How I cut my monthly cloud storage bill in half - with 5 tough decisions

My file and backup servers. That's a tenth of a petabyte right there. My security server is on a different shelf and the offline backup server mentioned below is upstairs. David Gewirtz/ZDNET Follow ZDNET: Add us as a preferred source on Google. ZDNET's key takeaways Cloud storage costs exploded after unlimited plans vanished Backing up everything was expensive and unnecessary Five strategic changes saved over $1,200 per year 60 terabytes. I'll never see that much cloud storage again. On one hand, I'm relieved. On the other, I'm deeply saddened. This is a story that began many chapters ago. It culminated with some serious self and infrastructure examination, a couple of hard conversations, and a few tough decisions. This is a story of online data hoarding in the guise of best practices, where too much became way too much, and the seemingly simple task of maintaining backups became an on-call emergency response second job. I changed all that this week. TL;DR There's a lot to this story, but if you want a quick summary, it's this: I moved from Google Workspace Enterprise to Backblaze B2. I dropped Time Machine and added Backblaze Personal for individual machines. My network of RAID servers help back each other up, and by the time I was done, I saved well over a thousand dollars. Also: The best cloud storage services of 2025: Expert tested But I had to give up a lot to do it. What follows is the necessary backstory, a description of the before-and-after network architecture, cost breakdowns, and an in-depth dive into how and why I made the decisions that led me to this new approach. Let's dig in. The before time Since what seems like forever, but probably dates back to the early 2000s, I've been a big proponent and practitioner of the tech industry's best-practice 3-2-1 backup strategy. The idea is to have three copies of every file, two of which are on different physical devices, and one of which is located off-site. Back before my parents passed away, the off-site was easy. I'd just shuttle a couple of hard drives back and forth to their place whenever we went to visit. In turn, they'd put me to work removing the thick hard-pack of computer viruses my Dad had picked up between visits. It wasn't entirely his fault. He was a jeweler, and back then jewelry sites were infested with malware. Also: Why some companies are backing away from the public cloud In any case, my storage needs were considerable, but not overly huge. Back when most people had dial-up, my self-hosted web servers lived in a converted linen closet. Those old servers used dedicated bandwidth provided by the only high-speed T-1 internet connection (which ran through my bathroom) between Rutgers and Princeton. At that time, my storage needs added up to less than a terabyte. Yes, that was a lot for the early days. But it wasn't 60 terabytes. Over time, my work moved me into three data volume intense areas: networked virtual machines, 3D modeling, and video. My local storage needs ballooned and have been growing ever larger for the past decade. Beyond the terabytes of networked VMs used for what we would now call AI agent development, I have all the video elements involved in producing fairly high-quality YouTube. A single YouTube video can come close to a terabyte during production. My multi-camera surveillance system around the house also eats up storage. Right now, I'm operating four local at-home file servers, with a total storage of 139.04 terabytes. Also: The best NAS devices of 2025: Expert tested That's a lot more than the 60 I have in the cloud, but there's a lot of the "2" part of the 3-2-1 backups on those servers. A little over ten years ago, storing in the cloud became not only practical but also cheap. You could buy all-you-can-store plans from a variety of online backup providers. And that's what I did. But by 2017, the never-ending feast of online storage came to an end. As vendor after vendor shut down their unlimited offerings, I found myself spending weeks moving my data to the next player still offering unlimited storage. I was trapped in a strange game of data warehouse Whac-A-Mole. Until I hit a wall. There came a time when those deals were no longer available, except for one offering from Google. In 2018, Google's G Suite (the earlier name for Google Workspace) offered an unlimited storage plan. If you bought five or more seats on a $12/mo business plan, $60 a month would get you unlimited storage. So, that's what I did. Then, in 2020, Google rebranded G Suite to Workspace and tightened its storage offerings. You could "request" unlimited storage, but having your request granted was another matter. For the next five years, my relationship with Google storage was rocky, increasingly demanding, and increasingly more expensive. You can read some of my pieces on this ongoing challenge here: By 2022, my two-person company was subscribed to a Google Workspace Enterprise plan, which did provide access to the storage I needed. It just took a lot of requesting to get Google to let me have that storage. That turned into something of an extended annual negotiation with Google's storage capacity gatekeepers. Every year or so, I'd get a notice from Google that I was approaching my capacity. Every year or so, I'd engage in weeks of back-and-forth support tickets "requesting" more capacity, citing reams of previous support transcripts. Every year or so, Google reluctantly agreed to a small capacity increase, along with an increase in how much I had to pay each month. A few weeks ago, I got another such warning notice. Instead of diving once more into diplomatic negotiations with the data overlords, I decided to rethink my life. Or, at least, the storage aspect thereof. Pre-migration storage architecture My email and storage costs have been linked ever since I started using Google Workspace. My main email account is a consumer Gmail account, configured to send and receive using my corporate domains. The corporate email domain relays are hosted in Workspace. The following graphic shows my pre-migration storage architecture. Our primary storage devices are the various computers my wife and I use, our file server (which holds all our primary documents as well as all my video backups), and a security server (which holds historical video from our security cameras). David Gewirtz/ZDNET The Macs were backed up to a local backup server using Time Machine. For the record, I dislike Time Machine. It's always a gamble whether you can restore anything from a Time Machine backup, and they fail regularly. Time Machine is an incremental backup. That means changes are added to the backup archive, and you (theoretically) can roll back to a previous day or week. The Macs also sync to our on-premises backup server (a different box from the file server) using ChronoSync, a very reliable Mac syncing tool with pretty excellent tech support. These backups are syncs, so anything that changes on the source machine (including deletions) is reflected on the server. Our big file server had four backup mechanisms. It backed up incrementally to Google Workspace. It also synced to Google Workspace, so Google Workspace had two complete backup architectures using storage. Also: Is OneDrive moving all your files? How to take back control of your Windows storage - 3 ways The file server also syncs to the backup server. Once a week, our offline backup server powers on, and the file server syncs to it. This provides an air-gapped backup just in case the worse were to happen. That offline file server is also pretty much fire and flood proof, so it should be able to handle almost anything. Keep in mind that all our servers implement RAID, so if any single drive fails, data won't be lost. I also keep spare drives in stock, so if there is a failure, I can replace the drive immediately and start rebuilding the RAID. Here's how our 3-2-1 strategy looked before the new architecture I've just implemented. David Gewirtz/ZDNET Pre-migration costs As with everything else in this world, the cost of cloud storage has been increasing regularly. In addition to the Google Workspace storage cost of $162.50 per month, we pay an additional two bucks each to extend the storage for our personal Gmail accounts. We also pay about $30 each month to extend our iCloud storage to 6TB. David Gewirtz/ZDNET Tough decisions As the chart below shows, our mission-critical document storage takes up only a sliver of the overall file storage on our big file server. David Gewirtz/ZDNET The videos in video production are all videos that have been uploaded to YouTube. Yes, the raw footage is stored on the server. I do pull from that footage from time-to-time. If I lost that raw footage, it would be a bummer, but not catastrophic. The same is true of media. That slice contains all of our imported DVD collection, along with music files, and other media, including my archive of the production graphics work for all of my ZDNET articles. It's an important folder, but as was the case with video production, most of it is archival. Since my wife and I subscribe to a ton of streaming video services, we almost never tap into that media library for video content. We haven't for years. Also: Your Windows PC has a secretly useful backup tool - here's how to access it As for the VMs, I keep a few live VMs on my super-charged M4 Mac Studio. The VMs stored on the server are from older projects and were used in massive simulations, projects I'm not doing right now. All of those VMs also use very deprecated versions of operating systems from times past. If something were to happen to them, it would be unfortunate, but it would not impact day-to-day productivity at all. All of those files are stored on one RAID server and backed up to a second local RAID server. It's highly unlikely that we'd have catastrophic failure across drives on both servers at the same time. So the real issue is what would happen if there were some sort of disaster. If something happened to the house and both servers were lost, then we'd lose those files. But if we had such a catastrophe, old VMs and video production archives would be the least of my worries. Tough decision #1: Stop backing up older media So my first tough decision was to stop backing up the three big slices to the cloud. Remember that not only was I backing all that up incrementally, but I was also syncing it all. Cloud data storage usage ballooned using that approach. Tough decision #2: Give up all that Google Workspace storage My second tough decision was to move off of Google Workspace for cloud storage. Giving up all that sweet, sweet storage brought me some sadness, but the constant debates with Google support had maxed out my tolerance meter. Tough decision #3: Stop using Time Machine My third tough decision was to stop using Time Machine. That's Apple's canonical recommendation for backup. It sucks. Since I'm already syncing my local Macs both to iCloud and to our backup server with ChronoSync, I felt I could shut down Time Machine. I then decided to move to Backblaze Personal for my wife's main Mac and mine. Backblaze Personal charges ten bucks a month or a hundred dollars a year for unlimited backups from one machine. You can have locally-mounted USB drives and it will back those up as well, but it won't backup network drives. Tough decision #4: Use Backblaze B2 for server cloud storage My fourth tough decision was to go with Backblaze B2 to back up just my file server documents. Backblaze B2 charges six dollars per terabyte per month, so I certainly wouldn't get the same storage deal as I was getting from Google Workspace. But if I kept my B2 needs to under 4TB (which is the case for the documents on the file server), I'd have a pretty solid budget-friendly solution. So far, as a new customer, Backblaze deserves points. I ran into a self-inflicted snag during my migration and the Backblaze support folks stepped in and got me out of trouble. Nice. Tough decision #5: No more incremental backups to the cloud My fifth tough decision was to forego incremental backups for the file server on the cloud. This would keep my Backblaze B2 costs down considerably. One neat feature of Backblaze B2 is that it retains files for a number of days. So I set file retention to 10 days. That means if I do delete a file I need, and the deletion syncs to the cloud, I have ten days to recover it. Post-migration budget and architecture So, here's where we now stand. This is the working architecture for the new normal. As you can see, a lot remains the same. But the file server backups to the cloud are substantially reduced, and Time Machine is no longer in the picture. David Gewirtz/ZDNET Here's how the new 3-2-1 backup strategy works out with this new approach. David Gewirtz/ZDNET I'm still keeping Google Workspace for corporate email relay, but that's a fairly inexpensive $14/month for our two accounts. And here's how it all nets out dollar-wise. David Gewirtz/ZDNET As you can see, I went from $196/month ($2,357 per year) down to $90/month ($1080 per year). That $1,200 yearly savings is pretty sweet. But it was a pretty challenging road getting here. Hopefully, in a year, I won't have another storage fight to chronicle. I'm hoping this new approach provides a few years of smooth sailing. Have you had to make hard choices about your cloud storage strategy? What services are you using to manage large amounts of data? Have you found a setup that balances cost, access, and redundancy? Have you considered ditching Time Machine or moving away from Google Workspace? How do you handle backups for media, VMs, or security footage? Let us know in the comments below. You can follow my day-to-day project updates on social media. Be sure to subscribe to my weekly update newsletter, and follow me on Twitter/X at @DavidGewirtz, on Facebook at Facebook.com/DavidGewirtz, on Instagram at Instagram.com/DavidGewirtz, on Bluesky at @DavidGewirtz.com, and on YouTube at YouTube.com/DavidGewirtzTV.