Leveraging of big data for marketing can be real-time, on the fly behavioral analysis for cross promotions, or batch analysis and segmentation for targeting and nurturing.

In the first case – real-time – the goal is to influence consumer behavior at the point of sale or on the site. The most common example is the recommendation engine, but check out couponing also belongs to this category. For years, large organizations such as Amazon or eBay have used recommendation engines to match and recommend products, people and advertisements to users based on analysis of user profile and behavioral data. These problems were actually some of the first tackled by big data and have helped develop the technology into what it is today. Now, much smaller organizations can benefit from the processing power of Hadoop to process these large data sets and provide real-time recommendations.

In the second case – batch segmentation – massive amounts of data are analyzed, which that was just not possible or practical with traditional relational solutions.  Organizations are now able to better identify a target audience and identify the right person for the right offerings. Big data allows marketing teams to evaluate large volumes from new data sources, like click-stream data and call detail records, to increase the accuracy of analysis. Indeed, the more information made available to a marketer the more granular targets can be identified and messaged – but the traditional issue was that too much data killed data. No longer!

Starting with marketing in The New Use Cases of Big Data may seem far-fetched. But that’s where it all started. We owe big data technologies to the marketing needs of large internet organizations.  And not to worry, we’ll dabble into other domains in the next posts.

Yves

Among a wealth of enhancements to Talend’s next-generation integration platform, there are certain aspects of this new release that are worth highlighting:

• Version 5.1 includes the official GA version of Talend Open Studio for Big Data, now available under the Apache license. It was possible since March to download a technology preview of the product, and we have had lots of very positive feedback on it. Now the real thing is there… download it now!
• After introducing BPM as a piece of the holistic integration puzzle with version 5.0, this version reinforces the links between BPM and ESB, with a new connector that provides point-and-click connectivity between a business process and web service running on the ESB.
• This version marks the first synchronized release of all the products: data integration, data quality, MDM, ESB, BPM and big data – all leveraging the Talend Unified Platform technology foundation.

All products from the Talend Open Studio family are available for immediate download. Try them out now!

Yves

Like many other open source vendors, Talend was founded on the promise to democratize integration. The means to achieve this democratization is through open source that accelerates development cycles, promotes adoption, and enables companies of all sizes to use enterprise-grade technology without mortgaging their future.

Hadoop, its Apache ecosystem, and a number of tools such as Talend Open Studio for Big Data are making this democratization possible. The same way open source has democratized several IT domains that nobody wanted to predict would open to such technologies – the MDM market is a good example – open source is the key to big data democratization, or Big Data for the Masses (a great Talend whitepaper, registration required).

Stay tuned for some new use cases of big data…

Yves

While not all open source companies have been striving like Red Hat (becoming the first open source vendor to reach $1 billion in revenue), or MySQL (sold to Sun for $1 billion), or Talend (100%+ growth sustained for several years), the results are quite encouraging:

• Only 2 of the 50 companies have either closed or are on life support.  4% failure rate: I’d venture to say it’s not bad if you compare it to the overall IT industry.
• 6 companies have been acquired, or 12% – and I would say that at least 3 or 4 of these are striving in their new homes. I have no idea how that ratio compares to industry average?
• 8 companies downplay open source in their positioning or have forsaken open source all together. I would venture that these companies never understood what it meant to be open source, they only wanted the marketing benefits (real or supposed) to be open source but were not ready to make a commitment to the cause. I have written about this in the past, and all I’ll say is: good for them if they were able to reposition successfully.

Overall, not a bad track record.  Open source companies are doing well, thanks for asking!

Yves

While most organizations have been amassing big data for years, it is worth noting that some of them have been “doing” big data for years. By “doing”, I mean processing, leveraging, analyzing, mining – anything else than just storing it.  It’s one thing for Wal-Mart to process in excess of one million transactions per hour and store this historical data, or for the US Census Bureau to collect demographics on 300 million Americans – and it’s another thing for these parties to actually process it, massage it, and extract actionable information.

