Visit FundSvcs | Print Page | Contact Us | Sign In | Join AASP
Best Practice in Bio Data Management

 

Best Practice in Bio Data Management

Date:  August, 2014                      

Prepared By: Terry Callaghan, Author PowerPoint, with Lynne Becker

Category:  Information Management                

Comments To: Terry Callaghan – callaghan@winants.rutgers.edu

Comment Period: September 1 to December 31, 2015

 

Description of Practice: Alumni database maintenance starting from scratch.

 

Prospective Users of Practice: Individuals who govern and steward the data in a variety of roles within Advancement; also those who manage and execute data acquisition and confidentiality protecting this asset.

 

Issues Addressed: Data seen as an asset needs clear definitions about what constitutes data from governance to stewardship including acquisition sources and solutions to the protection of the data through access and confidentiality policies.

 

Desired Outcome: A clear understanding about the steps required to establish and maintain the integrity, safety and functionality of the database.

 

AASP Recommendation: We will discuss the nuts and bolts of data including data’s value to us, data seen as an asset, data governance and stewardship, data sources, assessing data quality, data acquisition and data quality control.

1…Data’s Value

There are several components that contribute to the value of data. Data gives us proprietary knowledge. It allows us to synthesize information. Obviously we incur a cost to acquire data.  It provides us with a lifeline to our constituents as we are able to contact them. It allows and supports informed management decisions. We are able to measure performance. And we are able to do predictive modeling and plan for the future.

2…Data as an Asset

The value the data has to the organization (not the cost to re-create the data) includes having addresses, phones, emails, employment, occupation, philanthropic interests, wealth and assets, ratings and prospect summaries.

3…Data Governance and Stewardship

Data governance includes the policies and structure to ensure data quality and security. A data owner is a member of the business organization (prospect research, accounting, etc.) who is responsible for the quality of that data and dictates who should have access. Data governance also requires defined data roles such as view/update access and the actual release of data. Here is an example of a data access form:

Confidentiality policies play an important role in both of these roles: view/update access and actual release of data. Additionally, valid sources of data are established and documented. And all along data priorities are revisited and revised as the governance program matures.

 

Data stewardship involves those who carry out the data governance mission to ensure continued data quality and access. This involves daily “care and feeding” of data. We have a responsibility to keep it current. And through our commitment to data quality there is a constant effort to augment and improve the data.

 

4…Education about Available Data Sources

It is important to educate our business staff on the data that we can easily obtain from the organization and other sources.  The point to make is that any other data must be acquired with assistance from them.  Development and alumni relations staff usually have data within their organizations which can be very helpful when incorporated into our system of record. 

The graphic below explains how important current sources are for acquiring data in a higher education organization. Not only are the registrar and financial aid areas able to give information about current constituents such as alumni, Greeks and parents, scholarship recipients and market values, but also human resources for faculty and staff. The alumni directory provides information about employment, children, relationships and outside degrees. The US Postal Service delivers information about electronic address and phone updates (through the National Change of Address – NCOA). Athletics not only lets us know activities in sports but also season ticket holders. And, finally, wealth screening vendors provide asset data and modeling scores.  

5…Data Quality and Making an Assessment       

 

Defining Quality

Assessing data quality involves deciding what data we do have, how well we maintain it in the existing database and any inherited databases and reevaluating current maintenance procedures. This is important because if we can’t contact our constituents, we can’t engage, solicit and steward them. Approximately 15% of our population movers per year (US Census) and this varies by age as follows:

4% over the age of 65 will move to a new county

30% aged 20-29 will move to a new county

 If you have 443,000 alumni = 66,450 moves per year.

If you have   20,000 alumni =   3,000 moves per year.

This is increasingly important with our global economy, email, cell phones, etc.

 

Data Assessment  

What is our baseline? First measure it by determining our priorities and available resources. How much time and money do we have? Here is an example of a data inventory chart in order to establish our baseline.

 

 

Here is an example of a data quality/maintenance chart in order to establish our baseline.

