Fix New Outlook startup crash by renaming Olk and OneAuth caches

Background / Overview​

What’s failing and why the Olk + OneAuth reset generally works​

The two folders and what they contain​

• %localappdata%\Microsoft\Olk — Holds transient client artifacts: cache files, local WebView data and renderer artifacts, temporary logs and other non‑persistent startup state that the New Outlook uses to speed loading and enable offline behavior.
• %localappdata%\Microsoft\OneAuth — Stores local authentication artifacts: short‑lived tokens, cached sign‑in state and credential handshakes used to silently authenticate Microsoft 365/Exchange/Outlook.com accounts.

Technical plausibility and WebView2​

The quick fix — step‑by‑step (safe, recommended approach)​

• Close Outlook completely. Confirm there’s no olk.exe or OUTLOOK.EXE process left in Task Manager.
• Press WIN + R to open Run, paste: %localappdata%\Microsoft and press Enter.
• Locate the folders named Olk and OneAuth.
• Preferred — rename each folder to add a suffix (for example, Olk.bak and OneAuth.bak). Renaming preserves a backup copy for quick rollback.
• Start Outlook. Allow a few minutes for the client to recreate caches, reauthenticate and fully sync mail.
• If everything is stable for a session or two, you may delete the .bak folders to reclaim disk space.

$path = "$env:LOCALAPPDATA\Microsoft"
Rename-Item "$path\Olk" "$path\Olk.bak" -ErrorAction SilentlyContinue
Rename-Item "$path\OneAuth" "$path\OneAuth.bak" -ErrorAction SilentlyContinue

Safety notes and immediate caveats​

• Renaming is safer than deletion because it preserves a fallback copy you can restore quickly.
• This action resets local temporary state and cached tokens — it does not delete mail stored on Exchange/Exchange Online or Outlook.com. However, users may be prompted to reauthenticate (including MFA), and some UI settings or locally cached elements may need time to rebuild.
• If you rely on locally stored PST files that are not stored under %localappdata%\Microsoft, confirm where those PST/OST files live before broad operations. Back up any local‑only archives if you’re unsure.

When the quick fix won’t work — follow‑up troubleshooting​

• Launch the New Outlook in Safe Mode to rule out add‑ins (classic Outlook command-line method differs for the New Outlook).
• Repair the app: Settings > Apps > Installed apps > Outlook > Advanced options > Repair, then Reset if Repair fails.
• Reinstall using the offline/msix installer rather than the Store package; some users reported success by reverting or reinstalling a known‑good package.
• Collect Event Viewer Application logs and note the exact exception code, process name (olk.exe) and faulting module (ucrtbase.dll or similar). This information is critical if you must escalate to Microsoft support.
• If a new Windows profile opens Outlook normally, the problem is profile‑specific — keep backups of local-only data and consider migrating the user to a fresh profile.
• Run SFC and DISM to rule out system file corruption:
• sfc /scannow
• DISM /Online /Cleanup-Image /RestoreHealth

Enterprise considerations: how to deploy safely at scale​

• Test the rename approach on a pilot group first; verify sign‑in success and that any conditional‑access or MFA flows complete smoothly.
• Use user‑context automation (Intune Proactive Remediations, ConfigMgr run as user, or a logon script) to rename rather than delete folders. Renaming is reversible and safer for troubleshooting.
• Communicate to end users in advance. Warn them they may be asked to sign in again and that the first launch may take longer while indexes and caches rebuild.
• Keep backups of renamed folders for at least 24–72 hours during the pilot window so you can rollback quickly if the change produces unexpected outcomes.

• Capture Event Viewer errors and affected Outlook build numbers before remediation.
• Pilot the change for a small, representative user set.
• Deploy the rename remediation via user‑context management tools, not machine‑wide deletion scripts.
• Monitor for recurrence over 24–72 hours. If the issue reoccurs across the fleet, escalate to Microsoft with collected logs.

Risks, limitations, and what to expect after the fix​

• Reauthentication friction: Deleting OneAuth forces token reissuance and may trigger MFA or conditional access flows. Users who lack second‑factor devices or self‑service recovery paths may be temporarily locked out. Plan for helpdesk overhead.
• Temporary slow first launch: The app rebuilds caches and search indexes; large mailboxes on slow storage will feel sluggish until indexing completes.
• Not always permanent: For some users the fix is a one‑time remedy; for others the condition recurs until Microsoft ships a hard fix or a patched app is installed. Track recurrence rates during rollout.
• Local‑only PST edge cases: If users have PSTs stored in non‑standard locations under profile folders, verify backups—don’t assume all locally stored mail lives in cloud mailboxes.