As we have all experienced first-hand, or heard of, a number of real-life (or rumored) big data use cases have been in existence for years.  Among some obvious and famed “big” use cases:

• Credit card fraud detection: anyone who travels a lot for business is often led to some unusual spending patterns.  Like buying French train tickets online from a US IP address, minutes after ordering a Kindle book and just before paying for an intercontinental flight. So when my credit card company calls me to check, I can hardly complain (and I am actually glad they call).
• Retail: the decades-old beer and diapers mining story (allegedly a rumored one), recently supplanted by the teen pregnancy one (it will probably take time before we know if this one is real or not!). Less prone to urban legend, the long string of coupons that print out at the register reminding you that you need dip for these chips, or conditioner to go with this shampoo, or mulch to help your flowers grow, are the result of big data analysis.
• Yield management: with airfares that vary on an hourly basis and hotel room prices continuously adjusted, the travel & hospitality industry has been a master of shifting through massive historical data to get the best possible price out of a finite inventory of airplane seats or hotel rooms – without leaving any unsold.

What are the commonalities between these “big” use cases? Credit card companies, retail chains, airlines & hotel chains are for the most part large and wealthy organizations. They did not wait for the recent big data frenzy to “do” big data. They invested a lot of money into hardware and software, hired the best talents, and built their infrastructure and algorithms, by trial and error.  And for the most part, they have reaped the rewards of this investment.

What’s new is the democratization of big data. In future posts, we’ll be reviewing some of the less “big” use cases of big data, and how big data can help smaller, less wealthy organizations.

Yves

Big data is first and foremost data.

In this post I will try to clarify some of the basic concepts and tools for big data.

MapReduce

At the foundation of these technologies is a concept called MapReduce.  It provides a massively parallel environment for executing computationally advanced functions in very little time, on a grid of commodity hardware.  Exit, the Crays and other massively parallel, helium-cooled supercomputers.

MapReduce allows a programmer to express a transformation of data that can be executed on a cluster that may include thousands of computers operating in parallel.  At its core, it uses a series of “maps” to divide a problem across multiple parallel servers and then uses a “reduce” to consolidate responses from each map and identify an answer to the original problem.

MapReduce enables technology for big data such as Hadoop to function.

Hadoop

Started at Yahoo! as an implementation of MapReduce in 2005 and released as an open source project in 2007, Apache Hadoop has the basic constructs needed to perform computing: a file system, a language to write programs, a way of managing the distribution of those programs over a distributed cluster, and a way of accepting the results of those programs. Ultimately the goal is to create a single result set.

With Hadoop, big data is distributed into pieces that are spread over a series of nodes running on commodity hardware.  In this structure the data is also replicated several times on different nodes to secure against node failure. The data is not organized into the relational rows and columns as expected in traditional persistence.  This lends to the ability to store structured, semi-structured and unstructured content.

Hadoop is an Apache top level project.

The Hadoop Ecosystem

A number of projects have seen the light of day around Hadoop, aiming at providing additional features. The main ones for processing big data include:

• Pig, to write complex MapReduce transformations using a scripting language, Pig Latin, which defines a set of transformations on a data set such as aggregate, join and sort.
• Hive, a data warehouse infrastructure built on top of Hadoop for providing data summarization, ad-hoc query, and analysis of large datasets.
• HBase, a non-relational columnar database, which provides fault-tolerant storage and quick access to large quantities of sparse data.
• HCatalog, a table and storage management service for data created using Apache Hadoop that provides a table abstraction to where and how data is stored.
• Sqoop, a set of tools that allow Hadoop to interact with traditional relational databases and data warehouses.

This post is meant only as a high level overview of tools for big data. For more detailed information on these technologies (and more), I would recommend a very good Talend white paper: Big Data for the Masses (registration required).