 

           

 Developing a Plan

After determining baselines for data inventory and quality/maintenance, we begin by developing a plan to address our vulnerability. For instance under data quality and maintenance, let’s discuss address updates.

 

First – Do an Assessment 

In order to set goals and determine what data we would like to obtain, we should first assess where we currently stand.  For example, if alumni contact information is a priority initiative for our organization, we would want to determine some of the following:

 

How many lost alumni do we have?

What is our lost alumni rate?

What addresses do we think we have vs. how many are current and mailable?

 

The only true way to know is to mail to everyone. So once per year do a full mailing to the entire database and be sure to include “Request Returns” on the mailing. As a result of employing this protocol, Rutgers’ alumni magazine, which is mailed twice per year, has less than a 1% return rate.

 
Addresses – Return Rate and NCOA

The National Change of Address (NCOALink) has 160 million permanent change-of-address (COA) notifications in the last 18 months. This includes names and addresses of individuals, families and businesses who have filed a change-of-address. This database is maintained by the United States Postal Service. Access to it is licensed to service providers and made available to mailers. We must have our mailing list run by this service in order to qualify for Bulk Mailing Rate.

 Phone append services are also available through NCOA. Service vendors provide an enhanced service including additional databases for credit data, address verification and standardization, the phone append and, now, the cell phone append.

NCOA – Business Logic

Which addresses are positioned to be updated? All current home, domestic addresses of active entities with the exclusion of a handful of entities are updated. Those unique entities include coded incorrectly in NCOA like the wrong “Mary Smith” and a small number of “Do Not Contact Under Any Circumstances” (DNCAC).

               Then we have to decide which changes to accept. Always check dates and current status. If the NCOA effective date is more recent than the database address date, if address status is “Last known” or “Lost”, and if there is no home address, we accept the change. Standard protocol includes that after updates are made abbreviations are expanded and we verify and append phone numbers.

 

                As an example here is Rutgers University Foundation.

 

      700,000 addresses are sent for update to NCOA every 2 months (every 90 days for Bulk Mail).

 

      There are approximately

     8,000 address updates

     7,000 phone updates

     1,500 cell phone updates.

 

      In addition every piece of returned mail is manually looked up.

 

      There are approximately 100,000 address updates per year or about 15%.  This 15% is consistent with census statistics on the portion of the population that moves each year.  Checking our number of updates against the national average of moves is one way to gage if we are properly maintaining our system.  For example, if we only updated 5% of our database per year we might ask ourselves if we are missing some updates. 

       

International addresses have NCOA only available for six countries: Canada, United Kingdom, Germany, Australia, Switzerland and Austria. The international address format is different from US format. For other countries we can manage address updates through alumni reporting, some mail returns, social media/Web, and local clubs. 

It is very difficult to maintain address formats for potentially hundreds of countries for which we have addresses in our system.  In order to determine where to focus, we must ask our development, alumni relations and/or university global outreach staff what countries are most important.  It is better to maintain addresses for ten key countries really well than for 100 countries poorly. 

 

6…Data Acquisition

 

There are a myriad of ways to acquire data. We will present a Case Study to highlight particularly effective ways to obtain employment data.   

 

Employment Data Acquisition

            At Rutgers we obtained 50,000 employment records in one year. This was a 40% increase. We established a deliberate process and employed several sources including the alumni directory initiative, vendor appends, the internet and sometimes “bribery”, etc.

 

Here is a chart that outlines our sources with the percentages each provided for 45,890 updates during 2010-2011.

 

                  

The new employment data sources include 18% self-reported, 10% staff, 3% records initiatives, 56% Harris Directory, 2% corporate events, 5% Web search tools, 5% research and 1% records Web search tool.

            The alumni directory survey was sent to 350,000 alumni. 69,000 responded. The data obtained included 26,000 with employment, 27,000 with email addresses, 1,000 with other degrees and 2,000 with marital and children information.

Self-reported data is the most accurate data. And 8,000 self-reported outside of the directory survey through alumni emails and Website forms.

The Foundation and alumni relations, donor relations, records, faculty and staff’s initiatives resulted in 4,500.

