Most small businesses don’t need “the cloud.” They need a Linux box on the internet that doesn’t fall over when their landing page lands a podcast mention. The way the industry talks about cloud infrastructure assumes you’re Netflix; the reality for a six-person shop is much smaller.
I’ve been deploying small-business workloads to cloud providers for years. WordPress sites, FusionInvoice installs, Nextcloud servers, internal tools, the occasional Django app. The pattern that’s held up best across all of them is: DigitalOcean Droplet, Ubuntu LTS, plain old SSH. Here’s why, plus the workloads where I actually do reach for AWS or GCP.
What I actually run for clients
Real deployments I’ve set up in the last 18 months, with what they cost and what they hold:
| Workload | Droplet size | Monthly cost | Notes |
|---|---|---|---|
| Single WordPress site | 1 GB / 1 vCPU | $6 | Most small-business sites fit here easily. Cloudflare in front. |
| 4–6 WordPress sites, shared host | 2 GB / 1 vCPU | $18 | Virtualmin or plain nginx vhosts. Common for consultancies hosting their own client sites. |
| FusionInvoice + DB | 1 GB / 1 vCPU | $6 | Plus ~$5/mo Spaces for nightly DB dumps. |
| Nextcloud (5–10 users) | 2 GB / 1 vCPU + 100 GB volume | ~$28 | The volume cost dominates. Add Spaces if storage grows. |
| Internal Django app | 2 GB / 2 vCPU + managed Postgres | ~$45 | Managed Postgres is $15/mo and worth every cent vs. Self-managing. |
None of these are big numbers. None of them needed Kubernetes. The biggest one on this list runs cheaper than a single QuickBooks Online subscription.
Why DigitalOcean keeps winning
It’s not that DO is the most powerful or the cheapest. It’s that the friction is the lowest, and for small-business workloads that matters more than anything else.
- Predictable pricing. A $6 Droplet costs $6. There’s no surprise egress bill, no NAT Gateway hidden line item, no "you forgot to detach the elastic IP" gotcha. (See why DigitalOcean handles IPs differently for the most expensive version of that gotcha on AWS.)
- You can hand the dashboard to a non-engineer. When a client needs to reboot their own droplet at 9pm, they can. Try that with the AWS console.
- The Marketplace one-click images actually work. WordPress, GitLab, Nextcloud, Plesk. The images are maintained, current, and don’t leave you on a five-year-old Ubuntu.
- Backups are $1.20/month per Droplet as a checkbox, and they actually restore.
When I do reach for AWS or GCP
Three honest cases where the upcharge is earned:
- The workload needs a service that only AWS has. Specific examples I’ve hit: SES for high-volume transactional email, Kinesis for event streaming, SageMaker for managed ML training. If you need the service, you go where the service is.
- The client’s industry requires it. Healthcare clients running anything HIPAA-adjacent usually want AWS or GCP because the BAA paperwork is simpler. DO has a BAA available too, but the conversation is shorter on AWS.
- The team is already in that ecosystem. If everyone else at the company already lives in Google Workspace + GCP, fighting that for the sake of a $6 droplet is silly.
For everything else (which is most engagements I do) DigitalOcean wins on time-to-running and total cost of ownership.
What the boring setup looks like
The base recipe I run on a new Droplet, in roughly this order:
- Spin up the smallest Droplet that fits the workload. You can resize up later (CPU/RAM, with a reboot; disk is one-way).
- Ubuntu LTS, SSH keys only, no password login.
ufwfor the firewall. Allow 22, 80, 443 and nothing else.- Cloudflare in front of the public hostname; origin only accepts HTTPS from Cloudflare IPs.
unattended-upgradesfor security patches.- DO weekly backups enabled.
- App-specific stack on top (nginx + PHP + MySQL for WordPress/FusionInvoice; Docker compose for anything more complex).
That’s it. Across most client deployments, I touch the server maybe once a quarter for a major version upgrade or a config change. The rest takes care of itself.
When you don’t need any of this
Worth saying out loud: if your one WordPress site gets a thousand visits a month and your shared host is keeping up, stay there. Cloud servers aren’t a status symbol. The reason to move is a real bottleneck. The host is throttling you, the site is slow, you need software your shared host won’t run, or you’re paying enterprise managed-hosting prices ($30–$100/mo) for something a $6 Droplet would do better.
If you want help
I set up and manage these stacks for clients across Omaha and remote. One-time setup, or ongoing management with monthly patching and monitoring. Drop me a line and we’ll talk through what your workload actually needs.