This is the fourth post of a multi-part recap of the 2013 e-Records Conference. Some presentation materials from the e-Records Conference are available on the e-Records 2013 website.
By Marianna Symeonides, Government Information Analyst
Kay Steed is the RMO for the Employees Retirement System of Texas (ERS), the state agency that administers and manages retirement plans and benefits for state employees. ERS may only be a medium-sized agency, but it’s one that generates a huge amount of records: over the years, they had accumulated around 1.5 million documents on their shared drive…records that nobody was reviewing for disposition. ERS decided that their solution to this all-too-common problem would be to implement an enterprise content management (ECM) system. Microsoft SharePoint would be the platform through which the ECM would be established.
SharePoint at ERS
ERS began the project in 2010 using SharePoint 2007. They chose SharePoint because it’s all about “collaboration, collaboration, collaboration”; plus, it was easy to deploy, had a familiar Windows interface, promised low IT overhead, and used SQL servers. In other words, it would “play nice” with ERS’s existing technological infrastructure. When SharePoint 2010 came out, IT went ahead and implemented it for the divisions that hadn’t yet implemented SharePoint 2007, in addition to migrating any records from SharePoint 2007 into SharePoint 2010.
The implementation of SharePoint was planned meticulously and included meetings, division-by-division demos and cleanups, training in the agency’s records retention schedule, and something called crawls. A crawl is the process of collecting information — through automated crawling software — about all of the documents on the shared drive: the file path, source path, and the title of the document.
Before each crawl, all of the documents would be frozen so that users could not go back and add, delete, or change anything on the documents. Then the IT department would perform the crawl, collecting information about all of the documents on the shared drive. That data would then be exported into a .csv file known as the crawl document. The ECM team would then add five columns to the crawl document that they would use to enter metadata later in the process. That data was subsequently migrated into SharePoint.
Ms. Steed illustrated the SharePoint implementation process by sharing the step-by-step experience of her division, Operations Support. First, they reviewed their documents on the shared drive and counted 9,000 documents. They had meetings for most of the summer of 2010 in which they opened up every single document. For each document, they determined who the document’s owner was. The owner then helped them determine which record series the document belonged to. Then, a crawl was performed on the whole division’s documents, and after the crawl, Operations Support staff entered metadata about each document into the five columns they’d added to the crawl document. After the metadata had been entered, they submitted the crawl document to IT, who then migrated all of the records (with their metadata) into SharePoint. Finally, the crawl document was used as a sort of validation list to make sure that all of the documents were successfully migrated into SharePoint.
Of course, a huge undertaking like this is not without its hurdles. A few challenges they faced included:
- Choosing the right records series. Classifying records according to a retention schedule is challenging anyway, but the difficulty is compounded in a digital environment. Many of the records series on their retention schedule were created long ago, when records were in paper, and classifying their electronic records according to that retention schedule often “felt like trying to fit square pegs into round holes.”
- Controlling access to sensitive records needed by multiple user groups. There wasn’t always a one-to-one relationship between a record series and a security group. For example, two user groups needed to be able to access Security Access Records: the group who handles the badges and keys that get you into the building, and the group who handles the keys that unlock desk drawers, doors, and filing cabinets. Both groups are dealing with the same type of record (which would be stored in the same place), but certain records within that folder had to be restricted to certain users.
- Too many records series. Although ERS “only” has 281 series now compared to 1100 records series ten years ago, it’s still too many series. The Finance division alone had 67 records series. It’s definitely more difficult to decide which “bucket” to put a record in when you’ve got 67 buckets to choose from!
- Too many retention codes. Anybody who is familiar with TSLAC’s retention schedules knows that we have a lot of event-based retention codes, such as AC (after closed), LA (life of asset), or US (until superseded). This multitude of event-based codes makes it difficult to automate the processes — someone must physically enter the event date for the record (e.g., when the record is considered “closed”).
Tips for Successful SharePoint Implementation
Going forward, ERS will focus on developing a taxonomy, maintaining consistency across divisions, and implementing disposition in SharePoint. But even though their work continues (and really, when is records management work ever done?), Ms. Steed had some advice for organizations looking to implement SharePoint or another ECM system.
Before you start:
- Define your problem – identify what isn’t working and why.
- Make sure you understand your business processes.
- Do a cost-benefit analysis of your options to help choose your solution. Do a needs assessment if necessary.
- Develop the plan – and plan each step in great detail. Develop a realistic timeline for how long it will take to follow the plan. It will almost always take longer than you think it will.
- Develop a governance plan, which will serve as a guideline for how the ECM solution will be used. Such a plan would include (but is not limited to) roles and responsibilities, site design, rules for use, navigation, and training.
When you migrate:
- Plan, plan, plan. If a project isn’t well executed, employees won’t use it.
- Think carefully about your Library (a SharePoint feature) structure – will you have a separate library for each division? Who needs to have access and rights to it? Will the library follow your agency organizational chart or will it be more collaborative?
- Clean up your shared drive before you perform the crawl. That way, you won’t gather data on or export any duplicates or non-records.
- Before the crawl, make sure to freeze the documents and identify any documents that won’t be migrated.
- Make sure your taxonomy and classification are consistent. Metadata terms need to be consistent: “Otherwise, you call it ‘potatoes’ over here and ‘sweet potatoes’ over there.”
Overall, Ms. Steed and ERS seem satisfied with their SharePoint implementation and they have seen some other benefits. Ms. Steed conducts a face-to-face training class four times a year, and the evaluations have been very favorable (which is not surprising since Ms. Steed was one of our panelists for our webinar on Developing Effective Records Management Training Programs for State Agencies). The SharePoint implementation has also resulted in a greater interest in the retention schedule, which is a dream come true for any RMO!
And finally, as is probably already apparent, RM and IT at ERS are working together. Collaboration between the two divisions was essential in order to be compliant with the retention schedule and to make information more accessible to users – neither department can do it alone. With the right mindset and a team of dedicated employees, implementing an ECM is a feasible undertaking that can potentially make records management much easier for an organization.
One thought on “eRecords Conference 2013: From Shared Drives to SharePoint”
“…[t]oo many retention codes. Anybody who is familiar with TSLAC….”
…multitude of event-based codes makes it difficult to automate the processes …”
Indeed, I was part of an RM group from San Antonio in late 1998 where we visited Austin and brought up this and several other issues to TSLAC which frustrated applying automation to the records management process.
Most of the original crew have long since retired or moved away.
The use of MS Word and PDF on the website did not faciltate automation. These were document meant to be read. The schedules were converted into *.CSV files by a protracted, tedious, manual process. The resulting files were imported into SQL tables. Appears to still be a challenge
Back then, my version of a crawl was an old fashioned MS DOS batch file using the DIR command. (e.g. dir i:/s/b/t:c/O:N >I_Drive_DB\dirimport.txt)
One major problem with a crawl or the old fashioned method is the lack of enforced naming conventions. Luckily — nowadays there are bulk rename utilities…B-)