I’ve been ranting on and off for a while that Information Management has failed because we haven’t met the needs of the user. This is leaving the market open for the Cloud vendors to try and disrupt the Content Management market.
What I haven’t delved into is that the primary reason we have been failing is also the key to the potential success of the Cloud vendors…Records Management.
Been There, Done That
If you’ve ever been on a Records Management project, you know the routine. There are lots of rules to setup and test. 
Users have to be trained how to properly classify a document and coached on why this is a good thing.
What invariably happens is that the project timeline stretches as the complexities of setting-up a full-blown Records Management system on top of implementing key business processes come to light. The Records Managers seem to always want more granularity and the actual users of the system want their lives to get easier and not more complicated.
In the end, all sides fail.
What does work is focusing on the business problem and agreeing to not delete any content until the next phase which will roll out Records Management functionality. Lucky for the end-users of the system, that phase usually gets cut to save money.
End the Misery
Records Management, at least as it is done in the United States, is a painful process that seems to live to make the lives of everyone involved harder. The DoD 5015.2 “standard” is criticized by many as being too complex, and outdated. Some past advocates are even voicing opposition to the standard as a whole.
I agree with those opinions. I strongly believe that the complexity of implementing a system designed for DoD 5015.2 easily triples the chance that your Information Management project will fail. Even when many of the requirements are ignored, it is too burdensome to utilize.
If we ever want to succeed, we need a new approach.
Removing the “Work”
Cheryl McKinnon recently wrote a piece outlining Five Trends Reshaping Records Management. The last point is probably the most important.
Autocategorization becomes viable and approachable
This is the key that will change things. Autocategorization can, with proper setup, correctly classify a majority of documents. The tools can also learn from documents to become more accurate in making their determinations.
These tools can also do this without impacting the people creating and editing the documents.
Think of a Records strategy that didn’t depend on everyone in the organization doing a little extra work to classify a document. Think of a Records Strategy that works even when a document is saved in the “wrong” place. Think of a Records Strategy that already has all your content categorized and protected so that an eDiscovery request can be met in days.
It is possible. It will just take people looking in from the outside to make it work.
This is where the Cloud vendors come into play. They have a desire to penetrate the Enterprise market. They have a strong focus on keeping the User Experience clean and simple in order to boost adoption. They aren’t bound by all the preconceived notions by which older vendors are bound.
It is time to step back from the huge list of functional requirements and get back to the core business requirements.
Word of Pie 
I’d say that Cheryl’s 4th point (Open standards and open source change the sourcing landscape) is probably more important in the context of cloud vendors than the others. Auto-Categoritagging can be replaced by workflow, since we’re really talking about mundane types of content.
LikeLike
I disagree Chris. I do think that point 4 is important, but I think autocategorization can replace what we are doing with workflows. Plus, not all content goes through workflows so it can’t catch everything. The very act of creating, defining, and operating workflows is also part of the problem we are trying to get away from as we move forward. Workflows definitely have their place, but requiring every record to pass through one is a no-win scenario.
LikeLike
Autocategorization works well for high-volume, low-touch content. Whether or not the content is a record at the point to categorization is irrelevant. The types of content that users are going outside the “sanctioned” solutions and using tools like Box are not high-volume, low-touch; it’s content that requires more thinking and collaborating. Workflows, combined with user and content profiles, classification models that make sense, and big bucket retention schedules will help. I’m not advocating that every piece of content needs a workflow, but you and I have both worked with tools that can be used to create routing & tagging type workflows pretty easily.
The big unknown to me is how to integrate the cloud based tools like Box, Dropbox, Google Drive, etc with the RM systems. I don’t have the answer, and at the moment I suspect the cloud vendors don’t either (note to cloud vendors – I’m happy to help you figure it out for a fee).
There’s definitely a place for autocategorization, but it needs to be deployed carefully. A true enterprise solution is going to require a blend of approaches and tools to be successful. What the blend is is going to be determined by a variety of factors and it’s going to change from org to org and even within orgs, based on geography, business unit, user needs, etc.
LikeLike
True story, after one and a half years just recently walked away from a client where the definition of that “granularity” had been on-going for five years. Implementation time the RM analysts wanted a one-to-one-one mapping from the content document class to the record class to the folder the document was filed in. Inevitable, gruesome implosion is a foregone conclusion. Too bad so much time and money will be wasted in the meantime.
On the subject of autocategorization, classification and workflow. We have multiple mechanisms wherein we can identify a document and slot it accordingly, yes? This is not a “one size fits all” paradigm or implementation. In all likelihood, in larger sites, we probably use one or two mechanisms primarily and maybe a handful more based upon the content composition and file plan’s governance.
