Reference Manager Data Migration
By Dr. Zubair Khalid, DVM, MS, PhD ·
Quick Answer
- Export your reference library to a portable format before changing tools, then verify the file opens correctly in the destination program before deleting anything from the source.
- Preserve folder structure by exporting collections or groups separately and confirming that the destination tool supports hierarchical folders, since some tools flatten nested organization.
- Notes, highlights, and file attachments often do not transfer through standard export formats, so plan to manually reattach PDFs and export notes as separate text files.
At a Glance
| Migration Task | Common Export Format | What Transfers Reliably | What Needs Manual Work |
|---|---|---|---|
| Bibliographic data | RIS, BibTeX, XML | Authors, titles, journal names, years, DOIs, volume and issue numbers | Field mapping differences between tools, especially for conference papers and preprints |
| Folder and collection structure | Tool-specific export or XML | Collection names and hierarchy in some tools | Tags and nested folders when the destination tool uses a flat tag system |
| Notes and annotations | RIS or separate note export | Plain text notes in some formats | Highlights, sticky notes, and annotations attached to specific PDF pages |
| Attached PDFs | Linked file export or manual | File paths if the export preserves them | Physical file transfer when the destination tool stores files in its own library folder |
Understanding What a Reference Manager Actually Stores
A reference manager holds more than a list of citations. Each entry contains bibliographic metadata, your personal notes, tags, links to full-text PDFs, and sometimes annotations embedded in those files. When you migrate, you are moving several distinct layers of data, and each layer has different transfer behavior.
The bibliographic record is the most portable layer. Fields such as author names, article titles, journal names, publication years, and digital object identifiers follow standard formats that most tools can read. The RIS format and BibTeX are the two most common interchange standards, and nearly every reference manager can export to at least one of them.
Your organizational structure is a second layer. Folders, collections, and tags represent how you have arranged your library for a project, a thesis chapter, or a grant application. This structure is often tool-specific. One program may call a group a collection, another may call it a folder, and a third may rely entirely on tags. The export format determines whether that structure survives the move.
The third layer is your personal work product. Notes you have typed, highlights you have made in PDFs, and the PDF files themselves are not part of the standard bibliographic record. Some tools bundle these into their own export format, but the generic RIS and BibTeX formats do not carry them. You must plan for this layer separately.
The fourth layer is your identity and workflow integration. Your reference manager may connect to your ORCID record, to a manuscript writing tool, or to a shared lab library. These connections do not transfer with the data. You will need to re-establish them after the migration.
Why Folder Structure and Metadata Are the Hardest Part
The most common complaint after a reference manager migration is not lost citations. It is lost organization. A researcher who has spent three years building a folder system for a dissertation may find that the new tool imported every reference but flattened all folders into a single list.
The cause is a mismatch between how tools store hierarchy. Some tools store folders as a tree, where a folder can sit inside another folder. Other tools store only tags, which are flat labels that can be applied to multiple references. When you export from a tree-based tool into a tag-based tool, the hierarchy is lost unless the destination tool can read the export format and reconstruct the tree.
The RIS format has a field for keywords, and some tools use that field to carry folder names. But the behavior is inconsistent. One tool may export the full folder path as a keyword, another may export only the leaf folder name, and a third may ignore folders entirely. You cannot assume that folder names will survive an RIS export.
The same problem applies to tags. If you have used tags to mark items as read, unread, or assigned to a specific chapter, those tags may or may not appear in the export. Some tools export tags as keywords, and others do not. The only reliable way to know is to export a small test library first and inspect the resulting file.
Pre-Migration Checklist
Before you install new software or delete anything from your current tool, complete a structured checklist. This checklist protects you from the most common data loss events.
Inventory Your Current Library
Open your current reference manager and record the total number of references, the number of folders or collections, the number of tags, and the number of attached files. Write these numbers down. You will use them after the migration to verify that nothing was lost.
Check for duplicate references. Many libraries accumulate duplicates over years of importing from different databases. A migration is a good time to identify them, but do not delete them before you have a backup. If you delete a duplicate and later discover that one copy had a note the other did not, you have lost data.
Identify the Export Formats Your Destination Tool Accepts
Look at the documentation for the tool you are moving to. Find the list of import formats it accepts. Most tools accept RIS and BibTeX, but some also accept XML exports from specific tools. If your destination tool accepts a direct import from your current tool, that is usually the safest path because the direct import format is designed to preserve more data than a generic interchange format.
Test with a Small Sample
Create a small test library in your current tool. Include one reference with a folder assignment, one with a tag, one with a note, and one with an attached PDF. Export that test library and import it into the destination tool. Check whether the folder, tag, note, and PDF survived. This test takes ten minutes and tells you exactly what your full migration will preserve.
