Backing Up Your Work

July 7, 2026 18 minute read

There are few things as devastating to an author as to lose the novel you’ve poured hours of blood, sweat, and tears into. Hard drives crash, computers break, cloud services fall…there are a lot of things that can go wrong with digital data. This is, ultimately, why it’s important to have a backup strategy.¹ Of course, we rarely think about this in the moment. It’s easier not to worry, to just keep continuing forward—but then we regret it when disaster strikes. But the sooner you get a backup strategy sorted out, the sooner you can have peace of mind, knowing that even if your laptop decides to take a dip in the local pond, you still have your words.

To that end, in this post, I will be walking through some guidelines on establishing an effective backup strategy, give some suggested options, and then share my current setup and talk a little about it. Hopefully, this ends up being a helpful little jaunt for you, and a springboard to form your own backup strategy.

3-2-1 Rule

It can be a little overwhelming when approaching the prospect of backing things up, but thankfully, decades ago, people settled on a useful mnemonic known as the 3-2-1 Rule for backups. It is, in short:

Three copies of the data
Across two separate types of media
One of them in a different physical location (that is, offsite)

Though as can often be the case with mnemonics, if you’re not as familiar with what it’s referring to, it can be confusing. So let’s step through what each these mean.

The first rule is to have three copies of the data. Three copies means exactly that, though in practice one of them will be the actual file that you work on and edit; the other two are the backups of that primary file.² So after writing in a document, for instance, you would copy it to two additional separate locations. These three locations can have a priority order: I have this primary file, but if I lose it, I first look in place one, and if it’s not there, place two. The core of this is that making one just one backup copy is insufficient. Ultimately, this rule mitigates against things like “I accidentally deleted the wrong file”.

The idea of two different kinds of media made a bit more sense in the era of tape drives and floppy disks; it’s a kinda weird phrasing these days. But, it does speak to the fact that different forms of storage and different devices have different lifetimes and different expectations for use, and you want to mitigate that. In practice what this means is to make sure at least one copy of your data is in a different place than the others—these days, these locations would be things like internal hard drives, external hard drives, and cloud storage; I go into this in a little more detail later. Ultimately, this rule mitigates against “my hard drive crashed and I lost all of the files on it”.

These days, an offsite copy generally means cloud—historically, it often meant a copy on a tape drive that was literally physically stored in a different city. It is worth noting that not all cloud storage is created equally, and you want to think about what sort of encryption you want. Alternatively, there are ways of putting NASes in other physical locations and running a backup over the Internet. Ultimately, this rule mitigates against “my house burned down with all my electronics in it”.

Note that these are all minimums—the more copies you have across more media across more physical locations, the better protection you have against losing your data.

But if you care about your data, and want to make sure you don’t lose it, you should be able to identify these things: how many copies you have, what media they’re on, and what physical locations they’re in.

Versioning

Some backup systems such as Time Machine also back up versions of your files, so you can recover if you accidentally delete a file and it takes you a while to realize it, or if you need an earlier version of something. I don’t do this particularly much, but it has saved my bacon on occasion.³ In general, I think it’s reasonable that one of your backups is able to do this; in practice, you probably don’t need all the version backups replicated, but your mileage may vary. Another way to get around this entirely is using a Git repo, but I wouldn’t recommend that unless you have a compelling reason.

Restoring

Obviously, no backup is worthwhile unless you can restore files from it. I highly recommend regularly making sure that you can access/download/etc. your backed-up files, and making sure they work. I personally do this with spot checks every month or three, under the reasoning that if I can get these random files, the rest should be fine. Some people would recommend attempting a full restore to verify the integrity of everything. You can find the balance that seems appropriate for your tolerance of risk and your level of executive function. But just setting it up and assuming it works without ever testing it is a recipe for disaster.

Software

At this point, you’re probably going “that’s so much work, constantly copying files around!”, and you’d be right! Thankfully, there are a number of programs that have been written now that automatically back files up to a location you specify. They do often take a bit of up-front setup (what files, stored where, how often) but after that, you can just let it run on its own and barely think about it again. However, I personally like still getting notifications whenever a backup succeeds, just for that peace of mind of “this is still working”; there are few things worse than thinking you’ve been backing stuff up with no errors, and then when you need to restore you learn that it just hasn’t been running at all. Your tolerance for this may vary.

