Contact Us

Why Master Data Management and Machine Learning Go Hand in Hand

Ramon Chen, Chief Product Officer, Reltio

Reltio’s inclusion in the The Forrester Wave™: Machine Learning Data Catalogs Q2 2018, by Michele Goetz with Gene Leganza, Elizabeth Hoberman, and Kara Hartig, Forrester Research, June 2018, sparked (pun intended) several questions. Such as why was Reltio included, how did we receive such strong marks, and why were we the only Master Data Management (MDM) vendor in the Wave?

The simple answer is that the Wave’s qualification criteria includes several key areas in which Reltio is naturally strong. As the only Master Data Management platform recognized in the Wave we believe our core MDM capabilities contributed to our strong showing. In fact, Reltio had already been included, together with 23 other excellent companies, in Forrester’s Now Tech: Machine Learning Data Catalogs, Q1 2018 report preceding the Wave.

That report outlined key 3 characteristics of Machine Learning Data Catalogs (MLDC):

1. Interpret, define, classify, link, and optimize the use of disparate data sources

Reltio is used by companies globally to define logical business schemas, capture and discover relationships through the Reltio Self-Learning Graph, suggest ongoing improvements, and to organize and bring siloed data together across the enterprise to meet their business objectives. This continuous reliable data foundation feeds better operational execution, predictive analytics, and sets them up to evolve towards a self-learning enterprise.

2. Reconcile policies across data use

Reltio Cloud’s built-in data security and privacy, regulatory, life-cycle, and data quality policies coexist to adapt data for multiple uses through a powerful audit trail, and role based access to data. This is critical in the face of evolving compliance measures such as GDPR. Flexibility and agility to ensure that you can track not just where the data originated, but how it’s being used and the outcomes it generates, is a critical component of any forward looking ML strategy.  

3. Democratize data to the edge of business

Reltio is particularly well suited to meet this requirement through frontline business user facing data-driven applications and workflow and collaboration capabilities that come OOTB with Reltio Cloud. It allows teams to submit comments, suggestions, filter and easily segment information through a UI that’s as easy to use as Facebook and LinkedIn.

Data science teams are then able to use Reltio IQ, with Apache Spark to run their algorithms without the pain associated with cleaning and onboarding data in separate environments. This is increasingly important as enterprises deploy machine learning systems, with data scientists requiring relevant, curated data sources to train algorithms to improve results.

As this video illustrates, the true value comes from being able to synchronize ML-algorithm derived IQ scores back into master profiles as aggregate attributes. Making them available for segmentation by marketing, sales, and even data stewards and compliance teams. Teams can then continuously reconcile results to recommendations in a closed loop to self-learn and improve outcomes.

We are tremendously proud and honored to have been included in the MLDC Wave as it reflects our core belief that machine learning cannot be used in isolation from the overall data organization and management needs of the business.

Whatever your desired outcome, MDM forms the backbone of high quality, reliable data which allows ML to thrive.

ML in turn provides unique capabilities to improve and increase the efficiency of data quality, and enterprise data organization operations. Like the graphic I selected for this post, they go hand in hand, and are interconnected across all points of the data continuum and life cycle.

Four Ways to Use GDPR as a Strategic Driver

Ankur Gupta, Sr. Product Marketing Manager, Reltio

Post May 25, 2018, per the General Data Protection Regulation (GDPR), companies with business ties to the European Union need to comply to GDPR standards. The cost of non-compliance is huge, but the regulation is meant to benefit individuals as well as businesses. Let’s look at what it can unlock for you and your brand if you approach it in the right way. What about being able to say that you are the safest enterprise in your marketplace when it comes to data? How about if you can not only reduce operational cost but can also create new revenue streams by being compliant to GDPR and other upcoming regulations?

1. Replace Legacy Systems by Future Proof Cloud-based Applications

When companies are taking steps to comply with GDPR, they are required to perform a ‘spring clean’ of their data, which can in turn lead to multiple efficiency gains. Organizing data improves the way firms carry out analytics and take business decisions. To comply with the regulation, companies must be able to illustrate the entire data flow – how data comes into the company; how they store and manage it; and how they treat it at end of life. This will encourage businesses to replace legacy systems by flexible cloud services to be more nimble and transparent especially when regulatory regime keeps evolving. In addition, most large enterprises have grown through M&A. Thus, they can look at GDPR as an opportunity to get rid of obsolete software and accelerate application retirement.