Yves

A clear definition of big data is also difficult to pin down.  What is big to one organization may not big to the next.  As the name implies, big data is characterized by the size or volume of data records, but other attributes need to be considered, such as velocity, variety, or complexity.  Most scholars of big data today tend to converge on the “3Vs” definition (Volume, Velocity, Variety), some will agree that these need to be augmented with the “C” of Complexity.

• V for Volume: this one seems pretty obvious. But it isn’t. Take Wal-Mart, whose stores routinely handle 1 million customer transactions per hour. Most organizations’ IT systems wouldn’t be able to cope, but Wal-Mart has been doing this for decades. What’s big for Wal-Mart and what’s big for – say – Consolidated Autodealers of Middle America, Inc. just isn’t the same. But it’s equally critical.
• V for Velocity: also called latency. How often do we need to refresh the data stores? How long are we willing to wait for new transactions to become available for analysis?  Velocity has a huge impact. Updating 1 million records every minute is a lot more difficult than updating 100 million records every day.
• V for Variety: unstructured and semi-structured (or poly-structured) content appears in all kinds of places; for example web content, Twitter/Facebook posts, customer comments, etc. And even with structured data, the type of sources can quickly become a challenge: all kinds of databases, SaaS applications, files, Cloud systems, etc.
• C for Complexity: let’s put it this way – if it was simple, if we knew how to do it with traditional technologies, why invent new paradigms? Big data opens new doors to what we can learn, extract from it. Translating this into computer programs is where the complexity resides.

One thing is certain: big data is not defined by a set of technologies; rather it defines a set of techniques and technologies that are used to make sense out of it.

In a future post, we will be looking at the tools and technologies for big data.

Yves

That makes Red Hat the second billion dollar open source company.

Who was the first billion dollar open source company? Am I such a dinosaur in the open source industry that nobody else remembers the headlines when MySQL was bought by Sun?

• Infoworld: MySQL: Sun’s Billion-Dollar Baby
• Techcrunch: Sun Picks Up MySQL For $1 Billion; Open Source Is A Legitimate Business Model
• Wired: Sun Pays Up the Wazoo – $1 Billion – for MySQL
• Internet News: Sun to Nab MySQL For $1 Billion
• Computerworld: Sun pays MySQL ONE BILLION DOLLARS

Funny also that, in 2008, Infoworld called MySQL a Billion Dollar Baby, and that yesterday, just before the Red Hat results were released, Wired titled Red Hat Becomes Open Source’s First $1 Billion Baby.

Of course, you can’t compare a billion dollar valuation with a billion dollar revenue. Even I, in marketing, know that. And I don’t intend to downplay the tremendous work that the Red Hat teams have been doing, building a large, profitable and sustainable open source business. Red Hat is clearly the poster child of commercial open source, and we are all proud to have them lead the pack.

Still, the parallel is hard to miss. 4 years ago, a billion dollar in open source could only mean one thing: an exit, over a very hot technology or trend. Today, a billion dollar in open source is not different from a billion dollar in traditional software.

All it took was four years. (Red) hats off.

Yves

Not only working just as in “we trade logos for our sites” (even though we do that too), but working as in “we have tested, and proven that our solutions work well together”. In the case of MapR, our colleagues from R&D spent lots of time in the lab, testing all the connectors against a number of configurations.

For MapR’s distribution, we provide the HDFS, HBase, Hive, Pig & Sqoop connectors (like for the other distros), But MapR also includes a unique feature: Direct Access NFS, which enables high throughput streaming to HDFS. Combined with Talend’s real-time capabilities such as Change Data Capture (CDC) or even Talend’s ESB, it enables real-time streaming of pretty much any data to Hadoop, instead of batch loading. This can find some unique applications for real-time analytics, and Talend is very happy to also include NFS support in the set of MapR connectivity.

Another side benefit of this certification is that it “carries over” to EMC’s Greenplum MR: Greenplum’s Hadoop platform is indeed based on MapR’s distro. Of course, Greenplum has been a Talend partner for a long time, and we already provide support for Greenplum HD. The addition of Greenplum MR to our portfolio of technology makes our partnership even more valuable for our joint customers.

