtual reality: Our segmentation approach offers tangible support for managing customer-related processes. In addition to the probability of switching and customer value (future contribution margin), it now also integrates relevant customer preferences. To this end, we make predictions about the needs of a business partner in relation to its energy supplier and service provider. The prediction is based on information about purchasing decisions such as tariff changes or the use of energy-related services by similar customers in the recent past. This gives energy suppliers the opportunity to address communication content and offers in a very target group-specific way and to reduce wastage.
2021/01 - Technology
The turn of the year provides an opportunity to take a look at the technical foundation of our work. Data Science does not only take place in spherical heights, but always within the limits of technical possibilities. In practice, calculations require up to several days of time and a lot of green power, depending on the objectives. The effort for continuous maintenance, which is necessary to keep the system running safely, is unfortunately often underestimated. Here you can find more information about the hardware of our data manufactory.
2020/12 - Corona sucks - strategy remains in demand
Already in spring it became clear that new planning was needed. News from politics, business and society show that the fundamental lack of empirical knowledge and the fallibility of all forecasts in combination with danger in delay represent a separate set of problems with which we are poorly coping.
At the same time, the realization is maturing in many places that this caesura could permanently change accustomed ways of thinking and that we will permanently adapt some of our target systems.
Since late summer, mxben has been taking an entrepreneurial approach to the crisis. In our 8th year of business, we are investing more than ever in the security and performance of our customer analytics platform. Shortly before Christmas, we put our newest server into operation, equipped with two specific high-performance and cutting-edge processors of the EPYC series from AMD. This will not only speed up recurring calculations, but also our development work considerably.
At this point, we would like to expressly thank all existing and new customers whose trust in our competencies has made these decisions possible.
We would also like to thank the German federal government and the state of Hesse for their offers of support in overcoming the Corona crisis and for promoting our willingness to invest.
2020/12 - Do-gooder
We support the TECH TO THE RESCUE initiative with a modest commitment. The goal of this innovative platform is to bring together aid organizations/NGOs transnational specifically in the Corona crisis with volunteer companies and know-how carriers in the digital industry to overcome emerging hurdles with technology.
Separate from this, by the end of the current fiscal year, we will have devoted just under 1.5% of our gross operating income to scientific and charitable causes in the form of recurring or cause-related donations.
2020/10 - Association vultures
mxben joins the AI-Frankfurt Rhein-Main e.V. With this we would like to support and promote the goals of the association: to bundle economic, scientific and social activities around "artificial intelligence" in a regional network, with our own experiences.
2020/09 - Customer Analytics integrated next-best-offer
On behalf of a customer, we are expanding our customer analytics approach to include a module for dynamically deriving sales proposals for individual customers (next-best-offer). The key is to select a mutually beneficial product offer from the intersection of sales interests and customer interests (win-win).
2020/09 - mxben meets OpenStreetMap
What Wikipedia is for the knowledge of the world, OpenStreetMap is for the mapping of the earth. With the support of experts from the Geofabrik in Karlsruhe, we are working on making this treasure trove of data permanently available to our customers starting next year.
2020/07 - 7 years of "maximum benefit"
Soon it will be so far - Already today many thanks to all who have contributed to it!
2020/04 - Customer Analytics integrates price acceptance analysis
Successful pricing policy is one - if not: the - supreme discipline. Therefore, we are pleased to have successfully completed a challenging customer assignment to analyze price acceptance, price sensitivity and price threshold perception when adjusting prices in nationwide sales.
The main focus was on the mapping of framework conditions of the decision situation of individual electricity and natural gas customers; furthermore, on the temporal and spatial localization of events and conditions as well as the derivation of economic consequences. All analyses could be performed uniformly on the basis of our 4D data model.
2020/03 - Corona: mxben sets up virtual classroom
We make part of our experience, time and technology available to the 7th grade of a Frankfurt high school and set up a "virtual classroom" for students, teachers and parents free of charge. In this way, we not only want to bridge the school closure, but also promote digital forms of collaboration and learning.
The positive experiences of the first week alone are encouraging and show that even in a crisis, companies don't just call for help, but can help themselves.
2019/05 - Customer Analytics integrates duplicate consolidation
Business partner duplicates are like obstacles in the way of intelligent customer relationship management - and yet they are at home in every historically grown system.
Customer Analytics à la mxben now integrates an intelligent, multi-dimensional detection and cleansing of duplicates, which we recently developed in a project for a customer. Whether the information gained in the process, which can be updated regularly, is merely used to increase the accuracy of our forecasts or is also used for enrichment in the source systems is up to our customers.
2019/05 - AI-German Federal Association does not welcome mxben - and still deserves praise
The young KI-Bundesverband launched its own seal of approval in March 2019. Companies in the AI industry that make a commitment to the association to follow the principles described when using artificial intelligence are allowed to use the seal.
