Rack of backup drives in a dealership IT closet with a technician checking the equipment
Resources / Blog

Testing Backups: Why a Restore Test Matters More Than the Backup Report

By Rex MarchantNovember 18, 2026 · 9 min read

Backup restore testing is how a dealership finds out its data actually comes back. A report that says the backup completed shows the job ran. Only a restore shows whether a file, a mailbox or a whole machine can be recovered, and how long that takes.

Last updated: November 18, 2026

On a November 10 content call, Brandon Dorsey, our Vice President of Technology Strategy, said it in one sentence: “A backup that says it’s completed is only a checkbox until you actually validate it that there’s restorable data there.” He went on to list what a backup plan has to cover: what’s backed up, the plan itself, how long the data is kept and how often it’s copied. Then the last part. “And need to be tested.” Our backup and disaster recovery for dealerships page covers the plan. This post covers the test.

I’m the Vice President of Service at Helion Technologies, where my teams run the service desk, field service, systems engineering and proactive managed services, and I’ve led managed services teams for more than 25 years. On a November 4 planning call I put it this way: “You can’t set it and forget it. You have to check it.” Backups are the plainest case of that I know.

IT technician checking a rack of backup storage drives in a dealership network closet

What Does a Backup Report Actually Tell You?

A “completed” status tells you the backup job finished. It doesn’t show that the copy opens, that nothing important was left out of the job, or how long a restore would take. A restore answers all three.

I don’t want to knock the report. A failed job is worth knowing about. The trouble is that a success can look the same either way. As Brandon wrote in our post on proactive IT support at the dealership, a backup job “can also report success every night while copying the wrong data, such as a folder nobody uses anymore after a server change.” The report is a record of activity. A restore is a record of results. Busy versus effective.

Dealership office manager and IT lead reviewing a printed report together at a back office desk

CISA’s #StopRansomware Guide tells organizations to “regularly test the availability and integrity of backups in a disaster recovery scenario,” and says it again more plainly: “Test backup procedures on a regular basis.” The same guide says to keep backups offline, because many ransomware variants “attempt to find and subsequently delete or encrypt accessible backups.” Testing and an offline copy go together in that guidance, and I’d treat them that way.

A completed job is a claim. A restored file is proof.

What Should a Restore Test Actually Restore?

A restore test should pull real data back out of the backup and open it. A practical way to plan the tests is by scope, from one file to one mailbox to one whole machine, because each one answers a different question about the backup.

ScopeWhat you restoreThe question it answers
A fileA document from a shared folder, restored to a different location, then openedWas it in the job, and does the copy open?
A mailboxOne user’s mailbox, rolled back to a point before a deletionDoes email in the cloud have a copy of its own you can restore?
A whole machineA server or workstation, restored onto spare hardware or a test machine, then startedDoes it come back working, and how long does that take?

At Helion, test restores run on a schedule, with the results written down where the dealership’s managers can see them. We time them too, so the restore time comes from a test and not from the real thing.

How often is a fair question, and the published guidance doesn’t agree on one number. CISA’s guide says to test regularly. The CIS Critical Security Controls are more specific. Safeguard 11.5, which CIS assigns to its Implementation Groups 2 and 3, reads “Test backup recovery quarterly, or more frequently, for a sampling of in-scope enterprise assets.”

Notice the word “sampling.” Not everything, every time. One way to handle it is to rotate what you pick, so every important system gets a turn over the year. Our backup page adds one more trigger: test again after any big change, like a server moving to Azure or a new store coming onto the network. A schedule on paper beats good intentions, so I’d ask any provider, including us, to show theirs, along with the dates of the last few tests and what each one restored.

Technician at a workbench with a spare desktop computer and external drive ready for a test restore

Always open it. A restore you didn’t open is a restore you didn’t test.

What Does a Failed Restore Test Teach You?

A failed restore test shows which part of the backup plan doesn’t hold up, while there’s still time to fix it. Brandon’s list from that call is a good place to start looking: what’s backed up, the plan, the retention and the frequency.

I can’t tell you which of these a test will turn up at your stores. Here’s where to look.