Decide What You Are Willing to Rebuild
If the test shows that notes do not transfer, you must decide whether to manually copy notes or accept the loss. If the test shows that folders flatten into tags, you must decide whether the tag structure is acceptable or whether you need to reorganize after migration. This decision is easier to make on a small test than on a full library.
Step-by-Step Migration Workflow
The migration workflow has five stages. Each stage has a specific output that you verify before moving to the next.
Stage One: Clean the Source Library
Remove obvious duplicates only if you are certain the duplicate has no unique notes. Merge duplicates if your tool supports merging, because merging preserves notes from both records. Check for references with missing titles or missing years. These will still export, but they may be harder to find in the destination tool.
Stage Two: Export the Bibliographic Data
Export your full library in the format your destination tool accepts. If you have a choice between RIS and BibTeX, choose the format that your destination tool documents as its preferred import format. Some tools handle RIS more reliably, and others handle BibTeX more reliably.
Stage Three: Export the Organizational Structure
If your tool supports exporting folders or collections separately, do that. Some tools allow you to export each folder as a separate RIS file, which preserves the folder name as the file name. You can then import each file into a matching folder in the destination tool. This is a manual process, but it preserves your structure exactly.
Stage Four: Export Notes and Annotations
Check whether your tool exports notes as part of the RIS or BibTeX file. If it does, your notes will appear in the destination tool. If it does not, export your notes as a separate text file. Some tools allow you to export a formatted bibliography with notes, and you can use that as a reference document while you manually re-enter notes.
Stage Five: Import and Verify
Import the exported file into the destination tool. Then compare the reference count in the destination tool to the count you recorded in the inventory. Check a sample of references to confirm that titles, authors, and DOIs are correct. Verify that folders and tags exist. Open a few attached PDFs to confirm they are not corrupted.
Options and Tradeoffs
Direct Tool-to-Tool Import
Some reference managers offer a direct import from another specific tool. This is the easiest path because the import is designed to map fields from the source tool to the destination tool. The tradeoff is that direct imports are only available for the most popular tools, and they may still lose notes or annotations.
RIS and BibTeX Interchange
The RIS and BibTeX formats are the universal interchange formats. They are reliable for bibliographic metadata, but they are not reliable for notes, tags, or folder structure. The tradeoff is that you get a clean bibliographic transfer but must manually rebuild organization.
XML Export
Some tools export a full XML file that contains the entire library, including notes and tags. This is the most complete export format, but it is only useful if the destination tool can import that specific XML schema. If the destination tool cannot read the XML, you have a complete backup file but no way to use it.
Manual Rebuilding
The most labor-intensive option is to export the bibliographic data and then manually rebuild folders, tags, and notes in the destination tool. This gives you complete control over the final structure, but it can take hours or days for a large library.
Records and Measurements
Keep a migration log. Record the date, the source tool and version, the destination tool and version, the export format, and the reference count before and after migration. Record any references that failed to import and any folders that were lost. This log is your evidence that the migration was complete.
After the migration, run a verification check. Compare the reference count, the folder count, and the tag count between the source and destination. If the destination tool has a duplicate detection feature, run it to catch references that imported twice.
Common Failure Patterns
Folders Flattened into a Single List
This is the most common failure. The destination tool imports all references but ignores the folder structure. The cause is usually an export format that does not carry folder information. The fix is to export each folder separately and import each file into a matching folder.
Tags Lost or Merged
Some tools export tags, and some do not. When tags do transfer, they may be merged with other fields. For example, a tool may export folder names and tags into the same keyword field, which creates a confusing mix in the destination tool.
Notes Imported as Separate References
Some tools export notes as separate records, which creates duplicate entries in the destination library. If you see references with titles like "Note on Smith 2020," you have this problem. You must delete these duplicates and reattach the notes manually.
PDF Links Broken
The destination tool may import the reference but not the PDF file. The reference will show a link to a file that does not exist in the new library. You must reattach the PDF manually or use a file transfer tool to copy the PDF files into the new library folder.
Duplicate References After Import
If you import a file that already contains duplicates, or if you import the same file twice, you will have duplicate references. Use the destination tool's duplicate detection feature to merge them.
Safety and Data Protection Context
Your reference library is a research record. If you are working on a grant-funded project, your library may be part of the data that you are expected to manage and share. The NIH Data Management and Sharing Policy describes expectations for managing and sharing scientific data, and a reference library is part of the research data ecosystem. A migration that loses notes or organization can compromise your ability to document your research process.