Tracer postcards went to select groups such as alumni without emails.

Business card drawings were also held for staff and events.

We received 1,000 at corporate sponsored events through sign-ins at events and through corporation provided lists.

Web search tools provided 2,300 through daily information push, Google, Market Watch and corporate sites.

Prospect research through wealth screening provided 2,300.

            Market Watch is a financial news Website with business news, real estate sales, etc. Alerts from Rutgers alumni are received with high level alumni promotions, hires and similar life event notifications. Here is an example of a MarketWatch Alert.

 

            The vendor data append helped with alumni who had no employment data. 10,000 records were sent from the School of Arts and Sciences, Business and Engineering. 5,705 updates were received from the vendor. Of these 2,010 updates were confirmed and entered. That is approximately a 20-25% acquisition rate. It is a very manual process with confirming via corporate sites, Web searches, etc.

            Business cards, event spreadsheets, faxes and emails also provide information.  And, of course, sometimes “bribery” helps. One summer we ran a contest for our alumni relations and development staff.  Entry in to the contest was a business card or other biographic update.  They placed their name on the back (or even emailed their entry).  Each week we drew one name to win a gift card.  It is amazing what people will do for the chance to win a cup of coffee!!  Spreadsheets with hundreds of data updates suddenly appeared in our inboxes. (And, yes, they received hundreds of entries into the contest!)

We send an attractive advertisement to our development and alumni relations staff offering to sponsor a “fish bowl” drawing of business cards at their events.  We offer a gift of two beautiful 13 oz. Rutgers engraved wine glasses and 1 deluxe Rutgers University stainless steel opener set in a decorative box.  Advancement Services custom developed these items which are not available elsewhere.   We ask the events staff to send all updated alumni business cards to the Records Department for employment and contact updates. Here is a picture of our brochure:

                  

                  

             Our initiative for reinstating “Do Not Mail” alumni brought down our numbers to approximately 30,000 which is less than 1% of our alumni. We had sparse documentation prior to 2001 with an unusual number of prospects on the list. We decided that possibly the designation was historically used for a “special ask” list to prevent telefund contact for high level donors or just promoting “protecting pools”. In 2006 we implemented standard policies and documentation procedures and follow them today.

            We decided to try some test mailings to this group.  We wanted the first mail pieces to be “positive” mail from Rutgers, something that would be seen as a benefit.  Our first mailing was alumni membership cards which offered benefits at various vendors and businesses.  This was sent to 3,000 alumni coded “Do Not Mail” prior to 2006. There was very little negative feedback from this.  We received only 15 complaints and, in fact, received 2 requests to be added to the mailing list. In addition, the alumni magazine was sent 16 times as well as a recent Big 10 mailing.  We have now removed the “do not mail” indicator from these alumni (with the exception of those 15 who confirmed this coding).  After six months we will add them back into solicitation.  

 

It is a good practice to review your “do not mail” and “do not solicit” coding periodically.  

 

Do Not Contact Specificity

            Being specific about “Do Not Contact” is important because there are many different intentions involved. Ask for details.

      Do Not Mail – send no paper mail

      Do Not Email – opted out of email (can opt out by category – athletics, solicitations, etc.)

      Do Not Solicit – any solicitation

      Do Not Solicit by Phone (also by time of day)

      Do Not Solicit for Merchandise – credit card, etc.

      Do Not Send Magazine

      DO NOT CONTACT UNDER  ANY CIRCUMSTANCE

International Data Acquisition

            The University has a China data initiative and it is important in our identification of high potential suspects and prospects. The challenges involve incomplete registrar data including addresses, Chinese names vs. American names, no international NCOA for moves and changes, limited vendor data acquisition, the political changes among China, Taiwan and Hong Kong, and the Chinese language only Websites.   

            Some of our very aggressive and labor intensive action steps have resulted in over 1,000 confirmed addresses. We standardized addresses by streets and regions and took into consideration the geopolitical changes which involved Singapore and Hong Kong. It is always helpful to have a foreign national or fluency in the language to assist in name standardization and the understanding of the regions.