With this new certification, we are very excited to be providing more options for our clients embarking on big data projects.

Yves

Probably trying to make a quick euro, Corp Events (virtually unknown in our space) decided to jump on the big data bandwagon. After all, big data is kind of a hot topic today, isn’t it? And it’s not hard to do a conference, right?  Sell sponsorships to vendors, find a suitable venue, sell sponsorships to vendors, round up some experts, sell more sponsorships to vendors, fill the rest of the agenda with sponsor panels, sell more sponsorships to vendors – done.

What worked? The selling of sponsorships to vendors did. Pretty much all companies that deal with data (big or small), and have a French office, paid their due to Corp Events (Talend included). Of course, no vendor can conceive a big data event where they would not be visible!  Problem is that, a big data conference without Cloudera, Hortonworks, MapR… it’s kind of Database World in the 90’s without Oracle, Sybase, Ingres and Informix!  Sure, they don’t have French offices, and it would probably have been more difficult to convince them to come. But did Corp Events even try?

The filling of (the rest of) the agenda with sponsor panels also worked. It was funny how vendors were trying very hard to be polite to their colleagues, while scornfully dismissing each other performance claims.

The rest was a miserable failure.

Most non-vendor speakers were utterly irrelevant, or just ignorant. Tell me why the “sociological impact of big data” matters to IT professionals.  Vendors landscapes “updated every other week” could maybe include relevant vendors (heard of Google search?). Examples of open data such as (real-time) train station locations or Orange selling traffic conditions (that’s open, right?) are simply missing the point. I could go on (take a look at my tweets if you are curious, but be warned, they are not very kind).

Logistic was a disaster. Paris is not known for its congress infrastructure, but I can think of a dozen other locations better suited for a conference this size than the Cité Universitaire. Expo space was neither attractive nor inductive to visitor traffic (see for example: peak hour at Big Data Paris). Coffee and food would run out fast. The free Wifi network crashed, probably after the tenth attendee tried to get on it. And there wasn’t even an official #hashtag for Tweets!

I have been to many conferences (probably too many!) in my professional life. Not all were stellar. But even many of the ones put together by volunteers or non-pros were a lot better than this one.

The problem is that such an event will leave a bittersweet taste for attendees and sponsors. Next time a real event company tries to run a big data conference, it will be an uphill battle for them to fight. The big data community deserved better.

Yves

• In Network World, Alan Shimel commented on how Talend is contributing to the democratization of big data: “There are many folks hoping that Talend can do for Big Data and Hadoop what it has done for data integration in general. A reoccurring theme I have heard recently around Hadoop is that it is too hard and not mature enough for many organizations.”
• SearchDataManagement’s Marc Brunelli also picked up the complexity of big data, having me commenting that: “Our vision is that it shouldn’t require a Ph.D in MapReduce to implement your transformations.”
• DJ Walker Morgan preferred to lead with the licensing angle in The H Open, discussing the importance of “Talend’s Open Studio for Big Data moving to Apache licensing” and thus becoming a closer part of the Hadoop ecosystem.
• Christopher Tozzi wrote for The VAR Guy that “The Hortonworks partnership introduces perhaps the most complete open source solution for Big Data management yet.”
• I also had the privilege to be interviewed by Dave Vellante and John Furrier on theCube, Silicon Angle’s TV channel broadcasting live from Strata Conf.
• And last – but not least – our good friends at the BeyeNETWORK invited me to a Q&A Spotlight on Big Data Integration with Hadoop. Definitely worth the read.

This is by no means exhaustive coverage, just a few highlights on different angles of these announcements.

Yves

Online activity was overwhelming too. Keynotes were streamed live, allowing people all over the world to join. When I setup the #strataconf hashtag in Hootsuite, I didn’t realize I would not be able to follow all the tweets. Stop paying attention for 15 minutes, and there was no way to catch up. Almost 600 attendees had declared their Twitter accounts in their profile (see great post by Jason Sundram analyzing these Tweeps and their audience), there were 684 people tweeting on the second day (see awesome graphs of top Twitter influencers by Ali Rebaie, and this network map by Benedikt Koehler), and if anyone can find a way to count the #strataconf hastag… with all these data geeks I am sure someone has already done it. And of course, a lot of attendees are active on the social media scene, so blogging went full steam.

