What does mdm mean? What is a “master data management system” and why is it needed?

We constantly face the fact that in the field of MDM (Master Data Management) there is a catastrophic lack of understandable introductory materials that allow you to quickly understand what it is, why and why it is important. Let's try to correct this to the best of our ability and explain it in business language.

There is still a serious lack of certainty in this area, and even Wikipedia, oddly enough, is not entirely sure what is what. Therefore, we will explain in our own words and “on fingers”, with examples.

What is what

So, all the data that modern information systems operate on can be divided into regulatory and reference information, master data and transactional data.

Regulatory reference information (RNI) or reference books and classifiers allow us to structure the world around us for the purpose of analyzing it. As a rule, directories and classifiers are built in the form of lists and trees. For example, we distribute all products into product groups to make them easier to manage. In world practice, this part is classified as Reference Data, with a corresponding class of applications and processes - Reference Data Management.

Master data typically reflects real-world objects and their basic properties, for example: customers, products, people, offices, vehicles. Their key difference from master data is that master data are attributes, properties, lists, and other classification planes, but the objects described by master data usually do not exist in real life. Thus, a car model is an element of master data, but a specific car with its properties is an element of master data.

Transactional data reflects our actions with world objects (read master data objects), events and interactions. At moment T, we shipped product B to client A in quantity X and at price Y. And then at moment T1 we received a payment from client A in the amount of Z, these records will be transactional data.

If we look at these three groups, we will see some hierarchy reflected in the figure. Master data is used to structure master data and transactional data. Master data defines the structure of transaction data. And only transactional data is the final result of the chain.

Thus, reference data structures both master data and transactional data itself. Master data structures transactional data. As a result, our analytical capabilities in relation to transactional data directly depend on how suitable a “framework” of master data and master data we were able to build for our purposes.

Naturally, each type of data has its own characteristics, which we must consider in detail.

Features of reference data

The key feature of reference data is their relative immutability, i.e. Reference data are initially built on the basis of business goals or from the stable objective reality of the world (for example, a directory of organizational forms or car models). Changes in master data usually occur in accordance with a certain procedure; the maintenance of master data is assigned to a specific department. Any change must necessarily go through a certain approval process that guarantees the integrity of the master data.

The sources of changes in the master data are changes in the surrounding reality, for example, sometimes new countries or currencies appear. Or, if we are talking about directories built around business goals, changes associated with changes in business goals, accounting policies, distribution of responsibilities among product groups, etc.

In fact, reference data form the basic planes of analytics, due to which the composition and structure of reference data can be quite complex, including some reference books can be used in others as attributes, forming complex structures that implement many analytical planes. For example, for the purposes of marketing, production, inventory management and financial accounting, completely different groupings of product items and different sets of characteristics may be used.

With sufficient centralization of reference data management, the directory system is usually stable and not subject to “noise”. However, in companies with insufficiently integrated applications, we see the same problem: master data is often built separately within each application based on the “local” business goals provided by this application. As a result, “parallel” sets of reference data are generated that are not combined into a single system, which greatly complicates the formation of analytical arrays.

Ideally, master data management should be centralized. Thus, as part of the IT solution, a system appears that provides the technical and organizational basis for the management of master data, in relation to which all other applications are clients of one-way synchronization of master data - from RDM to applications.

Master Data Features

Master data is more “moving”: new customers, new products, new people - all this movement comes directly from the life of the company.

The most important feature of master data is that it is here that the company accumulates knowledge about the subject of its work. All information that appears over time about a specific client must be economically compiled, structured and updated, including the entire history of changes. The same goes for any master data. Equally important is that master data also defines the relationships between different entities. What is a branch of what, who is the supplier of which product groups, which vehicles belong to which branches, etc.

However, the degree of control over changes to master data is significantly lower, because records and their updates are generated in a variety of departments, at different times, made by different people, and often also in different systems that make up the complex IT landscape of a modern company. As a result, each system often stores only some part of the information about the same entity; each system assigns its own identifier. As a result, the formation of a comprehensive analytical picture becomes more complicated, and in particularly advanced cases it becomes completely impossible.

Because of this, master data is highly susceptible to “noise” as a result of errors and the creation of duplicate records. As a rule, this is dealt with by organizational measures, rules such as “do not create a product item if it already exists” or “before adding a client, conduct a search in the current client base.” But when the scattered nature of application systems is superimposed on the variety of writings and descriptions, and this is all within the limits of human attention and diligence, then no organizational measures help.

The result of “noise” is a loss in the quality of master data, which entails the destruction of analytics, because what kind of analytics can there be if transactions regarding one real client are tied to three different client records made at different times in different systems by different people?

Large companies with a long history and complex IT infrastructure can easily have 15..20 million client records about legal entities in their systems, while in total there are less than 5 million legal entities in Russia. And this is quite a real-life case. There is no need to talk about high-quality customer analytics here.

Another example. Product catalogs are very susceptible to noise. Due to the lack of a single unique identifier and the variety of spellings, the operator is often unable to correlate new products with those already recorded in the system. As a result, an item gathering dust in a warehouse will be purchased again only because one product card corresponds to the warehouse balances, and another in the design specifications, although physically these are the same product.

It should be especially noted that most application systems do not contain any effective mechanisms for managing the quality of this data, limiting themselves to a trivial search by substring, relying on the integrity of the operator and his knowledge of the application area. Therefore, the problem of master data quality usually arises when the introduced distortions and inadequacy of analytics reach impressive proportions and are visible, so to speak, to the naked eye.