We now obtain birth country and country of citizenship data from our university registrar. This information is used to determine possible moves to China.  We employ a social media scanning service and manually use social media specifically LinkedIn and International Media (Chinese Facebook). Our contact and employment data for alumni living in China is now comparable to the statistics for our overall alumni population.

Email Addresses and Management

            Messages sent via email behave differently from letters sent to physical addresses. The physical address is static and does not change as frequently.  It is usually limited to one or two per alumnus. It is easier to determine if it is a valid address because mail will be returned. Consequently, this data is easier to maintain and validate.

            With email addresses they are very ephemeral and can change easily and often. They are unlimited in number and many people have three, four or more. Family members may or may not share. It is hard to tell if it is good or to whom we are mailing. There is a variety of potential delivery problems including server problems and out of office notifications, etc.  These problems are sometimes inconsistent with a “mailbox full” bounce being active only occasionally but still registering as a “bounce” on your email delivery reports.  Legal and anti-spam requirements present challenges.

            In order to effectively manage email addresses we need preferred vs. home information. We need to store separately form home/business address. Because there are many email addresses they may not be linked to a physical address so preferred email may not be linked to a physical address. And there are more email addresses than physical addresses. Additionally, families may not share and we need to manage different email addresses for spouses at the same physical address.  This requires some careful analysis of bounce data to determine valid email addresses. 

           

Types of Bounces

We have delivery problems and need to track “bounces” and bad email addresses differently from physical addresses. The number and types of bounces have to be considered. Is the address good or not? Did we reach the right person? Was it delivered? Did filters send it into spam, etc.? Additionally, we need to offer opt-out and opt-in options for email for legislative and anti-spam reasons.

Open rates rather than bounces are a much better way to gage whether we have delivery problems or not. At Rutgers we have performed several analyses of email bounce data. The example below is representative of our findings. 


In this analysis we sent three emails to 33,000 email addresses within one month. The bounces from each email broke down as follows:

Vendor – Message #1 – hard bounces: 7,777

Vendor – Message #2 – hard bounces: 5,678

University Listserv (all bounces): 6,666

The following chart shows that 3,333 bounces happened all three times.

                         Shared 3,333

 Analyze bounce back files periodically. Look at type of bounce such as Mailbox Full vs. No Such Address notifications.  Analyze bounces over time. An “out of office” or a “mailbox full” may exist for a few weeks, but if we attempt to mail to that address a month from now there may be no “bounce”.  In order to preserve the number of “emailable” addresses we have we provide correspondence preference options such as: do not solicit, no athletics, etc. As much as possible, customize content making it relevant to the recipient.

Email Acquisition Observations

Here are lessons learned from our email acquisition efforts with vendor append. Use individual match not household.  Under Can Spam legislation, the vendor must send an email offering an “opt out” to our constituent before providing us with that email address.  We can work with our vendor to have that message appear as if it is from our institution in order to alleviate concerns on the part of our alumni.  As an added measure we should consider sending our own opt-out email when we receive the email address.

Other email acquisition opportunities as we mentioned above for employment data acquisition include events registration, telefund and business cards.

 

Lack of Data as Data

            Lack of data is data for instance no children could mean no heirs and that is different from unknown vs. zero children. Always request an RSVP for an event. Even if they don’t attend we have data. When someone says not able to give now, but can give in the future, that is data. Lack of giving is a life situation. And the retired status is important data.

 

Incorrect Data

            The identification of incorrect data is data. No answer on the phone call, call to employer and they are not there, email bouncing, email opt-out, all inform us. We take this information and update our information by removing or marking “past”, search for an update or add to a bio research project.

We educate our alumni relations and development staff that “null data” as well as incorrect data is important information and should be communicated to the Records department.  This allows us to update the information (even if only to move the data to “past”) and also search for updated information.