In alphabetical order, here are some pieces of software that can set up automatic backup, with varying levels of configuration and complexity:

Arq (Windows, macOS)
Carbon Copy Cloner (macOS)
SuperDuper (macOS)
Time Machine (macOS built-in)
Windows Backup (Windows built-in)

Media

As mentioned earlier, it’s good to have your backup copies on different forms of media. I’m going to walk through several different forms of media, and try to give several examples of each (in alphabetical order, for fairness). If there are things I’ve missed here (or in the software section), please let me know, and I’ll try to get this updated.

Hard Drives

Hard drives are relatively straightforward. Your computer already has one in it—maybe even two. You can also buy external hard drives pretty easily—and I’m including thumb drives in this category, since they act like hard drives.

If you have two internal hard drives, using a different physical hard drive as a backup location counts as a separate form of media—however, if both of the hard drives are from the same batch from the same manufacturer, there’s a very good chance they’ll fail around the same time. I’d highly recommend switching up brands, or at least timeframes in which they were manufactured, just to avoid both of them potentially failing at the same time.

NASes

A NAS—short for network-attached storage—is exactly what it says on the tin: it’s a small computer that runs, connected to your local network, that acts as file storage.⁴ These have the potential advantage of being available to any computer on the network, so everyone in your family could back up to one. As well, since there’s some computing ability available, these can do lots of other useful things with your files that are outside the scope of this blog post. If you have an open LAN port on your network router/switch (these tend not to be wireless, for bandwidth reasons), they are often plug-in-and-go.

Some NAS manufacturers are:

File-Storage Cloud Services

By this, I mean the cloud services that try to serve up your files as a file system, so that you can easily work on your files in them, rather than being a distant storage solution. They have the advantage of making sharing files with your friends and family easier, not to mention allowing your primary document location have some safety net by being in the cloud; the disadvantage is that your data is generally not encrypted, which runs into the issues I detail in the Cloud-First Services section. As well, they often try to up-sell you on additional products and services; this is a big place for vendor lock-in.

Some major file-storage cloud services are:

Cloud Storage Services

These are services that are meant for storing files—they almost always explicitly exist to fill the backup niche. They tend to be storage-oriented, and don’t have much in the way of additional products and services to sell. Their business is in keeping your data safe—and usually encrypted, so they have no access to its contents. A lot of times, these will have their own clients for doing the backup, provided as part of the service.

Some cloud storage services are:

Git Repositories

If you don’t already know what Git is, feel free to skip this section. However, if you do already know how to use Git, and the files you’re wanting to back up are plaintext—such as Scrivener projects or Obsidian vaults—you can actually make your project a Git repository, and push the data up to a server this way. This has the advantage of giving you excellent versioning and being backed by a common-use tool; the downside is that the main available Git servers will not encrypt your data, running into the issues I detail in the Cloud-First Services section. You can, of course, also set up your own Git server, but that can be a lot of work and maintenance burden.

Some public Git services are:

Cloud-First Services

I have, until this point, been talking about digital files that you have on your device.⁵ However, there are a number of services—such as Google Docs or World Anvil—that are cloud-first, storing your data primarily in the cloud, and you simply manipulate it with a web browser or an app. This, of course, poses a problem for backups in general, because you don’t have a file on your device to back up in the first place!

Now, in the general case, I personally avoid using cloud-first services as much as possible, especially for my creative work. There are several reasons for this:

Information in the cloud isn’t covered by my usual backup apparatus.
The cloud service could go out of business suddenly, and I lose all of my data.⁶
The company could close my account arbitrarily, cutting me off from my data.⁷
The company can use my data for training AIs without my permission.⁸
The company can scan my data for things considered “objectionable” and report that to government authorities.⁹
A government can compel the company to hand over all my data.

Obviously, my reasons for avoiding cloud services goes above and beyond “there’s no obvious way to back this up myself”. But in my opinion, the fact still holds: you are either in control of your data, or you are not.