Alternatives and interim workarounds​

• Use Outlook on the web (Outlook.com / OWA) from Microsoft Edge or another browser. It’s fully functional for most mailbox tasks and avoids local client state entirely. This is the simplest temporary fallback.
• Revert to the classic Outlook client if compatible with your environment and licensing. In some incidents a switch back to classic Outlook was the only reliable workaround until a patched New Outlook build was released.
• Reinstall the New Outlook using the offline/msix installer instead of the Microsoft Store version; some reports indicate certain store‑delivered updates caused the problem for specific versions and channels.

Microsoft’s response and patch status (what the timeline shows)​

• Community and Microsoft Q&A threads identified the issue and the Olk/OneAuth reset as an accepted remediation within hours of widespread reports.
• In at least one event (August 2025) Microsoft pulled or reverted a problematic rollout and later delivered an updated New Outlook build that resolved the crash in telemetry. Independent reporting and community follow‑ups corroborated a fix being shipped in subsequent app builds.
• However, local remediation remains necessary for many end users while updates propagate; administrators should treat the Olk/OneAuth reset as a practical, immediate mitigation rather than a root‑cause correction.

How to automate the safe rename (example Intune/Proactive Remediation snippet)​

• Detection script (PowerShell) — check existence of folders and whether Outlook is failing to open (optional).
• Remediation script:
  
  Code:

  $path = "$env:LOCALAPPDATA\Microsoft"
  try {
  if (Test-Path "$path\Olk") { Rename-Item "$path\Olk" "$path\Olk.bak" -ErrorAction SilentlyContinue
  }
  if (Test-Path "$path\OneAuth") { Rename-Item "$path\OneAuth" "$path\OneAuth.bak" -ErrorAction SilentlyContinue
  }
  } catch {
  # Log errors to local file for review
  "$($_.Exception.Message)" | Out-File -FilePath "$env:TEMP\OutlookFix.log" -Append
  }

• Deploy to a pilot group first.
• Ensure remediation runs in the user’s context (so you target the right LocalAppData).
• Retain the .bak folders for 24–72 hours and then delete once you confirm success.

Practical recommendations — what readers should do right now​

• End users: Try the rename remedy first (WIN+R → %localappdata%\Microsoft → rename Olk and OneAuth → relaunch Outlook). Back up any local PSTs first if you’re unsure where they live. Expect an MFA prompt.
• IT admins: Pilot the rename via user‑context remediation, collect Event Viewer logs before and after, and prepare communications that explain temporary re‑sign in and potential initial slowness. Prefer rename over deletion for reversibility.
• If the issue persists broadly: collect representative logs and escalate to Microsoft with exact timestamps and client build numbers; check the Microsoft service health dashboard and Q&A for confirmed mitigations and patched build numbers.

Critical analysis — strengths of the workaround and the operational risk it reveals​

• Effectiveness: For a large proportion of affected machines the remedy is quick, reproducible and restores the UI without deleting server‑side mail or rebuilding user mailboxes from scratch.
• Low invasiveness: Renaming local caches is non‑destructive and reversible, making it a safe first‑line remediation before escalating to heavier fixes.
• Operational clarity: The pattern (silent launch failure + 0xc0000409 in Event Viewer + WebView/OneAuth artifacts implicated) maps cleanly to a cache/token corruption model, giving IT teams a clear triage path.

• Fragility of per‑user local state: Modern clients that depend on ephemeral local caches and token blobs are exposed to partial‑write, update timing, or serialization bugs that can cause hard startup failures.
• Upgrade and distribution risk: Store‑delivered updates and rapid rollout windows can deliver a defective build widely before remediation is in place, forcing IT teams into reactive firefighting.
• User impact from reauthentication: Deleting auth caches may trigger MFA flows and conditional access blocks, creating additional helpdesk load and potential downtime for users without recovery options.

Final takeaways​

• If New Outlook refuses to open on a Windows PC, the fastest, lowest‑risk first step is to rename the Olk and OneAuth folders under %localappdata%\Microsoft and relaunch. This has been validated by Microsoft community staff and many independent reports.
• The remedy addresses local cache and token corruption — it does not remove cloud mailbox content, but it will likely prompt reauthentication and initial resynchronization. Plan for MFA and short delays while the client rebuilds indexes.
• For enterprises, automate the rename in user context, pilot broadly, and preserve backups for rollback. If the problem is widespread, gather logs and escalate to Microsoft for a permanent patch; Microsoft’s past pattern has been to revert rollouts and ship a fixed build once the root cause is addressed.
• If you prefer to avoid the New Outlook entirely while issues are sorted, run Outlook in the browser or reinstall the classic Outlook client where your workflows permit.