Organizations typically holds product data in three different kind of applications:
In an ERP application together with all other kinds of master data and transaction data.
In an MDM (Master Data Management) application either as a Product MDM implementation or a multidomain MDM implementation together with other master data domains.
In a PIM (Product Information Management) application.
Each of these applications have their pros and cons and where an organization utilizes several of these applications we often see that there is no single source of truth for all product data, but that some product attributes are controlled by one application and some other attributes are controlled by another application. Recently I wrote a post on the Pimnews think forum with a walk through of different kinds of product attributes and if they typically are controlled in PIM or ERP / MDM. The post had the title Six Product Attribute Levels.
The overwhelming part of organizations still use ERP as the place for product data – often supplemented by satellite spreadsheets with product data.
However, more and more organizations, not at least larger global ones, are implementing MDM solutions and, also midsize organisations, are implementing PIM solutions. The solution market was before split between MDM and PIM solutions, but we now do see some of the PIM solution providers also encompassing MDM capabilities. On the Disruptive MDM/PIM List there is a selection of solutions either being more MDM-ish or more PIM-ish as examined in the post MDM, PIM or Both.
Not at least Gartner, the analyst firm, has touted this as one of two Disruptive Forces in MDM Land. However, Gartner is not always your friend when it comes to short, crisp and easy digestible definitions and explanations of the terms they promote.
In my mind the two terms MDM and ADM relates as seen below:
So, ADM takes care of a lot of data that we do not usually consider being master data within a given application while MDM takes care of master data across multiple applications.
The big question is how we handle the intersection (and sum of intersections in the IT landscape) when it comes to applying technology.
If you have an IT landscape with a dominant application like for example SAP ECC you are tempted to handle the master data within that application as your master data hub or using a vendor provided tightly integrated tool as for example SAP MDG. For specific master data domains, you might for example regard your CRM application as your customer master data hub. Here MDM and ADM melts into one process and technology platform.
If you have an IT landscape with multiple applications, you should consider implementing a specific MDM platform that receives master data from and provides master data to applications that takes care of all the other data used for specific business objectives. Here MDM and ADM will be in separated processes using best-of-breed technology.
Product Information Management (PIM) have over the recent years emerged as an important technology enabled discipline for every company taking part in a supply chain. These companies are manufacturers, distributor, retailers and large end users of tangible products requiring a drastic increased variety of product data to be used in ecommerce and other self-service based ways of doing business.
At the same time we have seen the raise of big data. Now, if you look at every single company, product data handled by PIM platforms perhaps does not count as big data. Sure, the variety is a huge challenge and the reason of being for PIM solutions as they handle this variety better than traditional Master Data Management (MDM) solutions and ERP solutions.
The variety is about very different requirements in data quality dimensions based on where a given product sits in the product hierarchy. Measuring completeness has to be done for the concrete levels in the hierarchy, as a given attribute may be mandatory for one product but absolutely ridiculous for another product. An example is voltage for a power tool versus for a hammer. With consistency, there may be attributes with common standards (for example product name) but many attributes will have specific standards for a given branch in the hierarchy.
Product information also encompasses digital assets, being PDF files with product sheets, line drawings and lots of other stuff, product images and an increasing amount of videos with installation instructions and other content. The volume is then already quite big.
Volume and velocity really comes into the game when we look at eco-systems of manufacturers, distributors and retailers. The total flow of product data can then be described with the common characteristics of big data: Volume, velocity and variety. Even if you look at it for a given company and their first degree of separation with trading partners, we are talking about big data where there is an overwhelming throughput of new product links between trading partners and updates to product information that are – or not least should have been – exchanged.
Within big data we have the concept of a data lake. A key success factor of a data lake solution is minimizing the use of spreadsheets. In the same way, we can use a data lake, sitting in the exchange zone between trading partners, for product information as elaborated further in the post Gravitational Collapse in the PIM Space.
If you haven’t yet implemented a Master Data Management (MDM) solution you typically holds master data in dedicated solutions for Supply Chain Management (SCM), Enterprise Resource Planning (ERP), Customer Relation Management (CRM) and heaps of other solutions aimed at taking care of some part of your business depending on your particular industry.
In this first stage some master data flows into these solutions from business partners in different ways, flows around between the solutions inside your IT landscape and flows out to business partners directly from the various solutions.
The big pain in this stage is that a given real world entity may be described very different when coming in, when used inside your IT landscape and when presented by you to the outside. Additionally it is hard to measure and improve data quality and there may be several different business processes doing the same thing in an alternative way.
The answer today is to implement a Master Data Management (MDM) solution. When doing that you in some degree may rearrange the way master data flows into your IT landscape, you move the emphasis on master data management from the SCM, ERP, CRM and other solutions to the MDM platform and orchestrate the internal flows differently and you are most often able to present a given real world entity in a consistent way to the outside.
In this second stage you have cured the pain of inconsistent presentation of a given real world entity and as a result of that you are in a much better position to measure and control data quality. But typically you haven’t gained much in operational efficiency.
You need to enter a third stage. MDM 3.0 so to speak. In this stage you extend your MDM solution to your business partners and take much more advantage of third party data providers.
The master data kept by any organization is in a large degree a description of real world entities that also is digitalized by business partners and third party data providers. Therefore there are huge opportunities for reengineering your business processes for master data collection and interactive sharing of master data with mutual benefits for you and your business partners. These opportunities are touched in the post MDM 3.0 Musings.