CrashPlan is software for the moment after someone says, “I think I deleted it.” It runs quietly on employee computers, servers, and cloud-work suites, making encrypted copies and preserving earlier versions. When a laptop fails, ransomware scrambles a folder, a researcher overwrites a dataset, or an employee leaves with work still trapped on a machine, CrashPlan offers a route backward.
That job sounds simple because the interface between panic and relief is supposed to look simple. Underneath sits scheduling, deduplication, encryption, identity, retention, storage geography, legal hold, audit trails, and the awkward physics of moving terabytes over an internet connection. CrashPlan sells the orchestration as a subscription. The customer buys time - fewer hours spent rebuilding a device, chasing a departed employee, or explaining why the only copy vanished with a hard drive.
A social network crashed. The backup survived.
The product’s origin is a useful warning against falling in love with the first idea. Code42, founded in Minneapolis in 2001 by Matthew Dornquast, Brian Bispala, and Mitch Coopet, began as an IT consultancy. The team used services revenue to fund product experiments. Around 2006, one experiment became a desktop application with Facebook-like ambitions. It grew too large and too impractical. The team kept the online-storage component, and CrashPlan appeared in 2007.
The accident found a market. Consumers wanted an inexpensive, automatic copy of family files. Small companies wanted the same convenience. Enterprises noticed that a tool designed to protect messy personal folders could also protect the unstructured work scattered across employee laptops. Code42 stayed bootstrapped until 2012, when Accel and Split Rock invested $52.5 million. An $85 million round followed in 2015, led by JMI Equity and New Enterprise Associates.
Then came the break. In August 2017, Code42 announced that CrashPlan for Home would shut down in October 2018. The company wanted business and enterprise customers. Existing subscribers received time to move and a referral to Carbonite, but strategy did not soften the betrayal felt by loyal users. What failed first was not the backup engine. It was the fit between a mass-market promise and a company increasingly built around enterprise economics.
The most consequential CrashPlan restore was the company restoring the product itself.From consumer exit to independent enterprise company
The carve-out gave the product a second job
By 2022, Code42’s attention had moved toward insider-risk security. CrashPlan remained valuable but strategically separate. Mill Point Capital acquired the backup business in an undisclosed deal and established CrashPlan Group as an independent company. It was a neat separation: Code42 pursued data-loss prevention; CrashPlan concentrated on having a clean copy after loss occurred. Dan Beer now leads CrashPlan from Minneapolis, with roughly 170 employees according to the supplied company dataset.
Independence changed the ambition. Endpoint backup remained the foundation, but work no longer lived only on endpoints. It lived in Exchange, OneDrive, SharePoint, Teams, Gmail, Google Drive, and servers. CrashPlan acquired Parablu in October 2024, adding Microsoft 365 protection and patented ways to use existing OneDrive for Business or Google Drive capacity as a backup destination. By 2026, its active platform covered endpoints, servers, Microsoft 365, and Google Workspace.
The difference from ordinary sync is the preserved past. Sync tools are excellent at distributing the present, including a bad present. Delete or encrypt a synchronized file and that change may travel. Backup keeps versions outside the live workflow, ideally with immutability and controlled keys. CrashPlan encrypts data with AES-256 at rest and TLS 1.2 or better in transit, offers customer-controlled encryption options, and maintains SOC 2 Type II and ISO/IEC 27001 credentials.
It sells less panic per seat
CrashPlan’s business model is recurring B2B software, licensed per user or device depending on the product. Plans can be monthly, annual, or two-year. Some endpoint packages are self-service; Enterprise, MSP, server, and SaaS-workload deals can require a conversation with sales. Public prices and plan names have shifted, which makes the current support matrix more trustworthy than an old review screenshot.
What did it cost? The honest answer changes with the year and the workload. CrashPlan’s public pricing page has advertised low-cost entry tiers, per-user endpoint subscriptions, and annual enterprise packages, while Microsoft 365 protection adds its own user and storage arithmetic. Procurement teams should price the whole recovery design: the protected people, their allowed devices, the years of versions, cloud capacity, administrator time, and the network needed to bring everything home. A cheap archive that takes a week to restore can be the expensive choice. A higher subscription that turns common mistakes into self-service can pay for itself in help-desk hours.