The good news first: We think the document is great because it provides practical answers to important questions and creates transparency. For background information, we recommend the report "Verbrauchergerechtes Scoring" (Consumer Scoring) by the German Council of Consumer Experts at the German Federal Ministry of Justice and Consumer Protection. The issues raised there concern us all.
We would have been happy to sign the voluntary commitment of the quality seal. The only thing missing was membership of the association. Unfortunately, our application for membership was rejected. Reason: It was not possible to "recognize a concrete AI focus". :-O
2019/04 - Mass and therefore also class
Statistical forecasting is at odds with everyday experience here: after all, the more objective factors are known that can influence our behavior, the more accurately characteristic patterns and predictions can be derived from them. As a result of the latest expansion of our database, around 520 individually processed characteristics are currently incorporated into the forecasting procedures.
2019/04 - Customer Analytics WHAT IF released
What concrete influence will
- Your alternative courses of action in terms of pricing and product design (revenue side)
- General conditions on the cost side (procurement, NNE, taxes, apportionments)
- Reactions of the competitive environment (market side)
within the next 12 months on the individual
- future contribution margin (prospective customer value)
- loyalty (probability of switching suppliers) and
- the segment affiliation
of your customers? WHAT IF systematically illuminates and answers these questions. The consequences of expected developments as well as possible decisions are forecast - bundled in scenarios - in euros per year (customer value), number of existing contracts and business partners as well as resulting sales volumes.
WHAT IF builds seamlessly on Customer Analytics BASIC. We pay particular attention to the modeling of internal switching processes (tariff changes), which - in addition to the supplier change - have a considerable influence on future contribution margins.
We would like to thank our partner, Thüga SmartService GmbH (Naila) - until 2019: Conergos GmbH & Co. KG (Munich) - for their support, which helped make this development step possible.
2019/03 - Customer Analytics BASIC receives interface to lima
Previous projects focused on SAP/IS-U and kVASy (SIV) as major source systems. We will soon start processing data from lima (rhenag) for a customer. We are pleased to be able to expand the interfaces of the Customer Analytics Service with this.
2019/03 - Is switching suppliers a bad weather decision?
Clear answer: yes, too. As part of an experimental analysis, we examined the influence of distinctive weather indicators on supplier switching. Indeed, energy supply contracts are terminated more frequently during cold and rainy months. The intensity of solar radiation seems to have little influence on this. Since price adjustments or bills are often sent during these months, the effects cannot be precisely separated.
2019/02 - Support for the work of UNICEF
From now on, we support the work of the United Nations Children's Fund UNICEF not on an ad hoc basis, but continuously as a sponsor with a monthly contribution.
2019/01 - Data transfer portal online
ftp - goodbye: We exchange analysis results and data with our customers not only securely, but also conveniently via the Internet. Our transfer portal based on Nextcloud also offers up-to-date interfaces for fully automated data exchange and - like almost everything at mxben - is operated in-house on our hardware.
2018/11 - Price data interface to SAP/IS-U
The determination and simulation of past, and above all future, revenues is the linchpin of every economic success analysis. On the other hand, our behavior as consumers has an enormous influence on what remains in our wallets.
We have recently developed an automated interface for the transfer of time period-related price components down to plant operands in SAP/IS-U.
2018/10 - WHAT IF...
How do expected market price level changes interact with variants of a future price adjustment to affect economic success and other key figures? We investigated this question in a pilot project for a customer using twelve different development scenarios. The results were surprising and point the way to a valid forecast of future effects of room for maneuver in pricing.
2018/08 - Customer Analytics: BASIC Package
Together with Thüga SmartService - formerly Conergos - we put together the BASIC package "Customer Analytics - bottom up". Our offer to energy suppliers includes all necessary project steps to support target group-specific marketing and sales activities with a churn analysis and continuously updated forecasts on supplier switching probabilities, customer values and a segmentation model.
2018/07 - 5 years of "maximum benefit"
mxben has turned five years old. For children (you'll excuse the limp comparison!), this birthday indicates that they are "out of the woods". This small anniversary gives us reason to thank those who have contributed to this success. In the first place we thank our customers, who have challenged us with questions and big tasks; in the second place our partners and suppliers, without whose cooperation this work would not be possible. We look forward to continuing our cooperation!
Project experience confirms that the initial effort involved in individually preparing and linking data from different sources is rewarded with exceptionally high forecast quality. We will expand this approach and make it usable for new application areas.
2018/05 - Entry into force of the General Data Protection Regulation (GDPR)
2018/04 - Thüga SmartService and mxben agree on cooperation
Thüga SmartService GmbH (Naila) - until 2019: Conergos GmbH & Co. KG (Munich) - and mxben are deepening their cooperation. Thüga SmartService offers comprehensive solutions to support all core processes of an energy supply company and is the leading IT service provider for the Thüga Group association.