  • What’s in the job. A job can copy the wrong data and still report success.
  • Retention. If copies aren’t kept long enough, the version from before the problem may already be gone.
  • What was ever switched on. Our backup page cites Microsoft: for servers in Azure, a backup only runs for the machines somebody selected and gave a schedule.
  • Where the copy lives. CIS Safeguard 11.4 calls for an isolated copy, offline, in the cloud or off-site.
  • Is the copy protected as well as the original? CIS Safeguard 11.3 says it should be.
  • The clock. How long the restore took, measured, not guessed.

That last one gets its own paragraph. Brandon covered time on the same call, using two terms worth knowing, recovery time objective and recovery point objective: “how much time do you want it to take if you had to restore everything? That’s your recovery time objective.” The range he described runs from immediate to “it’s okay if it’s a day or two,” and the second decision is how far back you restore from, “the last 10 minutes or the last day.” In his words, “all those come with a cost.”

I won’t promise a restore time here. It depends on your data and your systems. A timed test finds it.

Are we fixing the cause, or are we working around the symptom? Running the job again until it reports success is the second one. It fixes nothing. If a test fails, write down why before anyone touches the schedule.

Fix the cause first.

Who Should Be Checking, and What Should You Ask to See?

Someone specific should own restore testing, and they should be able to show you the date of the last test, what was restored and what happened. If nobody can name the owner, settle that first.

Our team’s view, from a November 17 content call, is that internal IT teams often don’t check that backups are actually running. I wouldn’t read that as a criticism, because dealerships change constantly, with renovations, new stores and new systems, and things get missed when an internal IT team is firefighting the problems in front of it.

That’s where co-managed IT fits. On the co-managed path your IT manager stays in charge, and Helion’s team adds the monitoring, the offline copy and the restore testing alongside them. The wider plan, with owners and fallbacks, sits in our business continuity plan template for dealerships.

Dealer principal and IT manager meeting at a conference table with a closed binder between them

Whoever owns it, here’s what I’d ask them to show you.

  • The date of the last restore test, and the one before it.
  • What was restored: a file, a mailbox, a machine, or all three.
  • How long the restore took, written down.
  • What failed, and what changed because of it.
  • How long copies are kept, and how far back you can actually go.
  • Where the offline or isolated copy lives.
  • Who owns the test on their side, and who on yours sees the result?

Helion supports 2,000-plus dealerships, and these questions apply at any size.

If the answer takes a week to find, that tells you something. Ask anyway.

Things Dealers Ask About Testing Backups

The report says completed every night. Why test at all?
Because “completed” means the job ran. Nothing else. A test is what shows the data was in the job, opens and can be restored. A clean report with no recent restore is an assumption.
Will a restore test disrupt the store?
It doesn’t have to. A test can restore to a separate location or a spare machine instead of over live data. Ask your provider where the restore goes, when it runs and who approves it.
Is there a standard for how often to test?
CIS Safeguard 11.5 says quarterly or more often, for a sampling of systems, in its Implementation Groups 2 and 3. CISA’s guide says to test regularly without naming a number. Test again after any big change too.
Our data is in the cloud. Isn’t that backed up already?
Not automatically. On a November 4 planning call, Erik Nachbahr, Helion’s founder and president, said: “Just because your data is in the cloud doesn’t mean it’s backed up.” Our backup page walks through what Microsoft keeps and for how long.
Nobody has ever tested ours. Where do we begin?
With one file. Restore it to a different folder and open it, then schedule the mailbox and the whole machine next, and write down what you learn from each one so the next test starts from a real list instead of a guess.

Where to Start

Ask for the date of the last restore test. Start there. If the answer comes back fast, with a name attached, you’re in good shape. If it doesn’t, our complimentary IT and cybersecurity assessment is a no-obligation review of your current technology, security posture and vendor spend, along with your FTC Safeguards compliance standing and how staff use AI tools, and the findings come to you in a written report. Our managed IT services for dealerships include the proactive team that tests and restores backups.

A backup you’ve restored from is a backup you can count on.

About the author

Rex Marchant

Vice President of Service, Helion Technologies

Rex Marchant is Vice President of Service at Helion Technologies, where he leads the service desk, field service, systems engineering, proactive managed services and VoIP. He has more than 25 years leading managed services teams and holds ITIL Foundation and COBIT 5 certifications.

Rex Marchant on LinkedIn

Confidence in your current setup and actual protection aren't always the same thing. The only way to know which one you have is to look.

Get Your Complimentary IT & Cybersecurity Assessment →