Wouldn't it be nice if there was an easy way for you to show your authors and readers exactly what was citing your content? CrossRef Cited-by Linking (formerly Forward Linking) is a service that allows you to discover how your publications are being cited and to incorporate that information into your online publication platform. Participation in CrossRef Cited-by Linking is entirely optional and there is no charge for it, but in order to participate, there is an important quid-pro-quo: in order to discover what publications cite your content, you must in turn submit metadata listing the works that your publications cite.

In practice, this is not hard to do. Reference metadata can be easily included within normal batch DOI deposits, and it is now even possible to deposit references for Cited-by Linking through the cut-and-paste Simple Text Query interface.

But, as is the case with all CrossRef initiatives, the utility of Cited-by Linking is directly proportional to the uptake of the membership and, although CrossRef staff actively promotes Cited-by Linking to its membership through webinars and email reminders, participation in the service has been limited. This, in turn, means that the utility of the "Cited By" information varies greatly across disciplines and publishers.

Although there are other services that provide cited-by facilities (e.g. Google Scholar, Scopus, ISI and some hosting providers.), we think that CrossRef Cited-by Linking complements these other services because:

1. It is controlled and managed by the CrossRef membership. As such, it enables direct primary publisher-to-publisher linking without the use of intermediaries.
2. It is not tied to any particular "metric", although the metadata provided by the service could be used by others to either validate existing metrics or as the foundation for creating new metrics.
3. It is not constrained by discipline and, as such, will prove useful for discovery of otherwise hard-to-find interdisciplinary citations and for use by publishers in fields that are undeserved by the other services.
4. It is not constrained by content type. CrossRef can accept reference data for journals articles, monographs, reference works, etc.
5. The CrossRef Cited-by Linking service is built on top of the DOI infrastructure and, as such, is very precise.
6. CrossRef is increasingly being looked to as a source for collecting authoritative publisher metadata. We are finding that those who seek to use CrossRef's metadata services are increasingly interested in collecting reference metadata as well. We worry that, if they can not get citation metadata from us, they will resort to using the much less accurate metadata that they can gather through screen scraping and services like Google Scholar.

Note: Because this is an optional service for our members and only a subset of the membership is currently participating, you will only be able to retrieve a partial list of the DOIs that actually cite your content. CrossRef cited-by links are not intended for use as a citation metric.


As of June 2008, publishers participating control roughly one-third of the total number of DOIs deposited in CrossRef:

• Articles with references deposited: 5,730,124
• Articles with at least one cited-by relation: 12,054,387
• Total number of cited-by relations: 78,213,626
• Journals with at least one cited-by relation: 15,074

To make use of this service, publishers will need to do the following two transactions with CrossRef:

1) Supply CrossRef with the reference list for each article deposited in CrossRef
2) For any given article you wish to show cited-by links for, query CrossRef (using its DOI) to retrieve the list of citing articles as identified by their DOIs.

Cited-by Linking Policies

1. Participation in Cited-by Linking is optional for CrossRef members.

2. Participants must deposit references from their current material and must retrieve data (i.e. retrieve “cited-by” links to link to other  participants).  Participants must be depositing references from current material before retrieving data. A Member will be deemed to have started participating in when they begin depositing references.

3. Members may only retrieve data to display in primary articles that they publish, whether for display on primary journal platforms or within their secondary products. Agents will be able to retrieve data on behalf of members who participate in and deposit references.

4. There will be no “public” interface for retrieving data. There is currently a form for free single DOI lookup on the CrossRef website but no equivalent form will be available for data.

5. There is no charge as of 2007 for participation in Cited-by Linking.

6. Cited-by Linking will not be available for secondary records, except as described in policy #3 above.

7. Cited-by Linking will only be available for member use in primary content.

Cited-by Linking information deposits are really nothing more than the list of references in an article's bibliography. Under CrossRef's normal linking service, members deposit an article's metadata and then -- in a seperate process -- query for the DOIs of the references. In order to do that querying, the references must be parsed or extracted from the article to enable construction of a CrossRef query. Cited-by Linking deposits simply combine the identity of the article  (the DOI in the metadata deposit) with its list of references (the list of queries).

Since the necessary data elements are already available, the only new step is to combine that data into a new CrossRef deposit. This can be accomplished in one of two ways:

1) Supply the reference list inside the metadata deposit.

Publishers currently supply their article metadata in an XML file formatted according to the CrossRef XSD schema version 2.0.5 or later (all are supported). For Cited-by Linking an updated schema has been created which is now at version 3.0.1. This schema provides for the optional inclusion of citation lists attached to the existing journal article metadata. Version 3.0.1 may be used by participating as well as non-participating Cited-by Linking members. Support for the older XSD versions will continue for the foreseeable future. The main body schema file now uses a common 'include' file that will be shared by variations of the CrossRef schema.

2) Supply the reference list separately from the metadata deposit.

Participating publishers who do not wish to modify their existing metadata deposit process may instead deposit the reference lists in a separate reference-only deposit transaction. A special reference only schema has been developed which allows for the deposit of references for an article that already has a CrossRef-registered DOI. This schema and the full metadata schema make use of same common 'include' file file to describe the references for an article.

In both approaches, an article's references are described using the <citation_list> XML element. This element contains one or more citation> elements (which is also common to both of the above methods). The <citation> element contains the same fields needed in a normal CrossRef query.

