It’s that time of the year when all of the prognosticators, futurists and analysts break out the crystal balls and announce their predictions for the coming year. Not wanting to miss the fun, I am taking a whack at it myself but with a slightly more irreverent approach … with a Top 10 of my own. I hope this goes over as well as the last time I pontificated about the future with Crystal Ball Gazing … Enterprise Content Management 2020.
I don’t feel the need to cover all of the cool or obvious technology areas that my analyst friends would. A number of social media, mobile computing and cloud computing topics would be on any normal ECM predictions list for 2011. I do believe that social media, combined with mobile computing, delivered from the cloud will forever change the way we interact with content but this list is more of my own technology pet peeve list. I’ve decided to avoid this set of topics as there is plenty being written about all three topics already. I’ve also avoided all of the emerging fringe ECM technology topics such as video search, content recommendation engines, sentiment analysis and many more. There is plenty of time to write about those topics in the future.
Getting this list to just 10 items wasn’t easy … I really wanted to write something more specific on how lousy most ECM meta data is but decided to keep the list to these 10 items. As such, ECM meta data quality is on the cutting room floor. So without further a do …
Craig’s Top 10 Pet Peeve Predictions for 2011:
Number 10: Enterprise Search Results Will Still Suck
Despite a continuing increase in software sales and an overall growing market, many enterprises haven’t figured out that search is the ultimate garbage in, garbage out, model. Most end-users are frustrated at their continued inability to find what they need when they need it. Just ask any room full of people. Too many organizations simply decide to index everything thinking that’s all you need to do … bad idea.
There is no magic pill here, search results will ultimately improve when organizations (1) eliminate the unnecessary junk that keeps cluttering up search results and (2) consistently classify information, based on good meta data, to improve findability. Ultimately, enterprise search deployments with custom relevance models can deliver high quality optimal results, but that’s a pipedream for most organizations today. The basics need to be done first and there is a lot of ignorance on this topic. Unfortunately, very little changes in 2011, but we can hope.
Number 9: Meaning Based Technologies Are Not That Meaningful
Meaningful to whom? It’s the user, business or situation context that determines what is meaningful. Any vendor, with a machine-based technology claiming that it can figure out meaning without understanding the context of the situation, is stretching the truth. Don’t be fooled by this brand of snake oil. Without the ability to customize to specific business and industry situations these “meaning” based approaches don’t work … or are of limited value. Vendors currently making these claims will “tone down” their rhetoric in 2011 as the market becomes more educated and sophisticated on this topic. People will realize that the emperor has no clothes in 2011.
Number 8: Intergalactic Content Federation Is Exposed As A Myth
The ability to federate every ECM repository for every use case is wishful thinking. Federation works very well when trying to access, identify, extract and re-use content for applications like search, content analytics, or LOB application access. It works poorly or inconsistently when trying to directly control content in foreign repositories for records management and especially eDiscovery. There are too many technology hurdles such as security models, administrator access, lack of API support, incompatible data models that make this very hard.
For use cases like eDiscovery, many repositories don’t even support placing a legal hold. Trying to do unlimited full records federation or managing enterprise legal holds in place isn’t realistic yet … and may never be. It works well in certain situations only. I suppose, all of this can be solved with enough time and money but you could say that about anything – it’s simply not practical to try to use content federation for every conceivable use case and that won’t change in 2011. This is another reason why we need the Content Management Interoperability Standard (CMIS).
Number 7: CMIS Adoption Grows, Will Be Demanded From All Content, Discovery and Archive Vendors
Good segue, huh? If federation is the right approach (it is), but current technology prevents it from becoming a reality, then we need a standard we can all invest in and rely on. CMIS already has significant market momentum and adoption. Originally introduced and sponsored by IBM, EMC, Alfresco, OpenText, SAP and Oracle, it is now an OASIS standard where the list of members has expanded to many other vendors. IBM is already shipping CMIS enabled solutions and repositories, as are many others. However, some vendors still need encouragement. None of the archiving or eDiscovery point solution vendors have announced support for CMIS yet. I expect to see market pressure in 2011 on any content related vendor not supporting CMIS … so get ready Autonomy, Symantec, Guidance Software, and others who are not yet supporting CMIS. The days of closed proprietary interfaces are over.