Features of Transactional Data

Transactional data is the most dynamic of all the listed types and is essentially generated transparently for most operators in the course of understandable actions such as “ship goods”, “download bank statements”. Typically, each element of transactional data contains a number of references to master data records of various types, plus a date and a series of numbers characterizing the transaction.

For example, the minimum element of transactional data generated by a shipment would consist of a customer reference, an item reference, a quantity, and a price. In reality, transactional data has a slightly more complex structure, for example, there is an invoice and a corresponding data element indicating the counterparty and date, and there are its lines, each of which contains a product item, quantity and price. Nevertheless, for analysis purposes, the number of transactional records will be equal to the number of lines in the invoice, and the record corresponding to the invoice itself does not cease to be transactional data.

It should be noted that all kinds of accounting registers containing, for example, current balances are essentially not transactional data, but a kind of analytical generalization. It’s just that this generalization is kept up-to-date by technical means and is aimed at optimizing performance. The important thing is that it can be recalculated from transaction data at any time, and this is its fundamental difference - it is secondary.

The important thing is that transaction data usually affects the interests of external counterparties, and because of this there are no special problems with controlling their quality: for example, if the operator doubles the shipment, this will affect the balance with the counterparty, lead to discrepancies, reconciliation, and the error will be corrected. Thus, mutual control between interacting counterparties, coupled with analytical tools and established business practices, ensure relatively high quality of transaction data

Why does master data require a special approach?

Having defined the concepts and considered the features of data types, we have come close to the question: why is master data management singled out as a separate function in modern IT architectures? There are several reasons for this:

1. The narrow functional focus of applications often leads to the fact that each application operates with its own set of attributes. The call center does not call the office address, delivery does not deliver goods to the email address. In other words, a complete set of information, the so-called. The "master record" for each application is redundant. In addition, in large organizations, different application systems are built for different business goals and often even operate with different master data. Developers and operators of functional applications simply do not feel the need for “foreign” master data and reference data within the framework of their goals, and all these problems ultimately fall on high-level business analytics systems, in which master data and reference data are combined with varying degrees of success . And citizens who are forced to solve them can talk a lot about the problems that arise at this point, colorfully and pretentiously, but as a rule not in a very literary way.
2. Specifics of algorithms for working with master data. The paradox is that effective master data management requires a whole range of approaches that, within application systems, are usually not spoiled by the attention of developers. So, in order to avoid duplication of records, very flexible and efficient search and comparison mechanisms and complex normalization-validation-comparison algorithms are needed. Moreover, from the point of view of these algorithms, the management of different types of master data is almost the same. In general, it doesn’t matter whether it’s a commercial tree, a list of vehicles, customer profiles or personal employee cards. The approach does not change, it is simple: clean up errors and typos, catch and stop duplicates, supplement and update as much as possible, save the entire history of changes, highlight “doubtful” cases for processing by analysts (data stewards).

Thus, a separate element appears in the IT landscape - as a separate function, an IT service, which “conducts” the entire set of master data of the company, providing all functional applications with a single, holistic, up-to-date and complete master data. In most cases, communication between application systems and MDM occurs through a data bus, messages, or SOAP/REST API.

The paradox is that any company trying to build high-level analytics automatically comes to the implicit formation of master data, even if it uses ETL procedures to “overload” data from several accounting systems into one analytical one, because high-quality analytics is fundamentally impossible without quality master data. The only question is how advanced the matching algorithms are using the ETL process, and what quality of master data is obtained in the target analytical system.

Thus, due to the specifics of master data, the inability to really qualitatively solve the problem of managing them within a separate application system, multiplied by a large number of application systems, gives rise to a separate IT domain - master data management.

What is MDM and what are they?

So, master data management or MDM is crystallizing as a separate area of ​​expertise and class of products. However, in the current English-language terminology, this is a general class of products and solutions, which includes, in particular, a number of application solutions and related business processes:

• , in general, this is the management of reference data. It would seem that this is not master data, this is a different domain. But the difference between master data and master data is only in centralization, and the same tools that we use to manage master data we can use for master data with the only difference: all systems in relation to RDM work in read-only mode, and all changes necessarily go through the appropriate business processes.
• , integration of customer data. This term covers systems that are narrowly tailored to work with the client’s master profile. It is understood that these systems aggregate the maximum amount of information about the client and provide all other systems with services for searching, creating and updating client records, as well as bi-directional synchronization.
• PIM, Product Information Management, product information management system. A specialized type of MDM, focused on information about products, their properties and attributes. For example, such a system could be the “authority source” of product information for an online storefront, a printed catalog, an ERP system, and a call center all at the same time, ensuring consistency of presentation and consistency of information.

There is no doubt that with the well-known love of the English-speaking world for the structuring of reality, specialization and acronyms, it is possible to unearth a good dozen more narrow application systems that essentially implement the functionality of master data management in one specific narrow area.

The emerging market for MDM solutions today is represented by a large number of companies offering solutions of very different quality. All “first-tier” vendors indicated the presence of MDM solutions in their portfolio, either separately or as part of their systems. However, as was said, most of them are focused on other functionality, so the quality of implementation of specific MDM functionality leaves much to be desired. At the same time, many smaller but more focused MDM companies offer solutions that are significantly more functional.

Do I need MDM?