The invoice is only part of the cost
Conceptual buying model, not audited company economics. The expensive variable is often the time required to recover, not the monthly line item.
That model works best where the saved files are unusually valuable and the users are unusually busy: universities, research organizations, legal teams, healthcare, government, finance, technology companies, and managed-service providers. In a Wake Forest University case study, every maintained computer had to run CrashPlan. The policy removed backup from the realm of faculty good intentions while limiting backup-data access to a tiny administrator group. The university’s verdict was wonderfully untechnical: time saved on panic and device exchange was “priceless.”
CrashPlan says its MSP business serves more than 900,000 users on 1.4 million devices, with renewal above 90 percent and average tenure longer than seven years. Those are company-reported numbers, not audited disclosures, but they explain the private-equity appeal. Backup is sticky when it works. Years of retained versions, deployed agents, tested procedures, and administrator training make replacement a project rather than a swipe of a credit card.
Where the plan can crash
No backup product repeals bandwidth. An initial upload of several terabytes can take days or weeks. A full recovery can become the real bottleneck, especially when the service does not offer shipped-drive restoration. Independent reviewers have found CrashPlan capable of fast small restores, while customers with very large archives have reported slower or fragile experiences. Both can be true. Recovering one deleted presentation and reconstructing a 10-terabyte workstation are different sports.
It is also file-level backup, not universal disaster magic. Buyers who need bare-metal system images, mobile-device backup, signature-based recovery from a damaged disk, or endpoint support for NAS and mapped network drives should check the exact product before committing. Retention differs by plan. Device allowances and legacy subscriptions differ too. “Unlimited” storage does not mean unlimited devices, unlimited bandwidth, or instant recovery.
The conditions that break the model are easy to name and easy to ignore. It will not work if the agent never reaches the files, if exclusions hide the important folder, if the only encryption key disappears with the incident, or if a legal retention policy outlives the purchased retention. It will disappoint when a multi-terabyte recovery has a four-hour objective but only an ordinary internet pipe. And it cannot repair data that was never captured. CrashPlan reduces dependence on users remembering to copy files, but the organization still has to deploy it, monitor failed jobs, and prove that a restored copy opens.
A backup you have never restored is a theory with a monthly bill.The operating rule worth copying
What another company can copy tomorrow
The useful CrashPlan playbook is operational, not proprietary. Make backup automatic. Let users restore ordinary files without waiting for IT. Keep administrator access narrow and logged. Preserve versions long enough to discover slow-moving damage. Store encryption keys where the incident cannot take them too. Separate backup from sync. Most important, rehearse restores at realistic scale.
- 01Restore one file, one folder, one departed employee’s device, and one ransomware-era point in time.
- 02Measure the full recovery in hours, not in marketing adjectives. Include bandwidth and help-desk labor.
- 03Verify retention, exclusions, device limits, storage location, and key ownership in writing.
- 04Keep another recovery path for the data whose downtime would cost more than the backup.
CrashPlan’s own 2026 updates show where the market is heading. Audit logs grew more detailed. Azure Key Vault became an option for external key management. OneDrive and SharePoint gained policy-driven archiving that replaces low-activity files with lightweight links, reclaiming primary storage while leaving a route back. App version 12 added choices for file conflicts during restore. These are not glamorous features. They are features born from administrators asking precise questions after imprecise disasters.
The company’s bet is that backup can expand into a control plane for resilience, governance, and storage economics without becoming another sprawling security suite. That positioning sits between simple consumer-style cloud backup and the heavier platforms sold by Druva, Veeam, Commvault, Rubrik, and Cohesity. Backblaze, Carbonite, IDrive, and Acronis remain alternatives for smaller teams. The right choice depends less on the logo than on workloads, recovery objectives, and whether the worst plausible restore completes before the business runs out of patience.
CrashPlan has already learned what happens when a company changes the plan on its users. Its second act will be judged by a sterner promise: when the customer’s plan fails, does the copy come back? Everything else is storage.