I “think” even for the Cloud vendors it’s going to be about time to delivery and how well they execute. We’ll see.
LikeLike
Hi Lawrence, good food for thought! I started a comment here but it ended up in a diatribe, so I moved to to my blog instead: http://4most.wordpress.com/2013/06/26/clouded-view-of-records-and-classification/
LikeLike
Some good points raised in your post George and in the other replies. Two things to add…
Auto-Declaration/Categorization is not a one-tech approach. It needs to leverage what is in the content and the metadata that is captured as a matter of business needs. These are things that in normal business processes, automated or not.
As for Cloud, George is correct, it isn’t a solution for this. It IS a clean slate. The vendors can bring an outside perspective that none of us truly have anymore.
LikeLike
Hi Laurence (and all),
Very interesting discussion thread.
I will add my 2 cts and try to bring a different view of the RM/IG world as my original roots are more in the WCM/WEM industry and as I am then perhaps less-biased by 10 years of DoD implementations and deployments.
When I came into the RM world a few years ago and attended to what was my first information governance webinar, I only kept one single very simple but compelling message:
“Every successful organization needs to apply the key basic asset management principles to all its corporate information to reduce costs and risks and increase business value over time.”
I did not retain the subtle distinction between what was an information item vs what was a record (strange paradigm for a French guy like me in a country where this word and concept merely does not exist at all) or pm all the possible derived benefits of a possible IG Program (improved enterprise taxonomies; improved search and eDiscovery readiness; improved content security). What I only retained was:
– First assumption: Corporate Information is an Asset and needs to be properly managed over time. The fact that a document needs to be frozen as a record is only one action which can occur along a content lifecycle like many others possible ones (Move; Sign; Encrypt; Transform; Hold; Delete;…), some can be perhaps automated (autodeclare based on a given context), some others are manual, but clearlythe act of “declaring” a content is not the main and only criteria to trigger the fact that an information is or is not a “Corporate Asset”. And if we want to make it simple for the business lines this should translate also into simpler corporate business rules such as: “every corporate information needs to have an information policy applied”. Such a policy could be a very simple “Keep 6 months and auto-delete” which automatically applied by the content repository itself or a very complex rule including some time-based event triggering, some metadata anonymization and some other complex content routine and orchestration processes managed by another 3rd party system. But ultimately a rule has to be systematically applied on every content object. This is much simpler to understand and address in a company and usually everybody agrees on it (you do not buy a corporate asset such as a new laptop or a new table without having it somewhere in your corporate accounts – this is the basis of corporate accountability).
– This results in the second assumption: “Corporate Information Policy Management needs to be centralized but policy enforcement should be done in-place.” We are in 2013. The vision of a unique single ECM repository is a myth inherited from the end of last century. And the cloud is amplifying this phenomenon with additional new content silos which appear on a nearly daily basis. Combined to the deluge of information we assist to and the need to optimize content storage costs over time and across several possibly distinct content repositories this should push every sane information manager to rapidly adopt a federated information policy approach rather than yet another silo’ed approach. Or his whole content lifecycle strategy is doomed to fail. This means that it makes absolutely no sense to heavily couple the IG capabilities to a single unique ECM repository. Organizations need a distinct and repository-agnostic ILM solution.
– This comes to the third assumption: “Information Governance should be treated as a shared corporate service like LDAP or MDM.” IG should rapidly become a transparent and pervasive system available to any corporate content-enabled applications which is used by an organization.
Now if you combine that to the Cloud, you get a new value-added shared PaaS (Platform as a Service) layer, that we named IGaaS® at RSD, and which could be easily integrated to any content-centric applications be it in the cloud or on-premise.
I personally clearly do not believe we should repeat the error of the past and natively encapsulate all RM/IG capabilities into the new generation of cloud-based ECM system. We will just continue on the last 10 years of silo’ed approach. However it makes perfect sense to start considering IG as a supplemental and value added service available to corporate customers to RM-enable all their content and object stores.
Finally to close the loop on my first assumption, if information is an asset this also means that RM&IG will rapidly evolved towards a form of “Information Assets Management” tools which will ultimately not only deal with security classification, legal hold and defensible disposition but will support the overall “infonomics” (http://en.wikipedia.org/wiki/Infonomics) principles of a company. Record Managers and an adequate RM/IG solution are the best persons and tools in a company to make such a corporate information revolution happen. This is perhaps the real disruption factor for this market. Not the cloud per se.
LikeLike
Some excellent points and sound reasoning. Management in place is the current goal, but all the platforms have to support it before it can be realized. In addition, managing those records in place still requires being able to identify them without imposing on the users. That is currently very rare.
LikeLike