Strata is about Making Data Work. But this year, it was about Making Big Data Work. Sessions covered a very broad array of topics. It went from super technical topics, such as the architecture of Hadoop, how to write R, Javascript or JSON, and more – to enlightening presentations on why medical trial data is flawed, or how to show data using colors and saturation. Another recurring topic was the rise of data journalism – how today’s newsrooms rely on the parsing, analysis and display of (sometimes large) datasets – often publicly available. Data.gov seems to be the new Deep Throat.

The show floor was definitely the most technology driven part of the conference. Apart from outliers such as Facebook whose presence was not clear (selling private member data maybe?), all the key players of big data were there (sorry can’t mention everyone, review the full list here):

• Hadoop distros (Hortonworks, Cloudera, MapR, etc.)
• Big data infrastructure vendors (Amazon, EMC/Greenplum, Calpont, DataStax, Couchbase, Microsoft, etc.)
• Big data analytics (Jaspersoft, Datameer, Tableau, etc.)
• Big data integration/quality (Talend)

I have been to (too) many conferences and conventions in my career. Most of them are either technology, or business case driven. I have to say I am impressed by the mix at Strata. There was content for everyone. From the technologist who was looking for a crash course in scripting, to the senior IT executive who wanted to understand the technology stack and business cases behind big data, to the business user (such as the “data scientist” – more on that character in another post) who was looking for best practices and peer networking.

Of course, there was an important milestone for Talend at Strata since we announced both Talend Open Studio for Big Data, and a strategic partnership with Hortonworks. Interest for these news has been tremendous – more on this later.

Glad I was able to attend. Guess I’ll be back next time.

Yves

Why? Not because of individual positions. We all argue that we are misplaced and that we should be much higher and to the right. That’s part of the process. Some vendors are more gracious than others, some even complain publicly, and of course we try to convince the analysts that “this rating can’t be correct”.

No, the reason why this Wave is making waves, is that, for the first time, an open source player joins the leaders. The establishment of proprietary vendors didn’t like it when Talend entered as a Visionary in Gartner’s Magic Quadrants. But visionary was kind of OK, they could still dismiss Talend as a bunch of crazy developers with a quasi-good idea for small projects but no enterprise scalability.

I haven’t always said kind things about analysts, but I have always considered them to be fair and honest. In the case of Forrester, I spent 4 years complaining that the ETL Wave was out of date (and it was).  So I was thrilled when Noel Yuhanna and Rob Karel initiated the Wave project last summer. And also excited about the transparency of the process. The Forrester Wave is probably the most transparent of all similar reports, with all ratings and weights being published and explained.

At the end of the day, once we were done arguing about ratings (and I don’t necessarily agree with all of them, but as I was saying, that’s how it works!), the calculation placed Talend in the Leader Zone.  Then, as is customary, the draft of the report was sent to the vendors, as a courtesy copy.

This is when it takes balls to be an analyst. When VPs of Marketing, CEOs and Directors of AR start telling you what you need to do. And why there is no way you can have Talend as a leader. First try to convince you, then threaten to cancel their research contracts. I am really happy Noel and Rob stood their ground, and that the Wave was published without changes. It proves that our industry analysts are doing their work.

If you want to check out the Forrester Wave for Enterprise ETL, in which Talend is positioned as a Leader (did I mention this?), you can get it here, courtesy of Talend.

Yves

This post was edited on Mar 2 to comply with Forrester’s Citation Policy. The postscriptum was also removed.

