MDM: A Cloud Perspective
Following up on my previous article couple of weeks back, I am going to define what MDM means in the cloud age. And why I think there is still more to be done, and while some vendors are on the right track others are not. I will, in the spirit of a positive dialog, not name any vendor, unless giving kudos. My intent is to layout the premise, as best I understand, and not throw anyone under the bus.
Business move to the cloud was driven by cloud innovators, followed by desire to cut costs in traditional IT investments. So, as that happened, expectedly, integrators moved in to integrate the clouds; first with on-prem, but later cloud-to-cloud. Unfortunately MDM products have followed the same path to innovation, the integration play; and that’s really incremental innovation at best. It’s a fair point, when someone says, MDM can help integrate operations around the master data, either cloud to cloud or cloud to on-prem. But that cannot be a sole driver; what’s the differentiator? Why should a company buy expensive MDM license over mere integrators.
Some vendors have literally migrated MDM products to the cloud, adding cloud centric basic necessities on top. Some others, who pioneered the cloud integration, have built basic and critical MDM features on top of their integration platforms.
All good; and part of the evolution of making cloud work. But, I want to look beyond, into the near future, and see how MDM can “break out” into the cloud economy. In the rest of the article, I will try and enumerate some use cases that seem exciting and promising.
If you are a company moving to cloud, your primary drivers are cost and consolidation; not to forget using best-of-breed cloud solutions as well. But let’s focus on the majority use case, cost and consolidation. So, as you move to a single business solution, we can see why many vendors, like Oracle, have moved their cloud MDM right near their application itself as a basis of holding their cloud offering together efficiently, I think its smart design pattern in lieu of multi-entity MDM. As they build more and more vertical solutions, these MDM offerings become central repository.
Naturally, not all vendors offer all solutions, as cloud evolution is still underway. Besides best of breed solutions, like SFDC for sales (Sales Cloud), are well established. Moreover, if you have a website (no pun intended), you have hosted it somewhere and depending on your business model and industry, you need all that data, telemetry or product feature usage etc., all the way to the orders, also harnessed and exploited; not to mention connected to cloud ERP. (Think of newer industry jargons like CDP which is competing with MDM.)
MDM’s role in this multi-cloud scenario, of course is to connect these applications around the thread of master data. But if you are a MDM vendor, you will get challenge from the left (CDP, DMP) and right (Enterprise MDM), and middle (integration vendors). So, you must offer many more bells and whistles.
I will not talk about security, privacy, scalability etc., as these are given, if you want to survive; and reason why we flock to larger and global vendors.
I will also not talk about solution pathways targeted to a specific business function like HR, Sales. There is a small gap there, but you can’t build everything, and multi-entity MDM tools fill that gap; there is not much ROI for new entrants.
Lastly, I will stay away from industry vertical solutions. From general product positioning perspective, I do see some vendors targeting verticals like life sciences (Reltio comes to mind), manufacturing/supply chain (SAP, Oracle come to mind), to gain market share. I agree with this strategy, because this allows these vendors to gain thought leadership in these industries. However, I want to focus on more on the generally applicable use case across most of the industries in the post-cloud era.
So, here you go.
First and foremost, MDM vendors needs to write some serious algorithms to discover and connect master data across and in the cloud apps, at least around product and customers. It’s not the same for each entity, hence mentioning each one here. E.g. Product have NPI, PLM, obsolescence, bundling, BOMs, upgrade and other relationships. Customers have leads, opportunities, orders, CX, customer success, customer support related use cases. So the discovery and connection should consider the business critical nature of each use case, and consequently offer, from touch-less to minimum-touch consoles. We, the MDM folks, must start using machine learning to its fullest potential, not just in one or 2 places, like quality match-merge or customer identity. Our primary role is to make enterprise efficient through seem-less operational integration. So, start with structured and enterprise (B2B) data first! It sounds clichéd, but you will amazed at how may discussions I have every month with different business leaders amazed at lack of sophistication in MDM toolset.
Next, connect the un-structured and social data around the master data. This is where business outcomes become front and center, as CDP vendors are show-casing. I think that MDM must build this meat on the barebones of connecting master data. There is a huge market here, but arguably many vendors too; partnerships may be easier at first, and MDM vendors can bring their years of industry experience to these partnerships. (I hope to take a closer look at CDP industry soon)
Next use case is cloud SaaS. MDM has key roles to play here. SaaS by nature means, tracking customer from leads onward, as they play with your product, opt-in/out; their product usage; and the whole CX around accurate billing, reporting, and finally renewal/churn. Notice, that in each of these scenarios, it’s all about your product and customer. How can MDM help a SaaS vendor do this best? I will be writing more specifically on this in coming weeks; but you get the general point.
Do not forget Big Data. By now, every IT department has a big data effort, whether on-prem or in the cloud (increasingly). MDM must play its role in connecting back the data findings, first to the analytical world (traditional cloud analytics and reporting), and increasingly to operational worlds. We have to be careful in labelling everything as Big Data, but suffice to say there are cases, like IoT or manufacturing defect analysis etc, where there is a big data churning, and I see no connection back through MDM into operational and analytical world. One thing to note is that MDM is not only a connector, but also can create benefit for the enterprises from holistic view of product or customer. MDM must innovate and integrate to “ingest” the data results of Hadoop (batch) or spark (or other streaming clusters) quickly back into the enterprise.
Another big data related use case is using MDM as basis for data cataloging for data lakes. Although seemingly tangential, if a MDM vendor wants to survive in the age of AI, this use case being bed rock of AI, becomes a must. Another use case is DQ around data science. We all famously know about Data Science Pareto’s law that a scientist spends 80% time of cleansing data and 20% on analyzing; there’s a space for MDM vendors, especially ones who are ahead of the pack in above use cases to venture their and help data scientist. A quick note, these are not easy scenarios, I understand, but basis of my argument is that MDM must be at the table or will be on it soon. I will give a special mention to companies like Reltio, who have championed above and other big data related use cases.
IoT is worth a separate mention as well. So much data in IoT is around product (sensors, equipment, and device) and geometric data that MDM vendors have to focus more on that field. I hope to write about this in detail too, but to keep it short here, I see some promising results in this area, but these are pointed and not clubbed into wider MDM.
If I am right, some vendors are going to pick on some of these areas (interesting value creation) and some others will keep doing what they were doing pre-cloud (and commoditize MDM). I am interested in the former. Good news is that there are sparks out there, but nothing has caught on fire yet. Do let me know if you something exciting.