So, when should you definitely consider implementing MDM into your IT architecture? To a first approximation, the criteria are as follows:

1. You have a lot of master data. Tens and hundreds of thousands, or even millions of records in the area that we have designated as master data. Customers and products come first. No matter how good your systems are, such volumes of data inevitably become cluttered. And it will most likely not be possible to solve this problem using the standard means of your systems. Our experience shows that even with the stated “developed mechanisms for de-duplication and data quality assurance,” the number of duplicates is calculated in tens of percent of the number of records.
2. You have several functional systems with an overlapping set of master data. Yes, options are possible when the architecture is initially properly integrated, master data is maintained uniformly and data from different systems is correctly synchronized without any intermediary, and for each type of master data a primary system is defined with which all others are synchronized. However, this is the exception rather than the rule. More often it’s the other way around. And bringing this zoo to uniformity using an MDM system will be much easier than through integration without a dedicated MDM component.
3. Synchronizing data and bringing it together into a single picture requires a lot of effort, and the resulting analytical array is questionable. Many companies use a strategy that can be described in a nutshell as follows: “let it live in each system as it lives, and at the BI level we will reduce it no matter how” and use complex, cumbersome and not very effective ETL procedures in order to create BI storage. Unfortunately, this is a palliative solution, clumsy and very resource-intensive, forcing you to “tweak” the system every day, store huge amounts of “normalized” data, essentially doing a huge amount of work to compensate for the lack of MDM.

The last thing I would like to note in this section is that there is no better time to implement MDM than the moment of introducing new functional IT systems or migrating from one system to another. After all, the implementation process involves “cleaning” the accumulated data and linking the systems with each other. By combining these things, we kill two birds with one stone: we improve the quality of data in the new system and at the same time maintain continuity with the old one, because the MDM system stitches them together.

Where is MDM going?

What should you expect from this direction? What is the vector of development? We believe that the following trends will be relevant:

1. Development of “highly specialized” MDM products. The classes RDM, CDI and PIM are highlighted for a reason; these are the most relevant and understandable MDM tasks, the solution of which has the maximum effect for business. We recommend that you be careful, because in the current state of the market, “specialized” rather means marketing positioning than additional functionality. It is not a fact that “special-purpose MDM” from one vendor will solve the problem better than general-purpose MDM from another.
2. Improving MDM functionality in application systems. This happens at least through the purchase by “first echelon” vendors of smaller, but more advanced companies in this area. This will primarily affect companies that are trying to create a completely single-vendor IT landscape.
3. Mini-solutions and integration via standard APIs. Nowadays, most MDM solutions are heavy and expensive enterprise systems that are effective only at scale. They require complex configuration, trained specialists, and thoughtful integration. And although the effect of their implementation is enormous, the price of admission is also very high. However, like many technologies in IT, they will “go down”, over time they will become more and more available to smaller and smaller companies.
4. MDMasService. Clearly, there is no barrier to providing MDM as a service on par with other applications. The open interfaces of SaaS applications facilitate their integration, so it is quite expected to use many SaaS applications integrated with each other, among which will be SaaS MDM.

Conclusion

We hope that this material has given an idea of ​​what master data is, what data management systems are and why they are needed.

Definitely, for fairly large organizations with a complex IT landscape, MDM is an absolutely necessary element that allows you to “glue” together and make huge but highly fragmented data as useful as possible, simplify the architecture, radically reduce the labor intensity of creating business analytics and at the same time significantly improve its quality.

This is a proven and effective method of correcting the protective (adaptation) system at the level of the control centers of the brain. Normalization of the state of the adaptation system is accompanied by a restructuring of the activity of the whole organism, in particular, the quality of work of organs and tissues that have deviations from the norm or are affected by a pathological process improves. As a result of exposure to a weak electrical signal with certain parameters on brain structures, various parts of the neuroendocrine system are selectively activated, which, as it were, forces the body to use internal reserves and mobilizes the immune system.

The method was discovered by a group of Russian scientists in the 80s of the 20th century, but refined, tested and introduced into clinical practice in Europe. The technique has received positive reviews from leading medical centers both in Russia and abroad. The method was used to treat more than 200 thousand patients, including in pediatric practice. When monitoring patients for more than 20 years, no negative consequences were found. Today, MDM therapy rooms using the method of Professor V. Pavlov are successfully operating in Austria, Germany, Greece, the Czech Republic and Switzerland. In Moscow, the first MDM therapy room opened at the Semeynaya clinic (Serpukhovskaya medical center). Now patients who want to use this highly effective non-drug method have the opportunity to undergo treatment without traveling abroad.

How the procedure goes:

Mesodiencephalic modulation sessions are carried out using a computer complex MDM-2000/1, manufactured by ZAT ad., Czech Republic. Apparatus MDM-2000/1 registered in the Russian Federation and included in the State Register of Medical Products and Medical Equipment (Registration Certificate FS No. 2004/1128 valid until January 22, 2011), and also has a sanitary and epidemiological conclusion for use (License No. 77.99.28.944.D. 007272.12.04 dated December 28, 2004).

The procedure for performing mesodiencephalic modulation is simple and painless: a pair of fronto-occipital electrodes are applied to the patient’s head, through which specially selected therapeutic current pulses are applied, varying over time according to a predetermined program. The parameters of the electrical signals are programmed so that negative effects on the body are completely excluded.

The procedure lasts 30 minutes, the standard treatment course consists of 13 procedures (daily for 10 days)

A preliminary consultation and examination with a doctor who specializes in MDM therapy is necessary. In our clinic, all examinations and the procedure itself are carried out by certified specialists who have extensive experience and have undergone specialization abroad.