2. Gain Brand Loyalty and Attract New Customers

Companies can leverage GDPR to change the landscape from risk mitigation to improving their long-term competitive advantage. They can see early GDPR compliance as a competitive differentiator and position themselves as leaders of an emerging new normal. We trust those businesses who values our privacy beyond mere legal compliance. Thus, GDPR is an opportunity for businesses to get their data in order, get compliant and become consistently transparent with their customers. In a post-GDPR world, data sharing would be seen in the context of mutual respect and value exchange. It is an opportunity to re-connect your business with your current and potential customer base and start a new relationship based on mutual trust and responsible personalization.

3. Invest for the Future

The criticism that GDPR compliance might restrict innovations in AI ignores a subject’s right to privacy and consent. In fact, not being GDPR compliant would impose far more constraints on data collection and processing, slow down the ability to leverage innovations in AI and pay an opportunity cost such as market share losses in the future. Read this article for more details – Understanding GDPR and Its Impact on the Development of AI. In addition, in an era of data-driven innovation, business partners need to work together across the value chain. Data-driven innovation requires a clear understanding of the data to be collected and the reasons for collecting it. There are double opt-ins in such value chain: both partners need to be clear about what data they have about each other, and why. It’s very important that their data sharing practices are compliant with GDPR and other upcoming data regulations. As a first step to GDPR compliance, companies must define the scope of GDPR-relevant personal data that is collected, processed, and shared. Once a company identifies the scope of GDPR-relevant personal data, it should catalog all internal and external data sources that fall within this scope.

4. Execute A Delicate Interplay of Offense & Defense Data Strategies

In the post-GDPR era, personal data protection will become a data strategy issue. To comply, businesses need to have solid data organization and data governance in place. The GDPR gives companies the opportunity to holistically re-assess all their data, not just personal data. Data defense is about minimizing regulatory risk and preventing theft. Data offense focuses on increasing revenue, profitability, and customer satisfaction. Strong regulation in an industry (e.g. financial services or health care) would move the company toward defense; strong competition for customers would shift it toward offense. The CDOs and the rest of the leadership should see GDPR as an opportunity to establish the appropriate trade-offs between defense and offense to support company’s overall strategy. Read this blog post for more details – Is Your Data Strategy Defensive or Offensive? Why Not Both?.

Data is a company’s most important asset, and it’s constantly growing. Taking mandated compliance and turning it into an opportunity to personalize, delight and exceed customer expectations would fuel innovation reliably and responsibly.

How You Can Prepare for the Upcoming AI and Machine Learning Revolution

Ramon Chen, Chief Product Officer, Reltio

I was honored to be invited to participate on a panel discussing the evolution of Machine Learning (ML) at the Thomson Reuters Emerging Tech Conference. The panel consisted of luminaries such as moderator Jonathan Weber, Global Tech Editor, Reuters News, Asif Alam Global Business Director, Thomson Reuters, and Vikram Madan Senior Product Manager, AWS, Machine Learning.

We covered a lot of topics including deep learning, neural networks, image recognition, reliable data foundation as an ML imperative, digital personal identities, the increasing value of enterprise data, how you should safeguard your private data, GDPR, closed-loop as the last mile in ML, how LinkedIn is an example of the next generation data-driven application, autonomous data management and machine learning for data matching and correlation, classification of different types of data, Gluon and the Microsoft – AWS partnership, how elastic cloud computing with unlimited processing power makes ML a reality, and more.

Here are some key takeaways from the panel:

  • Machine Learning requires a foundation of continuous reliable data to ensure that algorithms are acting on the right information. Generating reliable data is usually the task of master data management (MDM) tools, that blend and correlate profile attributes across disparate siloed sources and applications. However MDM itself cannot deliver the complete picture as it’s missing the critical set of interactions and transactions that complete the 360-degree view. Today’s modern data management platforms go beyond MDM, and beyond data lakes that have been largely ineffective by providing a seamless feed of reliable data to maximize the potential of machine learning.

  • One challenge for, not just machine learning, but advanced analytics in general has been the friction of synchronizing data models between operational applications and data sources, and downstream data warehouses and lakes that are being used as the data pool for analysis and ML. Today it is possible to eliminate that friction by seamlessly transitioning information into Spark on demand, so that machine learning can operate on the latest, most up-to-date data, without the need to wait for data model updates and changes which have traditionally hindered business agility.