Hortonworks Data Platform, powered by Apache Hadoop, is a massively scalable 100% open source platform for storing, processing, and analyzing large volumes of data. It is designed to deal with data from many sources and formats in a quick, easy, and cost-effective manner. Hortonworks Data Platform provides an open, stable and highly extensible platform that makes it easier to integrate Apache Hadoop with existing data architectures and maximize the value of the data flowing through the business.

Thanks to Talend Open Studio for Big Data, users of Hortonworks Data Platform will be able to greatly simplify the deployment of Hadoop.  Talend Open Studio for Big Data abstracts the complexity of Hadoop and its “interfaces” (specifically Pig, HBase, Sqoop and Hive) by allowing graphical design of the big data integration jobs, and generating native MapReduce code. It alleviates the need for a deep, technical understanding of MapReduce and the different components of Hadoop. And, equally important, it brings to the table over 450 connectors to “the rest” of the information system – integrating enterprise data into Hadoop.

Hortonworks’ choice of Talend is a clear testament to the traction and recognition that Talend has been getting on the integration market, but also of the technical superiority of the solution. Our engineering teams have been working closely together to optimize the combined stack, and will continue to collaborate on further improvements to the solution.

Stay tuned for more.

Yves

Big data is getting easier

What does this mean?  Quite simply, that it no longer takes a PhD in massively parallel computing and data science to leverage your big data. Quite the opposite. Big data integration jobs can now be designed through a drag and drop user interface, and Talend Open Studio for Big Data generates the complex MapReduce code it takes to leverage your Hadoop clusters and run these complex transformations.

Big data is getting better

Talend Open Studio for Big Data is also part of the Talend Platform for Big Data, which adds enterprise-grade project management features and – very important – big data quality. Bringing data quality to big data is actually a first. The data quality dimension is not additive to the volume & complexity dimensions – it is multiplicative.  Big data exacerbates the data non-quality problems, to the point of rendering it unusable.

Big data is getting its data integration

Released under the Apache Software License, Talend Open Studio for Big Data is becoming a native part of the Apache Hadoop stack. Compatible with all Hadoop distributions, Talend Open Studio for Big Data is also becoming the key integration component of the Hortonworks Data Platform.  And that’s big, too. It makes Talend Open Studio for Big Data the de facto standard big data integration tool.

You can download now a technology preview of Talend Open Studio for Big Data. Don’t wait!

Yves

Out of the 37 vendors thus graded, only 3 got an A – and all 3 of them are independent, best of breed vendors. They are QlikView, Tableau Software and Talend. This Triple A is also a set of vendors that represent a newer generation of technology, one that is disruptive to established players, and that have been taking over the market at a very fast speed.  It is worth noting that none of the full stack vendors scored exceptionally well.  For complete results, the full report can be purchased from Decideo.

Unlike finance, when only extreme conservatism will get you the best grades, the BI market rewards innovation and disruption. Kudos to the Decideo members for recognizing this.

Yves

PS: in all fairness, despite lots of noise to the contrary, France did not really “lose its Triple A” – only one of the 3 major credit rating agencies downgraded the French debt. Still, it’s one too many.

I mean, who do we find there? Hadoop distros. Cloud. Management. Modeling. Data integration. Event streaming. Jim Kobielus, the Wave author, told Information Week that “It’s an apples and oranges collection of vendors”. I would go for a less fruity analogy: it’s more like an elephants and kettles collection (never heard this analogy before? come on…)

A number of players were omitted from this Wave. Microsoft, Oracle, Teradata, Informatica, Talend just to name a few – all have been offering Hadoop solutions for diverse amounts of time, some of them well before the Wave project even started – more than 6 months ago. And that puts us back at the heart of one of the critical issues plaguing analyst reports in general: latency. I have been arguing for a long time that such reports become obsolete before they are released.

The Hadoop Wave proves the point again. In a market that is moving so fast, how can you “freeze” offerings at a set date, and release findings more than 6 months later? I certainly understand that analysts have to juggle with many priorities and far too few resources, like the rest of us. But the system is inherently broken. This Hadoop Wave is already in urgent need of revision. Other markets are more mature – but this one needs real-time research.