Indications for MDM therapy

• Cardiovascular diseases: hypertension stages I and II; coronary heart disease: angina pectoris FC I-III, rehabilitation of patients who have suffered acute myocardial infarction
• Endocrine diseases: diabetes mellitus and complications of diabetes mellitus (ulcers, gangrene, retinopathy, neuropathy); insulin resistance, dysfunction of the thyroid and parathyroid glands
• Somatic diseases: tonsillitis, bronchial asthma, chronic bronchitis, COPD, peptic ulcer of the stomach and duodenum; gastritis with secretory insufficiency and hyperacid state; biliary dyskinesia; bronchial asthma; rheumatoid arthritis
• Neurological diseases: cardiopsychoneurosis; dyscirculatory and traumatic encephalopathy; hypothalamic (diencephalic) syndromes; pain syndromes in diseases of the peripheral nervous system
• Surgical diseases: preparation for operations, rehabilitation in the postoperative period; burn disease and frostbite; post-traumatic conditions, non-healing postoperative wounds,
• Peripheral vascular diseases: endoarteritis and atherosclerotic occlusion of peripheral arteries; venous and lymphatic insufficiency, including complicated by trophic ulcers
• Gynecological diseases: menstrual dysfunction: premenstrual and menopausal syndromes; chronic salpingo-oophoritis, neuroendocrine disorders complicated by infertility or uterine fibroids (no more than 8 weeks), polycystic ovary syndrome
• Diseases of the uroandrological sphere: impotence; chronic prostatitis; symptom of chronic pelvic pain, infertility
• Dermatological diseases: neurodermatitis; itchy dermatoses; non-bacterial forms of eczema
• In psychiatry: reactive states; agripnic syndrome; neuroses; asthenic and depressive states; withdrawal syndrome
• For preventive purposes: under stressful conditions and prolonged emotional stress; with mental and physical fatigue; for chronic fatigue syndrome
• In pediatric practice: It is possible to use electrical stimulation starting from the age of 5, for all indications listed in other sections; and:
  • enuresis;
  • logoneuroses;
  • night terrors and other neurotic conditions;
  • adaptation to school and preschool institutions;
  • tonsillitis;
  • sinusitis;
  • increasing the resistance of frequently ill children during seasonal outbreaks of ARVI

Contraindications to MDM therapy

• the presence of metal foreign bodies in the tissues of the head;
• intracranial hemorrhage, danger of intracranial bleeding;
• schizophrenia;
• epilepsy;
• acute psychosis with psychomotor agitation;
• skin diseases at the site of application of electrodes on the forehead and back of the head.

Learn more about the Mesodiencephalic Modulation (MDM) method

Mesodiencephalic modulation or MDM therapy refers to physiotherapeutic methods, but significantly exceeds all known methods in terms of therapeutic effect. Its basis is weak, but complex in structure, electrical signals that selectively influence brain structures, activating the work of the control centers of the defense system. Mesodiencephalic modulation method is based on previous research in the field of therapeutic transcranial (through the skull) electrical stimulation of brain structures. Transcranial therapy was first performed in patients in 1902. Since then, various modifications of devices have been used in practical medicine, using a variety of frequency and other characteristics of electrical signals. The most famous in Russia are electrosleep, electronarcosis and devices for TES therapy. Over more than 100 years of clinical use, a huge amount of material has been accumulated, which makes it possible, first of all, to determine the parameters of the electric current, do not have a damaging effect on the human body, but improve the course of many diseases.

Unlike its predecessors mesodiencephalic modulation method, focusing the impact on the subcortical-stem parts of the brain (mesodiencephalic zone), managed to achieve not only an analgesic effect, but also achieved selective activation of the main regulatory systems - hypothalamic-pituitary, adrenal, opioid, etc. Accordingly, frequency characteristics, pulse shapes, method of application, the polarity of the electrodes and other indicators differ significantly from previous methods.

Like a new direction mesodiencephalic modulation appeared in the mid-1980s on the basis of the Center for Emergency Cardiology of the Research Institute of Emergency Medicine named after. N.V. Sklifosovsky. Testing a device for electrical anesthesia developed at the Institute of Physiology. I.P. Pavlov RAS, a group of scientists led by V.A. Pavlov discovered that the analgesic effect of electrical stimulation is not the main one. It was found that exposure of the mesodiencephalic zone to a weak electrical signal with certain parameters leads to the release of biologically active substances into the blood - opioid peptides (in particular, beta-endorphin - the “hormone of joy”), pituitary hormones and insulin, which reduce the severity of stress reactions and increase the adaptive properties of the body. That is, the mechanism of action is as follows: as a result of a stressful situation for the body (trauma, infection, allergy, etc.), the connection between the central and peripheral neurohumoral regulation of organs is broken, which in turn leads to the inclusion of the damaged organ’s own ectopic rhythm and its withdrawal out of control of the central nervous system. Mesodiencephalic modulation allows for selective activation of the regulatory structures of the brain; hormones are released that normalize the activity of organs and contribute to the restoration of full functional activity.