The Committee on Publication Ethics Core Practices describe expectations for data and reproducibility. If you are preparing a manuscript, your reference library is part of your research record. Losing notes about why you cited a particular source can make it harder to respond to peer review or to verify your citations.
Your ORCID record is a separate layer of your research identity. A reference manager migration does not affect your ORCID record, but you may need to reconnect your reference manager to your ORCID account after the migration. The ORCID for Researchers page describes how researchers use ORCID to maintain their publication record.
Professional Escalation Criteria
If you are migrating a library that supports a funded research project, and the migration fails to preserve notes or folder structure, you should stop and consult your institutional research office. The NIH Grants and Funding page describes the expectations for grant management, and your institution may have specific requirements for research records.
If you are migrating a library that supports a manuscript under review, and you lose the ability to verify a citation, you should consult the journal's reporting guidelines. The EQUATOR Network provides a searchable database of reporting guidelines, and some guidelines require specific citation practices.
If you are migrating a library that contains confidential or sensitive data, such as patient records or proprietary information, you should consult your institutional data security office before exporting any files. The export file may contain metadata that you do not want to share.
Decision Framework for Choosing a Migration Path
Selecting the right migration method depends on three factors you can assess before you export anything: the size of your library, the complexity of your organizational structure, and the number of notes or annotations you need to preserve. A simple decision framework helps you match your situation to the appropriate workflow instead of defaulting to a single approach that may not fit your needs.
Factor One: Library Size
Count your total references and attached files. Libraries under 500 references with fewer than 50 attached PDFs can tolerate manual rebuilding because the time cost is manageable. Libraries above 2,000 references with hundreds of PDFs require a method that preserves file paths or copies files automatically, because manual reattachment becomes impractical.
Use this threshold to decide whether you can afford a manual rebuild or must pursue a direct import or XML export path. A small library gives you the freedom to accept some data loss and rebuild structure by hand. A large library demands that you preserve as much as possible through automated means.
Factor Two: Organizational Complexity
Count your folders and note whether you use nested hierarchies. A flat structure with fewer than 20 folders and no nesting can be rebuilt quickly in the destination tool. A nested structure with folders inside folders, or a structure where the same reference appears in multiple folders, requires a method that preserves hierarchy.
If your destination tool uses tags instead of folders, you must decide whether a flat tag system is acceptable. Tags can replicate a folder structure if you name them consistently, such as using a prefix for the top-level folder. This conversion is a deliberate design decision, not an automatic process.
Factor Three: Notes and Annotation Volume
Count how many references have notes and how many PDFs contain highlights or sticky notes. If fewer than 10 percent of your references have notes, you can manually copy them. If more than half of your references have notes, you need an export format that carries notes or a separate note export that you can import.
Annotations attached to specific PDF pages are the most fragile data layer. No standard interchange format carries them. If you have extensive annotations, you must either keep the source tool installed for reference or accept that you will lose the annotations and keep only the PDFs.
Applying the Framework
Use the table below to match your library profile to a recommended workflow.
| Library Profile | Recommended Workflow | Reason |
|---|---|---|
| Under 500 references, flat folders, few notes | Manual rebuild | Time cost is low and you control the final structure |
| Under 500 references, nested folders, many notes | Export each folder separately plus export notes as text | Preserves hierarchy and gives you a note reference document |
| 500 to 2,000 references, moderate structure | Direct tool-to-tool import if available, otherwise RIS export with folder-by-folder export | Balances automation with structure preservation |
| Over 2,000 references, complex structure | XML export if the destination tool accepts it, otherwise staged folder-by-folder export | Large libraries need a complete backup file before any manual work |
| Any size, heavy PDF annotations | Keep the source tool installed and use it as a reference archive | Annotations cannot transfer, so you need the original tool to look them up |
Record System for Migration Decisions
Create a decision record before you export anything. This record is separate from the migration log because it captures your reasoning, beyond the outcome. Write down the library size, the folder count, the tag count, the note count, and the annotation count. Then write down which workflow you selected and why.
This decision record serves two purposes. First, it forces you to assess your library before you act. Second, it gives you a baseline to evaluate whether the chosen workflow was appropriate. If the migration takes far longer than expected or loses more data than you anticipated, the decision record tells you which assumption was wrong.
Troubleshooting Method for Failed Imports
When an import fails, the first step is to isolate the problem to one of three layers: the export file, the import process, or the destination library. You can isolate the layer by running a controlled test.
Create a test file with five references in the source tool. Include one reference with a folder assignment, one with a tag, one with a note, one with an attached PDF, and one plain reference with no extras. Export this test file and import it into the destination tool. Check each of the five references individually.