Yves

PS: As far as the technical accuracy of the Wave is concerned, Curt Monash did a pretty thorough job of pointing out a number of technical aspects, and Derrick Harris from GigaOM offered an interesting taxonomy of Hadoop – saving me the embarrassment of trying to sound clever.

The following is an attempt to report via my tweetstream on 2 sessions held on Thursday morning at the #GartnerMDM Summit in London, as well as the guest keynote.

Taking Your Data Integration Competency To The Next Level In Support of MDM, by Ted Friedman

• #dataintegration is a critical piece of #MDM – the blood of the enterprise. @ted_friedman #GartnerMDM
• #dataintegration leverages separately designed data structures toward a common purpose. Good starting place for #MDM, right? #GartnerMDM
• Always baffled by this 37% of companies that have #dataintegration tools standards, but not enforced. #whatsthepoint #GartnerMDM
• Few orgs have more than 10 different #dataintegration tools. #relieved #GartnerMDM
• Canonical integration form is a path to #MDM – or it can be the other way. @ted_friedman #GartnerMDM
• 3 dims of #dataintegration: latency (real timeliness), granularity (batch vs trickle feed), physicality (moving vs virtual) #GartnerMDM
• Lots of M&A by big vendors combining app integration & #dataintegration, mixed results in terms of integration. #GartnerMDM
• Hype cycle highlights #opensource #dataintegration, #dataservices, convergence of #dataquality & data integration as key. #GartnerMDM
• “I can build best #dataintegration in the world, if I don’t include #dataquality I am just pushing garbage around” @ted_friedman #GartnerMDM
• Multiple vendors on #dataintegration #MQ include #dataquality as part of the stack. #Talend good example! #GartnerMDM

Applying Data Quality Best Practices To Maximize Trust In Your Master Data, by Ted Friedman

• So, how do you measure quality of data? @ted_friedman has a measuring tape on his slide. Wish it was that easy. #GartnerMDM
• Difficult to automate the measurement of data accuracy, but it is an objective assessment! #GartnerMDM
• #dataquality: Minimize risks, meet regulatory requirements #GartnerMDM
• #dataquality: Optimize costs, reduce waste #GartnerMDM
• #dataquality: Improve customer experience, maximize revenue. #GartnerMDM
• Most orgs have no idea of cost of poor #dataquality. Makes it difficult to justify cost of DQ program. #GartnerMDM
• #BigData makes #dataquality even more complicated. More sources, complexity, technology issues. #GartnerMDM
• RT @ted_friedman: #BigData is about more than “more”. Diversity, complexity, pace and other dimensions comprise the challenges. #GartnerMDM
• “Stay out of jail”: great message to justify #dataquality program! #GartnerMDM
• #dataquality workflow needed to support steward’s job. Think #BPM. Must be integrated – #holistic #integration. #GartnerMDM
• New entrant on #MQ: @Talend, coming at #dataquality from #OpenSource angle. Also has #MDM product. #GartnerMDM

Guest Keynote: The Art of Business Influence by Mark Jeffries

• Keynote by @markjeffries1 just starting: The Art of Business Influence. #GartnerMDM
• @markjeffries1 looking for sales people in audience. Aside from a few from vendors he won’t find many! #GartnerMDM
• We are all in sales, even though it’s a dirty word, says @markjeffries1 #GartnerMDM
• When somebody goes above & beyond expectations, you want to pay them back. Can be monetary – or not. @markjeffries1 #GartnerMDM
• Interactions always start with carefully balanced scales. Then they start tipping. @markjeffries1 #GartnerMDM
• LWAR: Listen Watch Anticipate React. In this order. @markjeffries1 #GartnerMDM
• Great example of a company that’s doing things right on Twitter: #SWA. @markjeffries1 #GartnerMDM
• And now, the 3 Cs: Consistency Clarity Confidence @markjeffries1 #GartnerMDM
• 7-8 NPM (nods per minute) are a sign to carry on. Nods stop – change topic. Btw it’s contagious. @markjeffries1 #GartnerMDM
• Creating jealousy effect is very good way to get people to come around to your thinking. @markjeffries1 #GartnerMDM

#TheEnd. Hope you enjoyed it. Don’t hesitate to leave comments!