Further development of electrical signal parameters, which was carried out under the leadership of Professor V.A. Pavlov on the basis of European scientific centers, made it possible to achieve a significant increase in this positive effect. The result of almost 30 years of scientific work was the MDM 2000/1 computer complex. Programs developed by specialists for the treatment of various types of pathologies allow patients to obtain pronounced clinical and biological effects:

1. Anti-stress. It allows you not only to cope with stress in an extreme situation, but also to prevent the exacerbation of a chronic process accompanied by depressive disorders. In addition, the anti-stress effect actually reduces the number of complications in the treatment of various diseases and makes their course easier.
2. Reparative. Acceleration of repair by 2 - 2.5 times compared to the most modern medicinal and physiotherapeutic methods. For example, with a stomach or duodenal ulcer (including a “kissing” ulcer), myocardial infarction, burns, fractures, trophic ulcers, etc.
3. Analgesic and anti-inflammatory. A powerful analgesic effect allows you to cope with most of the existing types of pain (migraines, radiculitis, radicular syndrome in spinal osteochondrosis, pain in arthritis, toothache, etc.). This not only reduces the intensity of pain, but also relieves the inflammatory process that maintains this pain.
4. Prophylactic. Prevention of exacerbations in chronic diseases: stable remission (no exacerbations) is often observed even in such serious diseases as bronchial asthma, hypertension, diabetes mellitus, etc. At the same time, the dosages of maintenance medications are significantly reduced (MDM therapy enhances the effect of most known medications), Side effects and allergic complications are significantly reduced. In cases of drug intolerance, as well as in cases of chronic renal or liver failure, MDM therapy can be used as the only method of treatment.
5. Polytherapeutic. In the presence of several chronic diseases, a simultaneous therapeutic effect is observed on the entire list, which is especially important in the field of gerontology.

Make an appointment with a physiotherapist

Be sure to consult a qualified specialist at the Semeynaya clinic. This service is not provided in all branches, please check with the administrators for more detailed information.

(click on the picture to go inside the publication)

As organizations develop, they implement more and more information systems in completely different areas: accounting, personnel management, warehouse management, etc. Systems live and develop independently of each other until the very moment when a company needs to look at its data as a whole. Data volumes are already reaching a critical point and it turns out that it is simply impossible to collate and compare data manually. Decisions based on contradictory and unreliable data lead to management errors, and duplicates and irrelevance of data lead to incorrect business decisions.

Of course, the problem described above is not new, and today we will discuss the classic solution - a master data management system.

(clickable)

Types of enterprise data: what is reference and transactional data

To understand what master data is and is not, let’s look at the main types of corporate data.


(taken from here)

Unstructured Data- text, mail, and other data that does not have a formally defined and described structure.

Semi-structured- data that does not have a specific scheme (or has a variable structure), but nevertheless has a formal description in the form of tags and/or specific markers. XML is an example of semi-structured data.

Structured (transactional) data- data that has a formally defined schema.

Metadata- this is data describing other data, for example, a customer database schema, a configuration file or a report template.

Master data- this is data containing key information about the business, including customers, products, employees, technologies and materials. Each of these groups can be divided into several subject areas: the people category includes client, seller, supplier. It can also have a set of validation rules that the data must satisfy.

Example of the general structure of master data and validation rules (clickable)

Why is it needed?

Historically, many data storage, analysis and visualization systems developed in parallel and are not compatible with each other. As a company grows, data integration becomes more important and, in many cases, critical, and according to Microsoft, mid-sized companies are already feeling the effects of working with disparate data.
Thus, one of the tasks of MDM systems is data synchronization, which simplifies the solution of related tasks, such as the preparation of financial statements.

An MDM system is one of the cornerstones in business architecture, together with ERP and BI systems, allowing analytics and business systems to have a single view of data, regardless of source and form.

Let's look at a few classic cases where it is necessary to use and implement a master data management system.

Zoo IT systems and consolidated reporting

Let the company have more than three data storage and analysis systems. They are filled and develop independently of each other. At some point, it becomes necessary to collect consolidated reporting and it is necessary to synchronize regulatory and reference information. For example, there is a company Romashka with a turnover of 1M and there are two records “General limit. Romashka" and "Romashka LLC" in different systems with a turnover of 400k and 600k, without synchronization tools, the reporting system will not be able to combine records.

Systems integration

Suppose there are several 1C systems in the company’s branches and the invoices issued by Romashka LLC need to be uploaded and analyzed in CRM. If there are several duplicates in CRM, for example Chamomile and General. Ogre Chamomile, then the question arises to which Chamomile in CRM to link these accounts to and is there the right one among these Daisies?

Unified database of counterparties

First of all, the creation of a unified database is necessary for high-quality and reliable information about counterparties. If a client who has already signed a contract receives additional N calls about the need to send documents that have already been sent (since “Romashka General Limited” and “Romashka LLC” are syntactically different companies), then this negatively affects the company’s relations.

Data cleaning and normalization

The cases described above are tasks of data cleaning and data quality.

Data cleaning and normalization are certainly tools, the goal is to increase customer loyalty (e.g. avoiding repeat calls), creating reporting (confidence in the correctness of analytics) and increasing the speed of task completion (we move through the sales cycle faster).

As a rule, the client comes to the need to implement a master data management system. For example, the need for operational control over the activities of an enterprise may require the collection of consolidated reporting, which in turn will lead to the need to synchronize master data into the IT system, which in turn will require the implementation of a master data management system.

Cases from life

Fourteen 1C-ok

One company N had fourteen 1C systems in its branches, and then one day they had to urgently submit reports on their activities to some chamber there. The lack of unified reporting threatened significant problems, and so M employees spent several weeks collating and reconciling the data. Or they might just physically not make it in time.

Trucks

A client from Astrakhan sent trucks to a customer in another region, and the logistics along the way were provided by company X, which did not have an MDM system and a unified database of counterparties. During the trip, the trucks were serviced in two regions - and at the end of the trip, company X billed the client for these regions according to the standard price list without the required volume discount, since the client was registered in these two regions under slightly different conditions and the system did not matched the names. The result is additional proceedings and deterioration of business relations.

Repeated calls

One day, a client received a call six (!) times after the contract was signed. Due to such incompetence, the client's loyalty and contract were at risk.

Solution methods

Let's look at the two most popular methods for solving the problems described above.