I do realize, of course, that this is difficult vis-à-vis collaboration, especially as a number of these platforms are designed specifically for that. Cloud platforms also make it easier to share creative work online; and some people also are phone-primary users, which pushes them very much towards cloud services. I do in practice sometimes use them, myself.¹⁰ In these cases, what I recommend is making sure to do regular exports of your data, and backing those up; this is harder, of course, because it requires manual action instead of happening automatically. Another option—especially if your only real usage of these services is for sharing documents—is to still write documents locally, but upload/copy them to these services as necessary for that sharing.

Exceptions

There is a bit of a shell game I have been doing this whole blog post, where I started out by talking about your stories or other works of art, but then started talking about backing up your data. Obviously, stories are part of that, but your personal data consists of so much more: financial statements, cute birthday party invitations, photos, home movies, research for your upcoming vacation, and so on. And if you really think about all of your digital data, that also includes your email, your contacts, your browsing history, and on and on and on.

The good news is that a lot of that actually is covered by the solutions I’ve talked about here. You can back up not only your stories, but also your photos and any other important documents.

However, you may have realized by now that there’s still a lot of data in cloud services that doesn’t easily translate to local files anymore, especially once phones are involved. And given my screen about cloud services in the previous section, you may be wondering how I manage to function. The fact of the matter is, I do have a lot of things in cloud services primarily. Some examples:

Passwords, in a password manager
E-mail
Discord chats
Facebook posts
Address book
Calendar

And many more besides.

Ultimately, as I noted above, I try to export data as I need appropriate to back up, just in case something catastrophic happens. In the end, I evaluate the risks and effort for each thing, deciding what I allow to be ephemeral and what I want to make sure I save.

At the end of the day, though, anything I create? That’s the most important category, and I make sure those stay out of the cloud. But if I lost my collection of past and future appointments, it would be annoying, but enough places send reminders that I could reconstruct it well enough—at least, the future, which is more important in that case. You can make the same judgement calls regarding your own data. You can decide what risks you’re willing to take, and what effort is worth preserving what data.

My Setup

Having said all of the above, I’m now going to talk through my personal backup setup, including some thoughts on software/hardware/configuration. It’s a little excessive for a backup solution, but I’m the paranoid sort about files. Your mileage may vary.

My primary computing device is a MacBook Pro.¹¹ I keep my files organized through general-practice methods; I might go into more depth on that in the future. Its hard drive is encrypted.
I do use Git repositories for each of my projects, though usually just for versioning purposes; I don’t have remote repositories to push to.
I use Carbon Copy Cloner as the software to do my backups. I’ve tried Time Machine in the past, found it too finicky and flaky for my purposes. I like using something that gives me a lot more insight into what it’s doing when, and a lot more control over which files are backed up where at what time.
I have an external USB hard drive that is my first-line daily backup. Theoretically this is versioned, but I’ve never had opportunity to really confirm that. This is a full-system backup, and the hard drive is encrypted.
I have a Linux computer on my network that acts as a mini-server for things like Jellyfin and Minecraft. It also acts as a NAS for a daily full-system backup. Its hard drive is encrypted.
I have a Synology NAS on my network that I back up to regularly. This backup is more piecemeal—so I can easily share some files with my family—so is several different jobs that run at intervals that make sense for the data. Music, for instance, is backed up a lot less frequently than my stories.
I run Backblaze on my Mac, with a private encryption key.
I run Hyper Backup on my Synology NAS with private encryption keys to back up my NAS to C2.

If you want to see what this all looks like, I have the following diagram:

---
config:
  flowchart:
    curve: stepAfter
---
flowchart TD
    classDef device fill:#fed8dc,stroke:#baaa92
    classDef location fill:#edddc5,stroke:#baaa92

    subgraph home[In My House]
    laptop[My Computer]:::device
    external[External USB Drive]:::device
    nas[Synology NAS]:::device
    linux[Linux Computer]:::device
    end
    home:::location

    subgraph offsite[Off-Site]
    backblaze[Backblaze]:::device
    cloud[C2/S3]:::device
    end
    offsite:::location

    laptop -->|CCC| external
    laptop -->|CCC| linux
    laptop -->|CCC| nas

    laptop --->|Backblaze| backblaze
    nas -->|Hyper Backup| cloud

Disaster

The real trick to it all is mindset, and working through the possibilities, and considering potential failures in every step of the chain. What happens if your computer fails? What happens if the cloud service disables your account? What happens if an electrical storm fries all the plugged-in electronics in your house? Certainly, some of these are more likely than others, and that likelihood needs to be taken into account¹². For any backup solution you have, run through a little exercise, like this:

