Plus ça change, plus c'est la même chose. 


I was reminded of this French saying (Jean-Baptiste Alphonse Karr) when I attended the Box user conference, BoxWorks 2014, in early September. BoxWorks is not only a user conference but a platform to make new product announcements, allow third-party vendors to show their applications and generally whip the attendees and analysts into a frenzy.

What made me think about the saying about change was that Box announced product enhancements that included workflow, retention management, integration with Office 365 (previously a sworn enemy to Box) and enhanced metadata capabilities. Each of these announcements was cheered, and the many pundits present made instant celebratory blog posts. It was as if “workflow” and “retention” were new concepts just invented by Box.

For many in the crowd, who are younger, they may not have known that document workflow was started in the late 1980s by companies, like Plexus, FileNet, Optika and others, and was such a big deal that the word “workflow” became like “Kleenex” for our industry. Even today I still hear people refer to enterprise content management

(ECM) companies as workflow companies. In the early 90s, workflow was literally an industry rage with all major document management companies trying to one-up each other with new features and functions. Also, independent companies appeared that were workflow-only companies that could be integrated with a document management system. Workflow technology grew into a new industry, business process management (BPM), which still exists today but is not generally equated to workflow any longer.

To many of us who have been around the track a few times, workflow is not just old hat, but it is an old, dusty hat that never lived up to its promise.

In my version of history, workflow was overhyped, overbuilt and oversold in the industry and became yet another costly failure for the companies that bought it and tried to use it. Many companies were sold on the idea that workflow would literally change how they did business and provide staggering efficiency gains. This never happened, and over time, “business workflow” was largely forgotten. There are still companies that provide workflow (FileNet, Documentum, OpenText, Hyland, etc.), but these workflows are generally confined to very structured and repeatable processes, such as accounts payable (AP).

The problem was that workflow processes, by their very nature, are not stable processes with many sub-processes and people that frequently change. In order to establish a workflow process in the beginning, extensive work has to be done, including mapping out the process, as an AS-IS process and converting that to a TO-BE process. Then the process is migrated to a workflow engine and tested. As I did with one customer years ago, by the time we finished the AS-IS and TO-BE workflow diagrams, the process itself changed and the people within the process changed. This was not a simple AP process but a whole department (banking) devoted to processing individual retirement accounts.

MORE: Why ECM Systems Are as Complex as We Think They Are, or Even More Complex

In addition, most regular office workers neither have the time or the skills to develop and maintain workflow processes, even though a “selling point” at that time was that they were so easy to build and implement that any business unit lead and/or supervisor could do the work. This didn’t happen, and many resources were essentially wasted in the effort. Companies learned quickly that workflow, like optical disk storage and records management (Box’s retention),
was something to stay away from.

By the way, I am not talking about small businesses that have limited resources. Workflow failed at almost all levels of business from the SMB to the Fortune 100. This failure was, I think, one of the many reasons why the ECM industry has been challenged and has continued to be of a less-than-stellar success. Conversely, it is also one of the many reasons why companies are moving to cloud content management systems and applications.

So I am interested in the Box approach to workflow and metadata and wonder how they will be different. Their workflow is built on the conditional "if this, then that," and someone has to not only develop and map out the workflows, that part has not changed, but also implement the workflow itself. If my 20+ years of experience tells me anything, it will not be the average user who does this.

Not only do you have to map out the workflows, but these workflows may depend on a file having the correct metadata tag so the workflow can recognize it. This brings in another level of complexity in that many companies do not have an established taxonomy and would be hard-pressed to justify the cost of building one. Assuming that a company does have a taxonomy, the second question is, "Who will populate the fields, and who will maintain them?" This has been a traditional problem with companies, and it applies not only to “tagging” documents but to things like records management.

Box is trying to compete with complex systems that are, by all accounts, not very successful.

The problem is: Users do not do the extra work to tag documents and companies decline to hire information managers who can not only help users but do the maintenance on the workflow and the taxonomy. The maintenance process is not automatic.

Which brings us to something of a conundrum: Many companies adopt cloud systems because they have reduced their IT footprint or have limited IT resources; if the workflow/metadata development has to be done by IT or a “Box specialist” within the company—that may be a problem. I don’t know the numbers, but I would assume that many Box implementations are done at the user-level and IT supplies limited resources, such as connecting to AD, establishing security and other backroom duties. After the initial install, IT bows out and may only pass on trouble reports from users.

So as Box continues to add functionality like workflow and metadata, I hope they are also planning on how companies will implement that functionality and also maintain it. As an example, workflows depend on a role-based email address, such as, and many companies that implement Box are not going to build role-based emails because that would count as a seat license. Also, if your workflows depend on regular email addressing, such as, who will remember to change the email when Mary moves on to another department or leaves the company? This may seem like a trivial matter, but when you build 20 or 30 workflows with five or more role-based emails, you have built a process that requires a high level of maintenance.

As an example, I audited a bank customer who had extensive workflows (from a legacy ECM system) that were slowly falling apart. They were alerted to this when several customers called in and asked why their checks were deposited but their account didn’t reflect the new amount. Long story short, a change in the workflow process, by IT, started sending the check’s deposit information to a different workflow queue and the input clerks in that queue didn’t think that the work was theirs, so they simply ignored it and went to the next workflow item in the queue. (By the way, the checks were separated and deposited.) There was nothing in the workflow process to alert management that items were sitting in a queue for weeks without being responded to and no one quality assured the process after the changes were made. Sounds like simple mistakes and corrections, but if Box workflows are being set up and administered by users, I foresee problems in this approach.

My take on this is that Box is trying to compete with complex systems that are, by all accounts, not very successful. If you think about it, systems, like FileNet, Documentum and OpenText, are the reason why companies have adopted Box and the concepts behind enterprise file synch and share (EFSS)–because it is simple, users can manage it and it works. By implementing features and functions that raise them to the same level as the big ECM legacy systems, Box runs the risk of becoming the very type of system that helped them to be successful.

It will be interesting to see how successful these new products are and what their adoption rate will be. In reading all the analysts’ reports on these new announcements, I didn’t read any cautionary notes or any real analysis of the long-term implementation needs that Box will have to prepare for. I hope this is not a case of “The Emperor has no clothes.”


Bud Porter-Roth has over 20 years of experience as an ECM consultant, with a focus on electronic document management, records management and paper document projects. Follow him on Twitter @BudPR.