Reviewed by MSAccessOnline migration team · Last updated May 2026
Splitting a Microsoft Access database into a front-end (forms, reports, code) and a back-end (tables on shared storage) is the standard first step toward multi-user stability. Done well, it reduces corruption tied to shared front-ends and simplifies updates. Done poorly, it only moves the pain to the back-end file. This guide covers correct setup, mistakes we see in reviews, and when to move to SQL or a web application.
What split database means in practice
Before split, one .accdb contains everything. After split, table data lives in a back-end file on a server share or hosted path. Each user opens a front-end that links to those tables. Queries, forms, reports, and VBA stay in the front-end, so you can redeploy UI fixes without copying the full data file to every desktop.
Back-end: tables, relationships, and optionally static lookup data
Connection: linked tables with consistent UNC or managed drive paths
Split does not change the Jet/ACE concurrency model. It organizes files so updates are safer. Heavy write load still hits one back-end .accdb unless you upsize to SQL.
When splitting is worth it
Split early if more than one person uses the database daily, you deploy form changes often, or IT needs a clear backup target (the back-end only). Stay unsplit only for true single-user tools or prototypes with no production SLA.
If users already report lock conflicts, read MS Access multi-user problems in parallel. Split fixes distribution problems; it may not fix contention by itself.
Correct setup steps
Full backup of the unsplit database with all users disconnected
Run Compact and Repair on a copy; resolve obvious corruption before split
Use the Database Splitter wizard or manual move tables to back-end, link in front-end
Place back-end on governed storage with nightly backup and documented UNC path
Distribute front-end copies per machine or via controlled installer
Test linked tables from each network segment (office, VPN, hosted desktop)
Document relink procedure for IT when the share moves
After split, add a startup routine in the front-end that verifies back-end connectivity and shows a clear error if the path is wrong. Silent failures frustrate users and create phantom support tickets.
Common mistakes to avoid
These patterns cause more downtime than staying unsplit:
One front-end file on a shared drive opened by multiple users simultaneously
Hard-coded drive letters (S:) that differ between laptops and Citrix sessions
No version control when a developer ships a new front-end without telling the team
Back-end on a slow VPN share while users save continuously from remote sites
Backups that copy the .accdb while users are connected, producing unrestorable files
Mixing production and test tables in the same back-end without access discipline
Split database deployment patterns
Pattern
Stability
Update effort
Verdict
Local FE per PC, BE on LAN
Good for small teams
Medium
Recommended baseline
Shared FE on network drive
Poor
Low until it breaks
Avoid
FE on Citrix, BE on same farm
Moderate
Low if scripted
Acceptable with governance
FE local, BE over VPN
Poor under write load
Medium
Plan SQL or web
FE linked to SQL BE
Strong
Medium
Best bridge architecture
Browser app, SQL BE
Strongest
Central deploy
Long-term target
Reliability practices that actually help
Publish front-end updates on a schedule; keep a rollback copy
Index foreign keys and fields used in WHERE clauses on busy forms
Shorten record locks: close forms after save, avoid continuous form scroll on huge sets
Run long reports from read-only snapshots or replica exports where possible
Monitor back-end size monthly; plan SQL before the 2 GB wall
Test restore, not just backup file existence
Pair split discipline with corruption awareness. If repair events are recurring, see Access database corruption fix for root-cause patterns that split alone cannot solve.
Unsure whether your split setup is production-safe?
Split is a file-organization strategy, not unlimited scale. Teams usually outgrow it when:
Concurrent writers exceed what the back-end file tolerates without lock storms
Remote users depend on high-latency file shares for every save
Back-end size or attachment volume approaches Access limits
Compliance requires row-level audit, encryption at rest, or granular roles SQL provides
Leadership wants browser access without installing Access on every device
At that stage, upsizing data to SQL Server while keeping Access forms is a proven bridge. The end state for many businesses is a web application on the same database, which removes front-end distribution entirely.
Migration path beyond split
A practical sequence minimizes disruption:
Stabilize split and backups; document peak users and lock-heavy forms
Migrate tables to SQL; relink Access front-end; validate reports
Convert the worst lock offenders to web screens first
Expand web modules by department; retire shared back-end .accdb for daily use
Decommission legacy front-ends after acceptance testing and training
Phased delivery keeps finance and operations running while architecture improves underneath. Pilot one workflow with real users before committing to a big-bang cutover.
What is the difference between Access front-end and back-end?
The back-end holds tables and relationships in an .accdb file on shared storage. The front-end holds forms, reports, queries, macros, and VBA, with linked tables pointing to the back-end. Each user should run a local or properly deployed copy of the front-end, not a single shared front-end on a network drive.
How do I split an Access database using the wizard?
Use Database Splitter on the back-end candidate after a full backup. The wizard creates a back-end file with tables and leaves linked tables in the front-end. Immediately verify linked table paths, relink if the share moves, and distribute updated front-ends to every user.
Why do users still get lock errors after splitting?
Splitting reduces front-end contention but the back-end is still a file-based database. Many concurrent writes, long-running reports, and VPN latency can still produce locks and corruption risk. SQL Server or a web app addresses concurrency at the engine level.
Should the back-end live on a network share or cloud drive?
Use stable, low-latency storage with documented backup and exclusive-access windows for maintenance. Cloud mapped drives and VPN file shares work only when IT enforces disconnect rules, antivirus exclusions, and monitored free space. Unreliable paths cause more damage than local office hosting.
How do we update the front-end for everyone?
Version the front-end file (date or build number in the name), publish through a script or installer, and block old versions from connecting if possible. Email attachments and manual copy-paste lead to schema drift, missing bug fixes, and mysterious errors.
When should we stop investing in split architecture?
When lock complaints persist after front-end discipline, user count grows past roughly ten concurrent writers, remote workers depend on VPN file access, or the back-end file approaches 2 GB. Those signals mean split was the right first step, not the final architecture.
Related guides
Data Safety
Access Database Keeps Corrupting? Causes and Permanent Fix
Learn the real causes of corruption in local Access environments and the safest long-term fix.