  • Another critical element of making sense of the output from machine learning and advanced analytics is closing the loop and bridging the gap between insight, action and outcomes. Today’s insights are still siloed from the actual actions that business teams will eventually take based upon the data. Further the outcomes of any actions take are rarely correlated back to the originating insights. The added value of a continuous feed of reliable data, relevant insights and recommended actions generated from ML is to have a closed loop where users can contribute to data reliability, and provide data on the outcomes, implicitly through their actions, or explicitly through feedback responses, so that ROI can be tracked, and ML has the historical data to learn and improve

It was a fun night with the audience contributing to the discussion. The future for AI and ML is a bright one. Everyone agreed that in order for such initiatives to deliver true value, a reliable foundation of data must be established, in order to ensure success.

Three Critical Ingredients for AI, Machine Learning & Cognitive Computing Success

Ramon Chen, Chief Product Officer, Reltio

You may already know a lot about Artificial Intelligence (AI), Machine Learning (ML), Deep Learning or even what some vendors call Cognitive Computing. Or maybe you are still trying to understand the nuances and differences between each term, and how they relate to each other.

Either way, it’s easy to be seduced by the “black magic” of technology that can solve a variety of your business challenges by just asking Watson, Einstein, Siri, Alexa, Hal (click for the iconic HAL 9000 scene in Space Odyssey) or other “humanizing” names.

In fact Gartner’s 2016 Emerging Technologies Hype Cycle has Machine Learning at the very “Peak of Inflated Expectations.” Those who are familiar with the cycle, know what is likely to come next <Trough of Disillusionment cough cough>. 

 

If you want to protect yourself from the hype, here are 3 critical ingredients for your consideration:

At Reltio we’ve been articulating a vision, which includes a pragmatic perspective of machine learning (ML) for over 3 years now. Realizing that not only does ML not offer a silver bullet, but there is still much to learn (pun intended) as to how such technologies can ultimately benefit both IT and business. Noted Big Data expert Bernard Marr provides a nice list of use cases that might be applied to your specific business and industry. The key is that a focused set of benefits for each users’ role, must be defined in order for it to be accurately measured so it doesn’t get labelled yet another (data) science project with limited value.

#1 Create a Reliable Data Foundation

Most companies are NOT ready for any form of AI, ML, or Cognitive Computing to help their business user, because their data is such poor shape to even attempt such an endeavor. Ironically, a great IT use case is to use ML to first help improve the consistency, accuracy and manageability for better data quality (DQ), uncovering patterns, anomaly detection and assisting humans, such as data stewards, to make their job more focused and efficient.

#2 Bring Analytics and Machine Learning to the Data

Just as the process of aggregating data to perform historical or predictive analytics is a cumbersome and expensive process, gathering and blending all of the right data that will guarantee machine learning is effective must be the in the DNA of any Modern Data Management Platform as a Service (PaaS).  

Bolting on AI or ML into legacy master data management (MDM) systems, or using such MDM tools to feed downstream disparate ML tools is putting lipstick on hosted managed services disguised as cloud. Reliable data, relevant insights and recommended actions via machine learning needs to be seamlessly combined into one, single multi-tenant cloud platform, architected from the ground up, for both analytical intelligence and operational execution, through data-driven applications.

Successful execution requires a closed-loop of all data, insights and actions, to ensure accurate metrication for continuously improved outcomes. Further, a multi-tenant cloud environment is the only way sufficient storage and processing capacity can be elastically accessed on demand to meet any business need.

Another benefit of a multi-tenant cloud PaaS is the potential to use a wide variety of anonymized data to help with machine learning across all industries. Having a large enough set of data is a critical factor for smaller companies to benefit from the right recommended actions, for common industry use cases.

#3 Don’t Go all in on one vendor

In a rush to market that “our tools do it too,” large vendors will unfortunately, over promise, and under deliver. It’s not their fault, as they must respond to the market, but many face an unenviable task of achieving ingredient #2 above, let alone attempting to now also execute on a plan to deliver their own AI technologies. 

An open ecosystem that allows you to choose and partner with the technologies, and domain experts of your choice is critical to getting the most out of a still young and evolving landscape. Most companies are already trying to evolve out of their legacy MDM platforms. Getting further locked into a single vendor, delivering both MDM and ML, through siloed disparate tools will not provide “clarity,” and may further complicate an already fragmented data management strategy.