Administrative decision

The administrative approach is to first clear out existing duplicates in IT systems, develop a coding system that can be used to compare entries in directories of different IT systems, and regulations. This method is relatively simple, but has a number of disadvantages - it will not prevent desynchronization of reference data in different systems, and regulations can always be circumvented.

Implementation of an MDM system

The technological approach is the use of a system that provides synchronization and a unified presentation of data. As a rule, most large companies implement various versions of MDM when manual consolidation of reference information and reporting becomes impossible, and the introduction of any new system forces changes in regulations and coding, only increasing chaos.

Of course, the one-time introduction of an MDM system will not solve all problems, and as business develops, the MDM system must also develop, and the type of MDM system itself may even change (the main types are covered below), however, as practice shows, MDM is the optimal business solution in similar cases.

Types of MDM systems

We will look at three main types of MDM systems - you can read more.

Centralized system

One IT system is selected; it can be either an existing IT system or a separate reference data management system. Reference data in this system will be considered standard, maintained in it and sent to other systems. At the same time, creating and editing reference data in other IT systems is prohibited. The advantages of this approach are:

• Ease of implementation;
• Ease of maintaining the relevance and purity of reference data in all IT systems, ease of administration and delimitation of rights;
• Current and clean reference data in all IT systems, which allows you to build clean local reporting in IT systems.

But this method has a number of disadvantages - in other systems it is impossible to create and edit records defined in the central system. That is, the internal business processes of the company change, which is often undesirable and sometimes unacceptable. The system is also unstable to communication interruptions and its performance critically depends on the current availability of the central system.

Analytical system

In the analytical master data system, all master data elements are created in client systems, from where they are sent to the master data system, where a master data directory record is formed from these elements. This allows you to quickly implement the system with minimal changes to client systems.

But since master data in a separate IT system is not synchronized with anything, there may be duplicates in the IT system itself and reporting may become blurred, so building operational reporting is difficult (local reporting is also said to be “dirty” - local records Master data may not correspond to records in the master data system).

Harmonized system

This system incorporates the best of centralized and analytical systems. It allows you to enter data in IT systems and then compare it with those already entered; it can search for potential duplicates, resolve conflicts associated with the simultaneous change of the same data in different IT systems, and synchronize master data in IT systems. In this way, business processes are not changed or disrupted, and manual work on reporting preparation is minimized - that is, local reporting is simply built. However, this approach is the most expensive, time-consuming and requires serious expertise to build, and may also require modification of client applications.

Examples of implementation of MDM systems

An example of an analytical master data management system is Navicon SalesOut, and an example of a centralized and harmonized one is different configurations of Navicon MDM.

Indicators of the need to implement MDM systems

Key: necessary integration various systems and unified reporting based on this data.

Particular prerequisites for implementation using the example of one of the clients

General indicators that make you think about the need to streamline master data and set up MDM processes:

• First of all, this is the presence or plans to implement several IT systems;
• Needs for automation of end-to-end business processes (i.e. processes that involve several IT systems) - need for integration;
• The need for consolidated reporting (i.e. reporting using data from multiple IT systems);
• IT strategy development. Many companies prefer to solve problems with master data before they arise. The longer reference data has been maintained in IT systems independently of each other, the more difficult it will be to verify, clean, and synchronize them in the future.

conclusions

Main theses and conclusions: synchronization of reference data facilitates 1) the introduction of new information systems into the company’s IT infrastructure; 2) integration of existing systems; 3) processing of corporate data; 4) reduces labor costs for updating data; 5) minimizes the risks associated with incorrect data. The implementation of a dedicated master data management system is not always mandatory, but it is always worth remembering the problems that may arise due to master data desynchronization when developing the IT infrastructure. Website

Story

On October 7, 2016, at the extraordinary general meetings of shareholders of BINBANK and MDM Bank, a scheme for the legal merger of BINBANK and MDM Bank was approved, according to which BINBANK will join MDM Bank, while the merged bank will continue to operate under the BINBANK brand (MDM Bank will be renamed BINBANK ). The process of legal merger of BINBANK and MDM Bank is planned to be completed by the end of 2016.

Owners and management

Chairman of the Board of Directors - Oleg Vyugin, Chairman of the Board - Mikail Shishkhanov.

Activity

PJSC MDM Bank provides a full range of services in the financial services market, including retail banking, services for small and medium-sized businesses, corporate, leasing and investment banking services.

Rating agency estimates

Write a review about the article "MDM Bank"

Notes

Links

see also

Excerpt characterizing MDM Bank

– Yes, I’m very happy about Nikolushka. He is healthy?

When they brought Nikolushka to Prince Andrei, who was looking at his father in fear, but was not crying, because no one was crying, Prince Andrei kissed him and, obviously, did not know what to say to him.
When Nikolushka was taken away, Princess Marya went up to her brother again, kissed him and, unable to resist any longer, began to cry.
He looked at her intently.
-Are you talking about Nikolushka? - he said.
Princess Marya, crying, bowed her head affirmatively.
“Marie, you know Evan...” but he suddenly fell silent.
- What are you saying?
- Nothing. There’s no need to cry here,” he said, looking at her with the same cold gaze.

When Princess Marya began to cry, he realized that she was crying that Nikolushka would be left without a father. With great effort he tried to return to life and was transported to their point of view.
“Yes, they must find it pathetic! - he thought. “How simple it is!”
“The birds of the air neither sow nor reap, but your father feeds them,” he said to himself and wanted to say the same to the princess. “But no, they will understand it in their own way, they will not understand! What they cannot understand is that all these feelings that they value are all ours, all these thoughts that seem so important to us are that they are not needed. We can't understand each other." - And he fell silent.

