The Role of Selective Dissemination of Information (SDI) in Tailored Library Services

Every researcher knows the feeling of drowning in information. New journal articles, conference papers, patents, and reports appear faster than anyone can read them. Selective Dissemination of Information (SDI) was created to solve exactly this problem. Instead of leaving users to hunt through endless lists of new publications, an SDI service quietly tracks what each person cares about and delivers only the items that match. It is one of the most user-centred services a library can offer, and it remains highly relevant in the age of automated alerts and recommendation systems.

What is selective dissemination of information? 🔗

SDI is a personalised, current-awareness service that automatically notifies users about newly available documents matching their stated interests. Rather than the user searching a document collection, each new document is matched against a stored record of the user’s interests, and only relevant items are pushed to them. This reversal, where the information comes to the user instead of the user going to the information, is the core idea behind SDI.

The concept was first proposed by Hans Peter Luhn of IBM in 1958, the same researcher who developed KWIC indexing and the Luhn algorithm. Luhn described SDI as a service within an organisation that channels new items of information to the points where they are most likely to be useful for current work. He envisioned the use of computers for this task, and the first mechanised SDI system based on his design, known as SDI-1, was implemented in 1959 at IBM’s Advanced Systems Development Division.

Luhn’s proposal was a direct response to the explosion of scientific and technical literature after the Second World War. As the volume of research grew beyond what any individual could track, libraries needed a smarter way to keep specialists informed. SDI provided that mechanism, and it has since become an essential service in special libraries, research institutions, and academic libraries.

Why SDI matters for personalised information delivery 🔗

The value of SDI lies in its precision. A general alert about “new publications in chemistry” is of limited use to a researcher working on a narrow topic like lithium-ion battery electrolytes. SDI narrows the focus to exactly that topic. This saves time, reduces information overload, and ensures that important developments are not missed.

SDI is particularly useful for researchers, scientists, students, and professionals who need timely access to information in a specialised field. Because the service runs at regular, fixed intervals, users stay continuously up to date without monitoring multiple sources themselves. For a research scholar pursuing a doctoral thesis or a faculty member tracking developments in their discipline, this kind of targeted support can make a real difference to productivity.

Components of an SDI system 🔗

An SDI system is built around two key records that are matched against each other, supported by the technology that performs the matching and delivery. Understanding these components makes the whole service much clearer.

User profile 🔗

The user profile is a structured description of a person’s information needs. It typically includes the user’s name and contact details along with a set of keywords, subject headings, classification terms, author names, or organisations that represent their area of interest. A profile may also specify preferences such as document types, languages, and publication years.

Building an accurate profile is the foundation of good SDI. In early systems, librarians or information specialists would interview each user to construct a detailed profile. The profile is not fixed; it must be reviewed and updated periodically as the user’s research direction shifts. A profile that is too broad delivers irrelevant material, while one that is too narrow misses useful documents, so finding the right balance is an ongoing task.

Document profile 🔗

The document profile is a bibliographic record of each new document that enters the system. As documents arrive in the library, their content is analysed and the main concepts are expressed using the same keywords, codes, or subject terms that appear in user profiles. This shared vocabulary is what makes matching possible. A document profile usually contains complete bibliographic details along with the descriptive keywords that capture the subject of the document, and any number of keywords may be assigned as the document demands.

The technology behind SDI 🔗

SDI depends on software that can analyse, filter, and match these two sets of profiles. Search engines and matching algorithms process both the user’s interests and the document metadata to identify the closest matches. Automation tools then handle delivery, sending alerts to users without manual intervention.

Modern systems increasingly use artificial intelligence to refine this process, learning user preferences over time and improving the accuracy of recommendations. The same logic that powers academic alerting tools, such as the email notifications offered by databases like ScienceDirect or the saved-search alerts in indexing services, is a direct descendant of Luhn’s original SDI concept.

How the SDI workflow operates 🔗

The SDI workflow is a systematic cycle that moves from understanding the user to delivering results and then improving the system. It can be broken into a clear sequence of stages.