Number 6: ACM Blows Up BPM (in a good way)
Advanced Case Management will forever change the way we build, deploy and interact with process and content-centric (or workflow if you are stuck in the ’90s) applications. Whether you call it Advanced Case Management, Adaptive Case Management or something else, It’s only a matter of time before the old “wait for months for your application model” is dead. Applications will be deployed in days and customized in hours or even minutes. IT and business will have a shared success model in the adoption and use of these applications. This one is a no-brainer. ACM takes off in a big way in 2011.
Number 5: Viral ECM Technologies without Adequate Governance Models Get Squeezed
In general, convenience seems to trump governance, but not this year. The viral deployment model is both a blessing and a curse. IT needs to play a stronger role in governing how these collaborative sites get deployed, used and eventually decommissioned. There is far too much cost associated with eDiscovery and the inability to produce documents when needed for this not to happen. There are way too many unknown collaborative sites containing important documents and records.
Many of these have been abandoned causing increased infrastructure costs and risk. The headaches associated with viral deployments force IT to put its foot down in 2011. The lack of governance around these viral collaborative sites becomes a major blocker to their deployment starting in 2011.
Number 4: Scalable and Trusted Content Repositories Become Essential
Despite my criticism of AIIM’s labeling of the “Systems of Engagement,” concept in my last blog, they’ve nailed the basic idea. “Systems or Repositories of Record” will be recognized as essential starting in 2011. We expect 44 times the growth of information in 10 years with 85% being unstructured, yikes! We’re going to need professional, highly scalable, trusted, defensible repositories of record to support the expected volume and governance requirements, especially as ECM applications embrace content outside the firewall.
Number 3: Classification Technology Is Recognized As Superior To Human-Based Approaches
For years, I’ve listened to many, many debates on human classification versus machine-based classification. Information is growing so out of control that it’s simply not possible to even read it all … much less decide how it should be classified and actually do it correctly. The facts are simple; studies show humans are 92% accurate at best. The problem is that humans opt-out sometimes. We get busy, get sick, have to go home or simply refuse to do certain things. When it comes to classification, we participate about 33% of the time on average. Overall, this makes our effective accuracy more like 30% and not 92%. Context technology-based approaches have consistently hit 70-80% over the years and recently we’ve seen accuracy levels as high as 98.7%. Technology approaches cost less too. 2011 is the year of auto-classification.
Number 2: Business Intelligence Wakes Up – The Other 85% Does Matter
It’s a well-known fact that ~85% of the information being stored today is unstructured. Most BI or data warehouse deployments focus on structured data (or only 15% of the available information to analyze). What about the rest of it? The explosion of content analysis tools over the last few years has made the 85% more understandable and easy to analyze then ever before and that will continue into 2011. BI, data warehouse and analytics solutions will increasingly include all forms of enterprise content whether inside or outside the firewall.
Number 1: IT Waste Management Becomes a Top Priority
The keep everything forever model has failed. Too many digital dumpsters litter the enterprise. It’s estimated over 90% of info being stored today is duplicated at least once and 70% is already past its retention date. It turns out buying more storage isn’t cheaper, once you add in the management staff, admin costs, training, power and so forth. One customer told me they’d have to build a new data center every 18 months just to keep storing everything. In 2011, I expect every organization to more aggressively start assessing and decommissioning unnecessary content as well as the associated systems. The new model is keep what you need to keep … for only as long as you need to keep it based on value and/or obligation … and defensibly dispose of the rest.
I hope you enjoyed reading this as much as I enjoyed writing it. I hope you agree with me on most of these. If not, let me know where you think I am wrong or list a few predictions or technology pet peeves of your own.