Prince Andrei's little son was seven years old. He could barely read, he didn't know anything. He experienced a lot after this day, acquiring knowledge, observation, and experience; but if he had then possessed all these later acquired abilities, he could not have understood better, more deeply the full meaning of that scene that he saw between his father, Princess Marya and Natasha than he understood it now. He understood everything and, without crying, left the room, silently approached Natasha, who followed him out, and shyly looked at her with thoughtful, beautiful eyes; his raised, rosy upper lip trembled, he leaned his head against it and began to cry.
From that day on, he avoided Desalles, avoided the countess who was caressing him, and either sat alone or timidly approached Princess Marya and Natasha, whom he seemed to love even more than his aunt, and quietly and shyly caressed them.
Princess Marya, leaving Prince Andrei, fully understood everything that Natasha’s face told her. She no longer spoke to Natasha about the hope of saving his life. She alternated with her at his sofa and did not cry anymore, but prayed incessantly, turning her soul to that eternal, incomprehensible, whose presence was now so palpable over the dying man.

Prince Andrei not only knew that he would die, but he felt that he was dying, that he was already half dead. He experienced a consciousness of alienation from everything earthly and a joyful and strange lightness of being. He, without haste and without worry, awaited what lay ahead of him. That formidable, eternal, unknown and distant, the presence of which he never ceased to feel throughout his entire life, was now close to him and - due to the strange lightness of being that he experienced - almost understandable and felt.
Before, he was afraid of the end. He experienced this terrible, painful feeling of fear of death, of the end, twice, and now he no longer understood it.
The first time he experienced this feeling was when a grenade was spinning like a top in front of him and he looked at the stubble, at the bushes, at the sky and knew that death was in front of him. When he woke up after the wound and in his soul, instantly, as if freed from the oppression of life that held him back, this flower of love, eternal, free, independent of this life, blossomed, he was no longer afraid of death and did not think about it.
The more he, in those hours of suffering solitude and semi-delirium that he spent after his wound, thought about the new beginning of eternal love that had been revealed to him, the more he, without feeling it himself, renounced earthly life. Everything, to love everyone, to always sacrifice oneself for love, meant not loving anyone, meant not living this earthly life. And the more he was imbued with this principle of love, the more he renounced life and the more completely he destroyed that terrible barrier that, without love, stands between life and death. When, at first, he remembered that he had to die, he said to himself: well, so much the better.
But after that night in Mytishchi, when the one he desired appeared in front of him in a semi-delirium, and when he, pressing her hand to his lips, cried quiet, joyful tears, love for one woman imperceptibly crept into his heart and again tied him to life. Both joyful and anxious thoughts began to come to him. Remembering that moment at the dressing station when he saw Kuragin, he now could not return to that feeling: he was tormented by the question of whether he was alive? And he didn't dare ask this.

What types of data are there?

Before moving directly to master data management systems, let's define what kind of data there are.

Below are the 5 key types:

1. Metadata;
2. Reference data;
3. Master data;
4. Transactional data;
5. Historical data.

Metadata is data about data. They are needed to understand and determine what data the enterprise operates on. Metadata defines structures, data types, access to them, etc. There are various schemes for describing metadata. For example, an XSD schema can be used to describe the structure of an XML document, and a WSDL schema can be used to describe a web service.

Reference data- This is relatively rarely changing data that defines the values ​​of specific entities used in performing operations throughout the enterprise. Such entities most often include: currencies, countries, units of measurement, types of agreements/accounts, etc.

Master data- This is the basic data that defines the business entities with which the enterprise deals. Such business entities usually include (depending on the subject industry focus of the enterprise) customers, suppliers, products, services, contracts, accounts, patients, citizens, etc. In addition to information directly about one or another master entity, master data includes relationships between these entities and hierarchies. For example, in terms of identifying additional sales opportunities, it may be very important to identify explicit and implicit relationships between individuals. Master data is distributed throughout the enterprise and is involved in all business processes. Typically, master data is perceived as a key intangible asset of an enterprise, because the effectiveness of its work depends on their quality and completeness. In Russia, instead of the term “master data,” the term “regulatory and reference information” is often used.

Transactional data– this is data that was generated as a result of the enterprise performing any business transactions. For example, for a commercial enterprise: sales of products and services, purchases, receipts/write-offs of funds, receipts at the warehouse, etc. Typically, such data is based in an enterprise resource planning (ERP) or other industry systems. Naturally, transactional systems make extensive use of master data when executing transactions.

Historical data is data that includes historical transactional and master data. Most often, such data is accumulated in ODS and DWH systems and is used to solve various analytical problems and support management decisions.

Master data management systems

Before moving on to the master data management system, let's define what master data management is in general.

Master Data Management (MDM) is a discipline that works with master data in order to create a “golden record”, that is, a holistic and comprehensive view of the master entity and relationships, a master data standard that is used throughout the enterprise , and sometimes between enterprises to facilitate the exchange of information.

Specialized master data management systems (MDM systems) automate all aspects of this process and are an “authoritative” source of enterprise-scale master data. Often MDM systems also manage reference data.

The situation when the MDM system is the only source of master data, all changes are made to the MDM system and only then transferred to consumer systems is called a “system of records”. This is an ideal situation for master data management. However, in real life things are not so simple: an MDM system will not always be a “system of records”. Due to the peculiarities of the business processes of a particular enterprise, the technical difficulties of specific systems, etc., it is necessary to create “copies” of master records. The system that contains a copy of the master data is called a "reference system". In order not to lose control, the “link system” must be under control and synchronized with the “record system”.