Profile creation 🔗

The process begins by ascertaining and analysing the needs of each user or group of users with similar interests. These needs are translated into keywords or codes and stored in a user profile file. Where several users share an interest, a single group profile can serve them all.

Document analysis 🔗

As new documents are acquired, their contents are analysed and indexed using the same controlled vocabulary. Each document’s subject facets are recorded in a document profile and stored in the document profile file. Maintaining consistency between the two vocabularies is essential, because the system can only match terms that are expressed in the same way.

Matching 🔗

At regular intervals, the system compares the document profile file against the user profile file. When a close resemblance is found between a user’s interests and a new document, the system records the relevant details from both profiles. This matching step is the heart of SDI, and the quality of the match determines whether the user receives genuinely useful information.

Notification 🔗

Once a match is identified, an intimation is sent to the user, usually as a list of bibliographic references or abstracts. Delivery may take the form of an email alert, a printed bulletin, or a customised feed, depending on the system and the volume of matches. The notification keeps the user informed without requiring them to search at all.

Feedback and refinement 🔗

A defining feature of SDI is its built-in feedback mechanism. Users assess the relevance and usefulness of the items they receive and report back to the system. This feedback is analysed and, where necessary, the user profile is adjusted to improve future matches. Over time, this loop steadily sharpens the accuracy of the service, which is something a simple one-way alert cannot achieve.

Difference between CAS and SDI 🔗

SDI is closely related to Current Awareness Service (CAS), and the two are often discussed together. In fact, SDI is best understood as a specialised, more advanced form of CAS. Both services share the same broad goal of keeping users informed about new developments in their field, but they differ significantly in scope and approach.

CAS is a generalised service. It circulates information about new publications, such as lists of recently received journals or contents pages, to a wide audience. Every user receives broadly the same information, and it is then up to each individual to scan the material and pick out what is relevant to them. CAS can be operated without a computer, for example by circulating the current list of periodicals among readers.

SDI, by contrast, is a personalised and pinpointed service directed at an individual or a homogeneous group. Instead of leaving users to filter information themselves, SDI states the exact information need in advance through a user profile and delivers only matching documents. It is fundamentally a computerised service that relies on matching the user profile with the document profile, and feedback is an essential part of the process. CAS does not require this matching or feedback step.

The simplest way to remember the distinction is this: CAS provides general information that the user must scan, while SDI provides a tailored list of documents selected specifically for that user. CAS answers the question “what is new in this field?” while SDI answers “what is new that matters to me?”

How the two services complement each other 🔗

Although they differ, CAS and SDI are not rivals. They work well together. CAS gives users a broad overview of trends and developments across a field, while SDI layers personalised, granular updates on top of that overview. A library that offers both can keep its users informed at two levels at once: aware of the wider landscape through CAS, and alerted to precisely relevant items through SDI. This combination allows users to stay current without being overwhelmed.

SDI in the modern library 🔗

While Luhn’s original vision relied on punch cards and early computers, the underlying logic of SDI is now embedded in everyday research tools. Database alerts, RSS feeds, citation-tracking notifications, and AI-driven recommendation engines all apply the same principle of matching stored interests against a stream of new content. For libraries, this means SDI is no longer a separate, labour-intensive service but something that can be configured and automated using existing database subscriptions and discovery systems.

For students and researchers, understanding SDI is more than an exam topic. It explains how the alerting services they already use actually work, and it highlights the librarian’s continuing role in helping users build effective interest profiles. As the volume of published research keeps growing, the value of a service that filters information down to what genuinely matters only increases.

What do you think? If you were setting up an SDI service for research scholars at your institution, how would you balance keeping profiles narrow enough to stay relevant against broad enough to avoid missing important work? And as AI takes over more of the matching process, do you think the librarian’s role in shaping user profiles will become more or less important?

References

1. https://en.wikipedia.org/wiki/Hans_Peter_Luhn
2. https://www.sciencedirect.com/topics/social-sciences/selective-dissemination-of-information
3. https://www.sciencedirect.com/
4. https://www.lisedunetwork.com/difference-between-sdi-and-cas/