My novel is a Word file on a laptop.
My laptop crashes and loses all data, what then?
I have a copy of the novel on an external hard drive.
My house burns down and I lose both my laptop and external hard drive, what then?
I have an additional copy of my data in Backblaze.
I lost my Backblaze password and private encryption key in the fire, what then?
I have a copy of my Backblaze password and private encryption key in an envelope at my best friend’s house, letting me get access.

And remember, backups only matter if you can restore from them. Sometimes doing the exercise is just mentally going through hoops—but sometimes, it’s worth taking some of the steps for real, and making sure you always have access to your data.

At the end of the day, it’s disaster that we’re trying to mitigate. A little forethought and planning can go a long way when things go wrong—and sooner or later, they will.

Conclusion

I hope this was a very helpful overview of backup options, and that my talking through my setup as well provides some context and inspiration. However, I’m not going to judge you for evaluating the risks, costs, and tradeoffs yourself, and making a decision that makes sense for you. Maybe you like living risky and not having backups, maybe you just need an external hard drive, maybe all you want is to just have some helpful cloud storage, maybe you want something even more complicated than my setup. It’s all good.

But, with any luck, I’ve convinced you that it’s worth looking into. You’re a unique, wonderful person who’s contributed wonderful, unique things to the broader tapestry of humanity. What you’ve created matters, and I’d hate for us to lose out on it because your computer crashes or your phone decides to go for a swim.

Happy backups!

In my opinion, this should apply to all digital files, not just things you’ve written. Photos, pay stubs, signed legal agreements, and so on should in my opinion also be backed up. See further commentary in Exceptions. ↩
Phrasing it this way can be a little confusing, but I think it’s so that it can make the mnemonic of “3-2-1”. ↩
One time that springs to mind is when I accidentally broke my music library file when trying to do some heavy-handed organizing; I pulled the previous version out of Backblaze, and was therefore able to rewind to before the disastrous edit. ↩
“But wait!” you say, “Don’t NASes have hard drives? Why don’t you count them in the hard drives section as an external hard drive?” The short answer is: because they’re more like computers with storage attached. And there are different considerations with network storage compared to USB storage—the latter you just plug in and it works, the former can be more finicky, in my experience. ↩
And been munging between “all your computer data” and “just your writings”. ↩
This happened to Code Spaces and Smack Jeeves. ↩
Both Apple and Google are known to do this. ↩
Google claims that they only train on “publicly available” documents; I personally do not trust them not to secretly violate this policy, nor do I trust them not to change it in the future. ↩
It is worth noting that content that has at some point been (and in many cases is still) deemed “objectionable” by various governments includes LGBTQIA+ content, criticism of the government or government figures, dancing, the word “pregnancy”, discussion of alternate forms of government or economic systems, modern art, interracial relationships, wedding dresses, and explicit descriptions of sex. As well, the rules for this can change suddenly and so a story you wrote—sitting in a cloud service, unencrypted—where someone has a lesbian relationship suddenly becomes possession of pornography and is a jailable offense. Just because you think your content is “safe” right now does not mean that if the political winds change suddenly, it will remain so. And even if you then delete the story, what are the chances the cloud service provider still maintains a backup? Or that they ran sentiment or content analysis ahead of that, such that they can flag to authorities that you might possess such “objectionable” materials on your personal devices? ↩
For instance, if I were to share a spreadsheet in this blog, it would be a Google Doc. I’m not as hard-line about it as it might seem. ↩
In case you were wondering why I have more macOS examples than Windows examples, this is largely why: I’m more familiar with the ecosystem. ↩
The more generalized version of this is a Risk Matrix. ↩

Megan Bessel