Yves (@ydemontcheuil)

The following is an attempt to report via my tweetstream on 2 sessions held on Wednesday afternoon at the #GartnerMDM Summit in London.

Improving Business Intelligence Analysis Results With Integrated, Active MDM by Mark Beyer & Roxane Edjlali

• #BI & #MDM – the chicken egg question. Roxane Edjlali & Mark Beyer at #GartnerMDM
• Show of hands: most ppl in room do #BI, some do #MDM, none have combined them. #GartnerMDM
• In #BI data flow is unidirectional, flows from sources to data warehouse. #MDM features a circular flow – like app integration. #GartnerMDM
• Addressing issues after the fact – #dataquality impact on #dataintegration, changes to endpoints, etc. – should architect it. #GartnerMDM

Organizational Issues in MDM: Critical Roles and Stakeholder Management, by Ted Friedman & Debra Logan

• Who pays for #MDM is key issue of sponsorship – sponsor may not have $$ but you still need their support. Don’t scare them! #GartnerMDM
• Are execs reluctant to support #MDM program because they’re afraid they’ll be asked to pay for it? #GartnerMDM
• RT @mosesjones: #GartnerMDM @ted_friedman ‘master data by definition is shared’
• Data stewards must have authority and relevance. Don’t matter if they belong to business or IT. #GartnerMDM
• @ted_friedman explaining RACI framework: Responsible Accountable Consulted Informed. Alleviates many ongoing debates. #GartnerMDM

#MoreToFollow.

Yves (@ydemontcheuil)

The following is an attempt to report via my tweetstream on 3 important opening sessions of the #GartnerMDM Summit in London.

MDM Concepts & Introduction, by Bill O’Kane

• Been attempting to unify/master data for 30 yrs. Guess what? Biz success happened anyway! O’Kane #GartnerMDM
• #MDM value added: Run the Business < Grow the Business < Transform the Business. #GartnerMDM
• “Don’t boil the ocean, start small” – very wise advice from O’Kane at #GartnerMDM
• #MDM often the first foray into #SOA: opportunity for adding real-time. #Holistic integration? #GartnerMDM
• Exec sponsorship changes often fatal to #MDM. Sponsorship by CFO usually effective – less turnover in position. #GartnerMDM

Opening Keynote

• 420 attendees from 23 countries at #GartnerMDM. No #dataquality issue in country names this year?
• Cloud, information, social, mobile: unstoppable forces, or more opportunities? Or both? #GartnerMDM
• #SocialData & #Mobile: vast amounts of information generated about your company. How do you analyze it? Can’t govern but link. #GartnerMDM
• Lots of barriers to moving #MDM into the #Cloud. I think it’s more about bringing Cloud/#SaaS data into MDM. #GartnerMDM
• No high profile story of #MDM failure… But does MDM have the potential to be controversial & raise public interest? #GartnerMDM

Panel

(On the panel, two end users and Andrew White played the “Agree/Disagree” game on some key questions surrounding MDM.)

• Panel arguing whether #MDM projects can prove hard savings ROI. #GartnerMDM
• Panel agrees that #MDM is not an IT initiative. Still, need to give a chance to IT to make it work or it’s wishful thinking! #GartnerMDM
• If you don’t have all your data in the ERP, ERP can’t drive your MDM! #duh #bestofbreed #GartnerMDM
• Key to resilience of #MDM programs is to demonstrate value, beyond original sponsors. #panel #GartnerMDM
• #MDM not mature enough to deal with unstructured data. Will come later. #GartnerMDM
• RT @Paul_MDM: #GartnerMDM, more than 50% of audience think their BI provides bad results! (Obviously depends on job responsibilities)

#MoreToFollow.

Yves (@ydemontcheuil)