If the plain reference imports correctly but the others do not, the problem is in the export format. The export file does not carry the folder, tag, note, or PDF information. You need a different export format or a separate export for each data layer.
If none of the references import correctly, the problem is in the file itself. The export may be corrupted, or the destination tool may not recognize the format. Try exporting again and check the file size. A file that is unexpectedly small may be missing data.
If all five references import correctly, the problem is in the full library. The full export may be too large, or it may contain a reference with a character that breaks the import. Run the import again and watch for an error message that names a specific reference. If you find one, remove that reference from the export and import the rest.
Common Failure Pattern in the Decision Process
The most common failure in the decision process is choosing a workflow based on the destination tool's advertised features instead of on your library's actual data. A tool may advertise direct import from your current tool, but the direct import may not preserve notes. The test library method catches this problem before you commit to the full migration.
Another common failure is assuming that a large library needs a fully automated workflow. Automation is convenient, but it does not guarantee completeness. A large library with a simple folder structure may migrate cleanly with a folder-by-folder export, even though that method is manual. The decision framework helps you match the method to the data, not to the tool's marketing.
Professional Escalation for Decision Failures
If your decision record shows that you selected a workflow based on incorrect assumptions about your library, stop the migration and reassess. If the library supports a funded project, consult your institutional research office before proceeding. The NIH Grants and Funding page describes expectations for grant management, and your institution may have specific requirements for research records.
If the library supports a manuscript under review, and the migration has compromised your ability to verify citations, consult the journal's reporting guidelines through the EQUATOR Network. Some guidelines require specific citation practices, and you may need to document how you preserved your reference data.
If the library contains confidential data, such as patient records or proprietary information, consult your institutional data security office before exporting any files. The export file may contain metadata that you do not want to share. The NIH Data Management and Sharing Policy describes expectations for managing and sharing scientific data, and your institution may have additional requirements.
Frequently Asked Questions
Will my PDF files transfer when I move to a new reference manager?
PDF files are not part of the RIS or BibTeX export format. Some tools offer a direct import that also copies PDF files, but you should test this with a small sample before you rely on it. If the direct import does not copy PDFs, you must manually copy the PDF files into the new library folder.
Can I keep my folder structure when I switch tools?
Folder structure survives only if the destination tool can read the folder information in the export format. Some tools export folder names as keywords, and some do not. The most reliable way to preserve folders is to export each folder separately and import each file into a matching folder in the destination tool.
What is the difference between RIS and BibTeX for migration?
RIS and BibTeX are both text formats that store bibliographic metadata. RIS is more common in the biomedical and social sciences, and BibTeX is more common in mathematics and computer science. Both formats store the same core fields, but the field names differ. Choose the format that your destination tool supports best.
Will my notes and annotations transfer?
Notes and annotations are not part of the standard RIS or BibTeX format. Some tools include notes in their export, and some do not. Annotations that are attached to specific pages in a PDF are almost never transferred. You must export notes separately and reattach annotations manually.
How do I know if the migration worked?
Keep a record of the reference count, folder count, and tag count before you start. After the migration, compare the counts in the destination tool. Open a sample of references to confirm that titles, authors, and DOIs are correct. Open a few PDFs to confirm they are attached.
Should I delete my old library after migration?
Keep your old library until you have verified that the new library is complete. The old library is your backup. If you discover a problem in the new library, you can return to the old library and re-export the affected references.
Can I migrate from one tool to another without losing anything?
A complete migration is possible, but it requires manual work. The bibliographic data will transfer cleanly. The folder structure, tags, notes, and PDFs require separate handling. You must test the export format, export each folder separately, and manually reattach notes and PDFs.
What should I do if the migration fails?
If the migration fails, stop and return to your old library. Do not delete the old library. Check the export file to see if it is complete. Try a different export format or a direct import. If the problem persists, contact the support team for the destination tool.
Related Articles
- Data Management Platform: What It Is and How to Choose One
- Data Management Software: A Guide to Choosing the Right Tools
- Statistical Tests: Choosing the Right One for Your Data
- Git for Research Projects: Version Control for Code, Data, and Analysis Notes
- Conda Environments for Bioinformatics: Managing Tools Without Version Drift
References and Further Reading
- Research Methods Resources. National Library of Medicine.
- EQUATOR Network. EQUATOR Network.
- Core Practices. Committee on Publication Ethics.
- NIH Grants and Funding. National Institutes of Health.
- ORCID for Researchers. ORCID.
- Data Management and Sharing Policy. National Institutes of Health.
- NCBI Literature Resources. National Center for Biotechnology Information.
- PubMed. National Library of Medicine.
This article is educational and does not replace institutional policy, professional advice, or applicable safety and regulatory requirements.