 

Data Sources and Tips

            Anonymous Donor Coding

            There are three levels of anonymous coding:

·         Anonymous Entity – the institution does not know the donor

·         Anonymous Donor – donor is known but their gifts are anonymous (not released/publicized)

·         Anonymous Gift – this one gift from the donor is to be kept anonymous (not released/publicized)

As a data tip about anonymous coding, let’s consider public information issues. Does the state require “Anonymity Requested” as a condition of the gift so that we are allowed to withhold that information when public information is requested?

 

Do our donor recognition levels include anonymous giving and do we consider it especially for cumulative giving? Donor Relations should discuss this with our donors who are coded anonymous or who have anonymous gifts.  A donor may wish the gift to remain anonymous, but might want to be included in a broad donor recognition level.  The donor might also request to be put into a lower recognition level, as opposed to being omitted entirely due to anonymity.  This should be discussed with them when donor levels are published (annually, end of campaign, etc.)

 

Do we have policies to remove anonymity?  Who can remove this coding?   In cases where a donor is anonymous, who knows their identity? It may be important for some Advancement Services management staff to be aware of this so that pledge payments, matching gifts, etc., can be better stewarded.  And who knows details of anonymous donor and gift guidelines - Development Officers, Donor Relations, Gift Processing?

 

            Push Tech Reports

            These specialized data reports are emailed to business units and development officers. Push tech reports include gift and pledge transactions, contact reports, to-do tasks and deceased information.

 

New Relationships Acquisition

            Dormitory information is important for building out relationship mapping. At reunion registration Rutgers has available 200 plus records by year. We will continue to collect this information annually because it can prove to be useful in the event we are trying to reach out to a potential high level prospect. 

 

Events Data

            All data on events for which registration is performed online are fed to our development system.  In cases of high level events that are managed more personally by our event planners, this data is maintained in excel spreadsheets and uploaded into our system.  Rutgers has attendees for over 1,500 events coded; this is well over 70,000 attendees. This information is extremely important for development visits and engagement scoring.

 

Unstructured Data

            We find data in unexpected places such as event programs, bios, giving levels, press releases, wedding/birth announcements, community flyers, election notices, school flyers, book readings, etc. Some less obvious sources include unstructured data such as relationships with family members, children, board affiliations, neighbors, community groups/activities (such as golf, boating, etc.), school flyers, newspapers, mutual philanthropic interests, and major donor events. This unstructured data is also often hidden in contact reports.

            Membership in dean’s committees, boards, student groups, volunteer groups as well as awards and honors are important data. 

We educate our development and alumni relations staff that this data is valuable and can help them in their job when incorporated into our system.  We encourage them to provide this information to Records regardless of format (paper, spreadsheets, etc.)    

7…Data Quality Control

            What do we do when we find something that doesn’t look correct? Have a policy in place that allows team members to email the helpdesk. Send changes to records. Contact the Director of Gift Processing and Alumni Records. Research everything. And then provide the team member with more information on the data. Explain the way the data appears on a report. Let them know if this was an incorrect entry. Modify a report if needed. The key take away point is that feedback is very important and all questions/issues will be researched and addressed. 

            When incorrect information is found, Records will correct the data. IT will search the database for any other possible instances of the error. IT and Records will implement preventative measures by adding edit checking to any entry, reporting issues to the database owners for a fix and providing records staff training. Also, we implement an integrity report which searches for any possible anomalies in the data (for example, email addresses without an “@’ or without a valid extension - .com, etc.) Rutgers conducts regular integrity checks running over 50 reports on a weekly or monthly basis. Usually these are blank, but they are very helpful edit checks for data quality.

 

Sample Policies & Procedures/Resources:

Rutgers University Foundation, Terry Callaghan, The State University of New Jersey, Callaghan@winants.rutgers.edu Associate Vice President, Gift Processing, Records Administration & Information Technology, Rutgers University Foundation & Rutgers University Alumni Association 848-932-2001 – office

 

The Data Administration Management Association – DAMA www.DAMA.org is a very helpful resource for information on data governance, stewardship and quality.  


 

Get in Touch

225 West Wacker Drive, Suite 650
Chicago, Illinois 60606
Phone: 312.981.6798 | Fax: 312.265.2908
Email: aaspadmin@advserv.org