Three dimensions of MDM systems

Let's consider the MDM system in three dimensions:

Typically, MDM systems are not implemented “in a hurry,” because their implementation is a complex process of consistent enterprise-wide transformations, from maintaining disparate data to creating a holistic, comprehensive view of the master entity. Therefore, the implementation of MDM systems is carried out sequentially with a gradual approach to the target result in the three specified dimensions.

Let's take a closer look at these measurements.

Domains

In the context of master data management, a domain refers to a specific area of ​​master data. The most common master data domains are the customer domain and the product domain. In Western literature, established terms have emerged for master data management within these domains: Customer Data Integration (CDI) for the customer domain and Product Information Management (PIM) for the product domain.

CDI traditionally includes not only clients, but also organizations or individuals, which may be called differently depending on the industry of the enterprise: clients, suppliers, banks, funds, patients, citizens, etc.

PIM traditionally includes: products, goods, materials, services, works, etc.
There are many similarities in the CDI and PIM approaches to master data management, but there are also many differences. For example, when deduplicating client entities, in most cases a simple parsing of entity attributes and their comparison based on probabilistic algorithms is performed, while in the product domain, semantic/ontological analysis of attributes is carried out with the inclusion of self-learning mechanisms. In addition, in the product domain, entities can have very different attributes depending on the selected category (for example, laptops have their own set of attributes, while washing machines have their own). All these features of different domains must be supported by MDM systems.

Recently, there has been a tendency to create multi-domain MDM systems with the ability to flexibly customize the metadata structure. This flexibility gives an enterprise the opportunity to describe master data specifically for itself, taking into account all the features and nuances, but at the same time it requires a lot of time and knowledge to competently design and configure such a system. There are also systems on the market with a “rigid” structure of master entities that already have correctly configured mechanisms, but the use of such a system is possible only by those enterprises that can adapt to it. Typically, such systems are well suited for solving the problem of master data management within a specific industry. In my opinion, the most promising systems are those with a flexible metadata model, but at the same time having models pre-configured for enterprises in different industries that can be quickly reconfigured.

Methods of use

Methods of use MDM (Method of use) determine what the MDM system will be used for in the enterprise. In other words, who will be the consumer of the master data (naturally, there may be several of them).

There are three main methods of use:

1. Analytical
2. Operational
3. Collaborative

The analytical usage method supports business processes and applications that use master data primarily to analyze business performance, provide necessary reports and perform analytical functions. This often happens through the interaction of MDM with BI tools and products. Typically, an analytical MDM system works with data in read-only mode; it does not change the data in the source systems, but cleans and enriches it.

The operational usage method allows master data to be collected, modified, and used during the execution of business transactions (operations) and serves to maintain semantic consistency of master data within these operations across all operational applications. In fact, in this case, MDM functions as an OLTP system that processes requests from other operating applications or users. Working in this mode often requires building a unified integration landscape using the principles of service-oriented architecture (SOA) and using enterprise service bus (ESB) tools. It is ideal if such tools are either included directly in the MDM system or are its continuation (there are vendors who have both MDM and ESB solutions in their line, deeply integrated with each other).

The collective method of use allows you to create master entities in cases where collective interaction between different groups of users is required during this creation process. Such reconciliation typically has complex “branching” business processes consisting of various automated and manual tasks. Manual tasks are performed by various data scientists (data stewards) in the order defined by the business process. Most often, the collective use method is used in the product domain. For example, when creating a new product, when there are several people responsible for entering different data, there is a lot of manual work and final approval. It is important that the MDM system allows you to configure arbitrary business processes to quickly support the business processes of a particular enterprise.

Implementation styles

There are usually three main implementation styles:

1. Registry;
2. Coexistence;
3. Transactional.

The registry implementation style involves creating a master data source as a “link system” to lower data sources. A registry MDM contains only the key attributes needed to identify and map entities. Registry MDM operates in read-only mode, with data entered at source systems and passed to MDM for entity resolution. Also, the registry MDM may store links to sources of non-key data, but the data itself is usually not transferred to the MDM. The registry style of implementation is usually used when choosing the operational method of using MDM (see above).

The coexisting implementation style involves distributed data input across multiple sources (business applications and MDM system). The MDM system in this case can be a “system of records” only for part of the attributes. However, a full-fledged master entity is formed in the MDM system, the changes of which are transmitted to other systems (perhaps not all). The coexistent style of implementation is quite simple and is often used as the first step to the next one - the transactional style, because does not require deep reworking of systems interacting with the MDM system.

The transactional style of implementation involves the creation of a full-fledged “system of records” in which all data on master entities is stored. The MDM system in this case is the “single source of truth” for all consumer systems. All operations for creating and processing data are performed at the MDM system level. Data entry at the consumer system level is prohibited. This approach is usually quite difficult to implement, because requires significant changes to business processes and subscriber systems.

Conclusion

In practice, the choice of one or another MDM implementation strategy is determined by many factors: the goals of the enterprise in the field of master data management, the degree of maturity of the enterprise, the degree of readiness of the IT infrastructure, the availability of investments for the implementation of the project and many other parameters. To decide on an implementation strategy, you need to conduct a thorough analysis of all these factors and draw up a detailed feasibility study of the project and a detailed schedule indicating the phases of project development. But this is another broad topic that requires separate consideration.

One thing is for sure: the implementation of an MDM system must be approached very carefully and progressively. Most MDM systems implementation projects fail precisely because they underestimate the complexity and volume of changes that MDM projects have to face.