Every EMR vendor pitch includes a migration timeline slide. It usually says something like "8-12 weeks from kickoff to go-live." In practice, that number is either optimistic by half or hiding a lot of work you'll end up doing after cutover. I've run three migrations and consulted on a dozen more, and the pattern is remarkably consistent: the technical migration is the easy part. Keeping the clinic running while you do it is where practices lose money, staff, and occasionally patients.

Here's what an honest timeline looks like, what actually consumes those weeks, and how to structure the cutover so your front desk isn't rebooking a full week of patients because your prescription queue went dark.

The real timeline for a mid-sized practice

For a practice with 2-8 providers, roughly 3,000-15,000 active patient records, and standard integrations (labs, e-prescribing, payments, a pharmacy or two), plan for 4-6 months from signed contract to a stable steady state. Not to go-live. To the point where your team stops saying "in the old system" every ten minutes.

Break it down like this:

  • Weeks 1-3: Discovery and data audit. You export samples from your current EMR, the new vendor maps fields, and you find out what's missing. This is where you learn that your custom intake form fields don't have a home, or that your old system stored allergies as free text.
  • Weeks 4-8: Configuration and first data load. Templates, user roles, schedule types, fee schedules, document categories, and a test migration of 10-20% of records into a sandbox. Your clinical lead should be reviewing charts in the sandbox weekly, not just at the end.
  • Weeks 9-12: Parallel testing and staff training. Full data load into staging. Front desk shadows the new workflow. Providers chart three or four real visits in the sandbox using yesterday's patients as material.
  • Weeks 13-14: Cutover window. The actual switch. More on this below.
  • Weeks 15-24: Stabilization. Fixing what broke, closing gaps, retraining on the parts everyone got wrong the first time, cleaning up duplicate records, reconciling anything that didn't migrate cleanly.

If a vendor tells you they can compress this to 6 weeks, ask them what they're skipping. Usually it's the audit, the parallel testing, or the stabilization phase. All three are where quality lives.

What actually eats the time

The hours don't go where the Gantt chart says they go. Data mapping usually finishes on time. What blows up the schedule:

Custom fields and free-text notes. If your old system let clinicians write in free text for anything structured (allergies, problem lists, medications, custom intake), someone has to decide how that translates. That someone is a clinician, and they don't have spare hours. Budget provider time explicitly, not hopefully.

Attachments and scanned documents. A practice with ten years of history often has hundreds of thousands of PDFs, faxes, and scanned IDs. Migrating these takes storage, indexing, and often manual re-tagging. Decide early whether you're moving everything or archiving pre-cutoff documents in read-only form. Both are valid; indecision is not.

Integrations. Lab interfaces, e-prescribing accounts, payment processors, and pharmacy connections each have their own setup timelines that don't care about yours. Surescripts registration for a new EMR typically runs 2-4 weeks. Lab interfaces can run 6-8. Start these on day one of the project, not after configuration.

Compounding and concierge specifics. If you're on the compounding side, formulary mapping, batch numbers, and pharmacy routing rules are their own project inside the project. Concierge practices tend to have complex membership billing and communication cadences that don't fit off-the-shelf templates. Both add weeks that generic timelines ignore.

Designing the cutover so the clinic keeps running

The cutover itself is 3-7 days of elevated risk. Your goal is to shrink the window where you're operating on faith and to make sure no patient falls through a gap. A few patterns that work:

Pick a low-volume window, then reduce it further

Most practices cut over on a long weekend. That's fine, but also block the surrounding schedule. Reduce Monday and Tuesday visits by 40-50%. Push routine refills forward the week before so the pharmacy queue is light. Front-load labs so results aren't landing in two systems at once. Your revenue for that week takes a hit; your sanity takes less of one.

Freeze the old system on a defined date

Set a hard read-only date for the old EMR. Everything after that date lives in the new system. If you try to run both live for a week, you will have chart divergence and staff will pick whichever system is faster to open, which is always the old one.

Keep read-only access to the old system for 12-24 months

You don't need to migrate every historical note if you keep the old system accessible for lookup. Negotiate this in your exit terms with the outgoing vendor before you sign anything with the new one. Contract leverage evaporates the moment you announce the switch.

Staff the first two weeks with a war room

Someone senior needs to be reachable in real time for the first ten business days. Not "check Slack when you can." Actually available. In our last migration we had a rotating lead on-site by the front desk for two weeks, and a clinical lead available to providers by phone during every session. The issues that came up were small individually and catastrophic in aggregate if left unaddressed.

Prescription and refill continuity

This is where clinics get hurt fastest. A missed refill window creates a patient call, a same-day appointment, or worse, a lapse in care that lands in a review. Before cutover, generate a refill risk report from the old system: any patient due for a refill within 30 days of the switch date. Push those refills through in the week before cutover if clinically appropriate, or flag them for proactive outreach immediately after. In compounding practices this matters even more because your pharmacy relationships have their own queues.

What to measure during and after

Numbers to watch weekly for the first two months post-cutover:

  • Time from intake form submission to chart-ready (should return to baseline by week 4)
  • Refill turnaround time (should not exceed baseline by more than 24 hours in week 1, back to baseline by week 3)
  • Patient message response time
  • Provider charting time per visit (expect a 30-50% increase for the first two weeks, then decay)
  • Claims lag: days from encounter to claim submitted
  • Denial rate on new claims (watch for coding template issues)

If any of these are still off baseline at week 8, you have a workflow or configuration problem, not a training problem. Treat it as such.

A note on AI-assisted migrations

Automated data mapping and AI-assisted chart summarization have genuinely changed the mid-migration phase. Structured field mapping that used to take 40 hours of analyst time can now happen in a few, with a clinician reviewing exceptions. Summarizing long free-text histories into structured problem lists is another place the tooling has matured. If your new vendor doesn't offer this, ask why. If you want to see how we approach migration and the agents we use for intake, refills, and record cleanup during the transition, talk to our team.

Frequently Asked Questions

Can we migrate without any downtime?

Effectively yes, if you cut over during a scheduled closure and freeze the old system as read-only. What you can't avoid is the productivity dip in the first two weeks after go-live. Plan for it in your schedule rather than pretending it won't happen.

Should we migrate all historical records or start fresh?

Most practices migrate structured data (demographics, active problems, medications, allergies, recent labs) and keep the old system in read-only mode for older documents. Full document migration is expensive and often unnecessary if lookup access is preserved. Confirm any record retention requirements with your own compliance counsel before finalizing what you archive versus migrate.

How much revenue should we expect to lose during cutover?

In our experience, a well-planned cutover costs 10-15% of a normal week's revenue across the cutover week and the following week combined. A poorly planned one can cost 30-40% and stretch over a month. The difference is almost entirely in preparation, not in the software.

When should we tell patients?

Communicate the change 3-4 weeks in advance if it affects their portal login, payment methods, or how they request refills. Keep the message short and practical: what's changing for them, what stays the same, and who to contact if something breaks. Don't apologize for improving your systems; do acknowledge that anything new has a learning curve.

What's the single most common mistake?

Underestimating provider training time and overestimating how much staff will "figure it out." Providers need protected hours in the sandbox before go-live, not a lunch-and-learn the Friday before. Staff need scripted workflows for the top ten tasks they perform daily. Skipping either turns week one into chaos.