<xsd:complexType name="citation">
<xsd:all>
<xsd:element ref="issn" minOccurs="0" maxOccurs="1"/>
<xsd:element ref="journal_title" minOccurs="0" maxOccurs="1"/>
<xsd:element ref="author" minOccurs="0" maxOccurs="1"/>
<xsd:element ref="volume" minOccurs="0" maxOccurs="1"/>
<xsd:element ref="issue" minOccurs="0" maxOccurs="1"/>
<xsd:element ref="first_page" minOccurs="0" maxOccurs="1"/>
<xsd:element ref="year" minOccurs="0" maxOccurs="1"/>
<xsd:element ref="doi" minOccurs="0" maxOccurs="1"/>

<!-- book/conf. specific elements -->
<xsd:element ref="isbn" minOccurs="0" maxOccurs="1"/>
<xsd:element ref="series_title" minOccurs="0" maxOccurs="1"/>
<xsd:element ref="volume_title" minOccurs="0" maxOccurs="1"/>
<xsd:element ref="edition_number" minOccurs="0" maxOccurs="1"/>
<xsd:element ref="component_number" minOccurs="0" maxOccurs="1"/>

<!-- Citation text as it appears in the article , future placeholder -->
<xsd:element ref="unstructured_citation" minOccurs="0" maxOccurs="1"/>
</xsd:all>
</xsd:complexType>

A placeholder has been included in the <citation> element to support a future capability to process unstructured citations.

Note: the <citation> element is modeled after the <query> element found in the CrossRef XML query schema, which can now be used instead of the legacy pipe delimited query format.

Either of these deposit formats may be used and the process for submitting them is the same as that used for article metadata. The files are simply uploaded to CrossRef using the HTTP protocol as described on the system's help pages.

Retrieving cited-by links (cited-by articles)

Publishers participating in Cited-by Linking will be able to retrieve a list of articles that cite a specific target article. Only publishers that own the target article will be able to retrieve the list.

The primary method of retrieving forward links is to submit a Cited-by Linking query using the CrossRef XML query schema. A Cited-by Linking query request is just the DOI of the target (cited) article.

The <_query> element shown here also contains a few attributes that control exactly how the query is run. Setting the 'alert' attribute
to true will cause the system to send 'spontaneous' forward link results (via email) each time a new reference is deposited that cites the target DOI (the DOI in the <_query> element). The start-date and end-date elements can be used to limit the range of dates in which the reference was received. Using these attributes and a polling scheme, a publisher could request only those references
received since they last polled, as an alternative to receiving alerts.

<Settlement name="_query">
<xsd:complexType>
<xsd:sequence>
<xsd:element ref="doi" minOccurs="1" maxOccurs="1"/>
</xsd:sequence>
<xsd:attribute name="alert" default="false">
<xsd:simpleType>
<xsd:restriction base="xsd:boolean" />
</xsd:simpleType>
</xsd:attribute>
<xsd:attribute name="start_date" >
<xsd:simpleType>
<xsd:restriction base="xsd:date" />
</xsd:simpleType>
</xsd:attribute>
<xsd:attribute name="end_date" >
<xsd:simpleType>
<xsd:restriction base="xsd:date" />
</xsd:simpleType>
</xsd:attribute>
</xsd:complexType>
</xsd:element>

The XML query is sent in as the URL encoded value of the 'qdata' parameter in this request:

http://doi.crossref.org/servlet/query?usr=<username>&pwd=<password>&type=q
&qdata=<x-www-form-urlencoded queries>

Optionally, it may be uploaded for asynchronous processing using an HTTP POST with the XML results being returned in an email.

In response the system will send an XML result formatted to the CrossRef XML query result schema. Shown below is the information returned for a journal citation. In addition to a normal CrossRef query, this result contains the article title and up to 10 contributors. Contributors differ from the typical <author> element in that both first and last names are supplied.

<xsd:element name="journal_cite">
<xsd:complexType>
<xsd:sequence>
<xsd:element ref="issn" minOccurs="0" maxOccurs="2"/>
<xsd:element ref="journal_title" minOccurs="0" maxOccurs="1"/>
<xsd:element ref="journal_abbreviation" minOccurs="0" maxOccurs="1"/>
<xsd:element ref="article_title" minOccurs="0" maxOccurs="1"/>
<xsd:element ref="contributors" minOccurs="0" maxOccurs="1"/>
<xsd:element ref="volume" minOccurs="0" maxOccurs="1"/>
<xsd:element ref="issue" minOccurs="0" maxOccurs="1"/>
<xsd:element ref="first_page" minOccurs="0" maxOccurs="1"/>
<xsd:element ref="year" minOccurs="0" maxOccurs="1"/>
<xsd:element ref="publication_type" />
<xsd:element ref="doi" />
</xsd:sequence>
</xsd:complexType>
</xsd:element>

An alternative to using the XML query format described above will be to send in a much simpler HTTP request as shown below. This format is limited to only one DOI in the query and does not support enabling alerts.

http://doi.crossref.org/servlet/getForwardLinks?usr=<username>&pwd=<password>&doi=<doi>
&startDate=<startDate>&endDate=<endDate>

The response to this request will also be formatted as XML.