At Reltio, we formed strategic partnerships with companies like QuintilesIMS to combine our strengths to jointly deliver on a vision for the next generation MDM and analytics capabilities for life sciences.

In summary, look to master your data in #1 for a reliable data foundation. #2 ensure that it covers all data types, sources and modes of consumption in a seamless feedback loop on a Modern Data Management platform architected from the ground up to avoid further siloing your data. Finally, #3 give yourself the openness and flexibility of your partners of choice to meet your business needs. 

You don’t want a HAL-like failure that prevents you from realizing your true goal of improving your business.

Why Cognitive Computing Solutions, Driven by Advanced Analytics will Replace Traditional Applications

Guest Blog post by: Judith Hurwitz, President & CEO Hurwitz & Associates

In this guest blog post, renowned industry analyst and author, Judith Hurwitz provides her POV on Cognitive computing, predictive analytics, reliable data and data-driven applications.

Like this post or add a comment to enter into a drawing for free copies of Judith’s latest book “Cognitive Computing and Big Data Analytics”

Cognitive computing is not a single technology or a single market. Rather it is an approach to solving problems by leading with data.  This means that rather than creating all the logic first and flowing data into a solution, you begin by analyzing the data to determine the patterns in that data. As more data is added, the cognitive system gets smarter and adapts to this new data. Therefore, it is no surprise that one of the most important aspects of cognitive computing is advanced analytics. Advanced analytics is defined as the collection of algorithms and techniques that leverages both structured and unstructured data sets to identify patterns.  There isn’t a single approach to analytics that is useful in a cognitive system. For example, information from neural networks, text analytics, sophisticated statistical models, predictive analytics and machine learning is all core to creating and managing a cognitive system.

By building the right predictive models that are able to react to changing business environments, companies will be able to prevent problems from happening before they occur. Analytics models must be able to take into account current information from customer interactions and respond quickly. The systems incorporate large sets of structured, unstructured and streaming data to improve the predictive capabilities. The sources of this data needed to make the models as effective as possible come from social media, customer relationship systems, web logs, sensors and video.

One of the most important issues for customers will be the trust in the data and in the results from a cognitive system. A cognitive system is only as good as the data that is ingested. Data needs to be analyzed so that the meta data is understood. This data needs to be refined so that its meaning is clear and the data itself is truth worthy. After all, the value of a cognitive system is that it creates an environment where industry experts can trust a cognitive computing system.

How are organizations beginning to use a cognitive computing approach? Given that cognitive computing is a young field, organizations are beginning with areas such as healthcare where there are huge volumes of unstructured data that cannot easily be analyzed or managed. Fraud detection is another important area where advanced analytics and the cognitive computing system can have a significant impact.  An insurance company may be faced with thousands of fraudulent damage claims. Using advanced analytics it is possible to create a model that can detect even subtle indications of fraud. The resulting cognitive system can detect new threats by learning from data even before they are even noticed by management.

While it is always possible to use analytics on a set of structured or unstructured data, a cognitive system can advantages.  A traditional application is based on an organization’s understanding of a process or business problem.  By the time that application has been written, debugged and implemented, the business may have changed significantly.  Contrast this to an approach the developer begins with a hypothesis and then trains available data.  If that hypothesis is not born out by he data then either the team will need to change the hypothesis or to bring in new data. The analytics process is allows you to understand relationships that already exist but have not previously been identified. Using machine learning greatly improves how effective predictive models are can improve accuracy especially when companies need to analyze large amounts of data sources that are primarily unstructured.

It is clear to me that a cognitive approach to advanced analytics will have a dramatic impact on hundreds of different market segments. When we have the ability to gain insights that is hidden and then apply learning to that data there is a potential to transform industries ranging from healthcare, to financial services, metropolitan area planning, security, and IT itself.  At the heart of business transformation is the ability to make sense of massive amounts of data. To be successful, this data has been refined in a way that creates trusted systems that have the potential to learn the secrets buried inside that data.

Judith Hurwitz is the President of Hurwitz & Associates, a consulting and research firm focused on important emerging technologies.  Hurwitz has co-authored eight books including Cognitive Computing and Big Data Analytics (John Wiley & Sons, 2015).