Together, the partners offer companies in the German energy industry an integrated product range that is future-oriented with regard to the digitalization of their business processes.
2018/03 - Customer Analytics: individual and multi-client capable
For the first time, we made our customer analytics solution multi-client capable for a customer. The main task of the pilot project was to combine the usual individuality of data import, preparation and forecasting with standardized but adaptable interfaces and scalability. The new development makes it possible to integrate other utilities working on a common SAP template into the solution with reduced additional effort and to make future enhancements more efficient for all participants.
Other projects will continue to benefit from an individualized approach tailored to the utility's situation and data situation, using our common data model.
2017/12 - Move to the new data center
After thorough preparation, our servers move into their new home at 23media in the Telehouse data center in Frankfurt. With Tier3+ quality level, the environment meets even higher requirements and offers additional space, which we use to make new investments in the security and reliability of our analytics platform.
What we also appreciate is that our hardware is now powered by green electricity in a CO2-neutral way and the noise level in the server room is so low that you can still talk to each other!
2017/11 - Quality leap in the accuracy of our forecasts
For the first time, we achieve a hit rate of 80% in predicting the termination date in the reference scenario, which is theoretically possible but unattainable by practical means. The result is achieved using an analysis data set with several million rows and around 400 individually calculated features. In follow-up projects, the result can still be slightly improved.
2017/11 - mxben supports the open source project IPFIRE
With great success, we have been using mature OpenSource solutions based on Linux in combination with specific commercial security solutions for several years. Because we benefit from this together with our customers, we give a part of this success back to the developer community of the OpenSource firewall distribution IPFIRE and support the project with a small donation every month.
2017/07 - mxben concentrates human resources
Since the beginning of the second half of the year, Andreas Hahne has been leading the company alone as managing director and shareholder. Exciting projects with new customers and data from different source systems (SAP, SIV, Microsoft and others) allow us to redesign and formalize the entire process chain of data preparation, integration of third-party data and final evaluation and forecasting.
2016/10 - GET and mxben agree on cooperation
GET AG (Leipzig) and mxben will cooperate more closely in the future. Mutual experience on the working level goes back to the year 2002 (not a typo). Therefore we know and appreciate each other. GET is our preferred partner when it comes to the inclusion of market information in forecast models or the determination of contribution margins.
2015/01 - Concept "Prospective customer value"
How can the likely achievable value contribution of a business partner in the time horizon of the next twelve months be described with a single, economically relevant number? And predict it? How can expected values for the continuation of current contracts, the realization of cross-selling and up-selling opportunities, and down-selling risks be brought down to a common denominator?
Inspired by approaches from controlling, statistical forecasting and sales management, mxben develops the concept of prospective (forward-looking) customer value in euros per year.
2014/09 - Development of our "4D" data model
Accurate forecasts of the termination date require the consideration of temporal relationships. mxben develops a new data model and a multi-level data preparation. All four dimensions relevant for predicting customer behavior:
- Sequence and time interval between events, e.g. tariff changes
- Start and duration of circumstances, e.g., contract durations
- Person-related characteristics, e.g., age
- Location-related characteristics, e.g. number of business partners in the connection object
are mapped uniformly. This results in new possibilities for forecasting and a considerable increase in quality.
2014/04 - Commissioning of our IT platform
The first server from mxben moves into the data center. The machine is small, strong and black - the last type of a series still built by IBM. Upgraded several times in the following years, the rock-solid machine works reliably until today.
2013/07 - Founding of the company
mxben is registered on 15/07/2013 as a limited liability company under HRB 97014 in the Commercial Register of the District Court of Frankfurt am Main. In order to provide services for the company's marketing, sales and IT divisions, the founders contribute their expertise, which spans more than two decades and was most recently practiced at a major German energy supplier.
The principles of our business activity have proven themselves and are still valid today:
- Independence in cooperation with partners and suppliers: mxben values the quality and usefulness of its partners' data and services. If they are used in customer projects, we do not receive any commission.
- Fairness towards our customers: Customer loyalty should be the result of outstanding performance and not based on dependence on a "black box". Customers of mxben can therefore continue to use the programs developed within the scope of a project even after the end of the cooperation.
- Transparency of our work: Statistics is not your hobby? You are not alone. If you wish, we can still give you a detailed insight into what we do for your company.
- Data protection and responsibility: We have a clear understanding of the responsibility we assume when processing data on behalf of our customers. That's why we forgo the cost advantages of the cloud and only use our own dedicated and self-managed IT systems in Germany. Our high technical standards stand up to any scrutiny to this day.