Morris County Library
E-ref Characteristics

January 2006 update: e-ref has risen 51.8% in the last two years--but, after seven years of practice, is still only 5% of our total reference service. We are seeing a migration from phone to e-reference. Other patrons still initiate by phone, but prefer an electronic reply (413 times this past year). Library web site had 293,757 visits in 2005 and 629,039 page views.

March 2005 update: a few more patrons are conversant with the service, tell their friends about it, but many are still astonished libraries offer this, are so "21st century?!" Ratio of "standard" reference:e-ref dropped from 3:1 to 2:1, 2003-2004. All reference, all formats, did not change year to year, but some shifted from phone/in person to electronic. Library business card introduced December 1, 2004 (going at 1,000/month) has increased visits to library web but has not produced commensurate increase in e-ref traffic. Questions seem to come 1:200, once in every 200 web visits. Only 1:10 views of the AskaLibrarian page results in a question. (We surmise that some patrons see the phone numbers and simply call.)

Oct. 2002 update: Very little of this has changed over the last five years. A small, regular cadre of e-patrons is now handled with dispatch; scanning to mail back has lately been added; Q & A are no longer saved, in the interests of patron privacy, but discarded at the end of each month, after statistical analysis. The questions are dissimilar, non-repeating enough that there is nothing to be gained by maintaining a FAQ file. Failed addreses (on mailback) are decreasing as more people seem to know how to express their e-mail addresses and are retrievable about 75 percent of the time. In 2002, e-ref is being counted in parallel with "extended reference" (questions take on phone or in person, written up for later call back--anything that can't be answered in five minutes or five sources, in other words.) Statistical measures are raw count, and service and efficiency rate(s).

These observations are based on three years of information referral in bulletin board format (10 hours/week), one and a half years in chat (3 hours/week, 20 person "room" capacity), and, at the Morris County Library, over eight years of e-ref via web form submissions.

  1. Why do people use this format?
  2. How does query tone differ from phone/in-person? Initial queries tend to be more explicit, detailed. Perhaps because patron is less obviously in position of supplication for help, queries tend to be less defensive or querulous.
  3. What are user expectations? Depends upon whether they are of your user community or not. If it is your native patron, service may be viewed as a right. If outside of your established service area, most are grateful for any response at all, much less assistance. Most do seem to be looking for at least acknowledgement of query within a 24-hr time frame (or by their next log-on, which may be longer.)
  4. Do they abuse the service? Repeat patrons are rare. A few, having established that the institution is accessible, helpful, then establish a more direct communication and come in or call. (Conversely, over 50% of requests for books via our online form have been placed by only two patrons.)
  5. What is best, first response? Attempt to answer their question, however fragmentary. If you want to engage them, to undertake a reference interview, give them something while you ask for details. Many are using this format precisely because they did not want to play 20-questions with a librarian!
  6. Length of response: do not over-reply; your answer may very well bounce. Many patrons do not know how to express their own e-mail, leaving off country code or domain name. Others with restricted, controlled lists (as AOL with parental controls on) do not realize that if they have not put your address on their permissions list, your mail will not get back in to them. Validate some channel of exchange before you spend more than 12 minutes on an answer!
  7. What about auto-responses? FAQs? Be very, very careful. If you need to resort to these, first ask yourself if you are staffing the function adequately? Few auto-responses read as genuine or engaging. Also, unless carefully coached, some staff begin to see which auto-response they can cram a question into, fitting the square peg into the round hole. It is not good service.
  8. Refer them to their local library? Rarely. They know nothing about it or are avoiding it or don't have one, which is why they are using your service in the first place. If you can confirm they have an adequate resource at hand then refer, but with specifics....titles, services to ask for. Explain ILL; give them the name of the directory or index to use. Give them the information they need to ask an explicit, effective question of their local information agency.
  9. What does staff need to know to be effective?
  10. Follow-up? Is it follow-up or is it stalking? Do a periodic sweep (once a week? twice a month?) of any loose ends, simply sending follow-up (this could be a template!). Response back has been rare. We don't follow patrons out of the building to their cars and we usually don't call back a week or two later to see if they are happy. Don't assume that because you can't see their faces you have to pester them to know if they are satisfied. Silence, to paraphrase Sir Thomas More, may betoken content.
  11. Count the stats? Why ever not? It is a visit to your institution, an attempt to use your resources.
  12. Volume? Generally vastly over estimated, even with publicity. Unique and national resources, of course, have a different experience.
  13. Positioning. Top level, top level, top level! Library should be a top level link, connection in the institution page or intranet and the query service should be top level within the library page, site.
  14. Think about: will you provide pointers, answers or documents (electronically)? We answer, advise, point, fax (to local patrons) and mail.) Will you e-mail the articles from the licensed database? How much service and material do you provide electronically? Can/will you scan material for mail back to the patron?
  15. If you open yourself up to a world of patrons, do you need to be multi-lingual? Have you the staff and fonts for this?
Relax and give it a try. Announce it as a trial, not a service, if you have doubts about how it will fit into your workflow. In most cases you need to get a sense of your user community, of demand before you can adjust policy (much of your extant reference and delivery policy will cover e-ref too), determine staffing, etc. As my blessedly flexible boss says "It's just reference in a different format."

The good news? It's fun, can create great goodwill for and knowledge of your institution and may contribute to the vitality and health of libraries as we meet and serve the next generation of users.

Statistical measure of e-ref
Service rate (SR)
Number of questions divided by (service population/100 or 1,000 or 1,000,000)
Which you use depends upon the size of your universe. Our county population served is 470,000, so we use 470. (470,000/1,000) for a SR per thousand population. For the US a better SR would be per million (288,315,865/1,000,000 = 288.315). A NJ Borough, population 1065, would do better to use 100s, as divisor 10.65, for SR per 100 population. Move the decimal back the number of zeros in your per population--2, 3 or 6 places. Service rate allows comparison of library services across formats and, even , across dissimilar (sized) patron universes. (Ex: in NJ, 2003: "standard" reference = 910 questions/1,000 (8.6 million people asked 7.8 million questions in 2003) and chat ref (including homework help) = 5.9/1,000). It may also serve as a gauge of best allocation of the library dollar. 2006: County SR increased by 2 questions/1,000 resident from 2003 to 2005.

Efficiency rate (ER)
This is meant as a review of process and training, a red flag as to collection strength and/or effectiveness of staff training. ER measures e-mail sent out for each e-ref received. If 25 questions were received and 75 sent out the ER would be 2: 75 total mail minus 25 original questions = 50 replies = 50/25 = 2. Some clarification and follow-up is expected in reference so a rate above 1 is fine, but if 2 pieces were going out for every question, the work process (are staff feeling too rushed to take time to compose one comprehensive answer?) and/or analysis of the original questions should be reviewed. An acceptable norm seems to be an ER of 1.3-1.5 (e-mails back for each e-mail in).

Sara K. Weissman
Reference Dept, Morris Cty Library
2 July 1997, updated 18 October 2002, 9 March 2005, 29 January 2006