Freelancers, do you host and maintain client websites? Why or why not?

Maintenance yes; hosting a definite no.

Maintenance is a straightforward relationship, be it either ad-hoc or a regular amout of work. You can provide value here - you know your client and their needs, and you can help them with that in ways that others' can't.

Hosting is problematic as you're reselling someone else's (the hosting provider's) product. They have processes to deal with hardware failures, people that don't pay, domains that expire, security vulnerabilities etc. They also have the volume of customers to cope with this. You can't provide any extra value here.

Hosting is quite commoditised these days. So I pick a service provider and recommend to a client that they sign up and make me an admin. The responsibilities are clearer all round.

But perhaps most importantly is that it gives both of us the freedom to walk away: the client can just stop giving me work, and if I can't take it on, they're in a position to employ someone else.

I haven’t freelanced in years, but have some thoughts.

I always found myself hosting stuff and regretting it later. I think I needed to either avoid this altogether or lean into it more as a core part of my offering. I wound up floating somewhere in the middle.

I have worked as an independent consultant on and off and it wasn't until recently that I started to host client applications. It's extra income, builds some loyalty and leads to additional paid work. However, if something goes down you are responsible to ensure it comes back up. At times clients assume that it was a bug in your work and don't want to pay.

This is a nightmare when the pipeline is complex and deployments have to be on a specific cloud provider. I've started to migrate clients onto their own setup as I prefer focusing my time on building applications and getting clients up and running on their own system versus playing devops.

I had answered a few questions related to freelancing on my #AMA that you might find useful.

I was a freelancer in my younger days and have returned to it now. When I started freelancing, I (like you) aimed to be a "one stop shop" for less tech savvy organizations. In my case, though, it was mostly artists and lawyers. I thought that I would do the hosting in addition to the design and coding of their sites and build loyal relationships and pocket a tidy margin for the hosting. However, I dropped that after about a year for a few big reasons:

1. I wasn't taking enough of a cut to make it worth my time. Unless you're going to be charging ongoing Squarespace-level prices, the overhead of staying on as an SRE for a client is unlikely to be worth any bonus you get to the relationship.
2. On top of that, I saw no bonuses to the relationships. I had hopes that it would help me close sales and help me gain repeat business from clients. The opposite happened, clients would be worried about me quitting one day and taking their site down with me. I had a much easier time selling my services when it included an account on a "reliable" hosting service that they would have the password and payment info for when we were done.
3. Finally, it actually wound up hurting my relationship with some clients. While a standard design & build contract had a defined "end" (when the site is accepted and goes live), hosting contracts never did. While this is, in theory, ok because your services should be limited to hosting after the initial build, my experience was that less technical clients didn't understand where the appropriate boundaries were. They failed to understand why I would jump into action if their site was down, but insist on charging them if they wanted me to make design tweaks. This wound up hurting the relationships that being a one-stop-shop was supposed to help and was what finally pushed me to stop pretending to be a hosting company.

All that being said, if you want to own being a hosting company (whereas I was more a design and build focused freelancer who was doing hosting as a "bonus feature"), you might be able to make it work. Just take my story as a caveat.

When it comes to maintenance, I think you as the developer are best qualified to do that, and it's best for your client that they don't let someone else do it - I've known a few situations where that goes horribly wrong, and ends in costly rewrites (who likes other people's code, after all?).

As their trusted developer, you can help your client navigate the jungle of different hosting options, and find the service that is best for them, but you probably should not undertake hosting yourself.

Do they need something that can scale to tens of thousands of users eventually? Do they need fast page loads, high uptime, mailing lists, or the ability to add more domain names easily for future campaigns? As the person who developed their site, you know best how to pick the right provider. Maybe you're comfortable setting up Wordpress and don't need the one-button install that some providers offer? Maybe some provider doesn't have experience hosting the Magento site you've built, and you don't trust them to do it well? Do you need lots of space for static pages, or a fast database? Does the site need email services, and what kind? Is DNS going to be hosted elsewhere, or do you need that, too? Is the site easy to cache, do you rely on pagespeed or any other Apache extensions? Will it be serving a single geographic region, or do you need a service with a global CDN?

The answers to each these questions are where your input provides value. But because they are likely different for each of your clients (unless you work in a specialized niche), the best hosting solution is also going to be different in each case.

Full Disclosure: I work for servebolt.com, where we provide hosting for e-commerce sites, with a focus on fast page loads. I'm obviously biased, but many of our clients have previously hosted their clients' sites, and are now moving to our service to focus on maintenance, and not have to deal with hosting.

I would not recommend it, and usually developers don't know much sys admin/devops in order to do a proper job, especially in security and performance settings. Also you have to consider the moral and legal implications of keeping their data, visitors and clients private data, and so on.

There is also a mix, having a hosting account (AWS or whatever) and both parties have access to it, and you can take extra responsibilities on it (like setup, maintenance etc), but you will take some of the responsabilities out of you, and you can always end the contract without any complications.

Still starting out in the freelance game (If anyone needs any WordPress work done, hit me up!) but from the couple projects I've worked on it's been easiest just to have them sign up for a Hosting package through NameCheap, or wherever, then I just have to load the site up and let their hosting provider deal with uptime and updates and security.

It'd be interesting to have some sort of auto-billing connected DigitalOcean droplets. Charge their credit card every month, if it fails try again after x amount of time, suspend the website after y days, if it keeps failing spin down the droplet after z amount of time. That way you only end up eating like 1 months worth of server costs for that one project. That way you can generate another revenue stream for your clients with only some up front time cost setting all of that up.

I feel like there are three approaches:

1) Don't host your client's websites. You can set them up for them, but afterward, you give them the credentials. Simplest and cleanest way to do it, and the way I prefer to work.

2) Host your client's site for a nominal fee. It feels like free money until it isn't. Getting an extra 20-30 buck per month/client will quickly not be worth it when the client is sending you an angry text at 2 am because their site is down.

3) Offering a recurring service package that makes it worth it for both of you. You aren't just providing hosting, you are providing availability and insurance. You could bundle this with an ongoing retainer that also includes some combination of support, consulting, or ongoing development. I wouldn't start considering an offering like this until we are talking $500/month minimum. At that point, hosting makes sense. But you still need a handoff plan in case the client ever decides to go in another direction.

Nope, I've never done it before. It never occurred to me honestly to propose I host their websites. They usually have their own infrastructure or I suggest using a commercial option. I've even referred clients to WordPress specialists because of requirements that didn't need a custom made web app (or even SquareSpace) :-)

Though I would probably start exploring the option of a static website in the future, with something like Hugo

I can recommend and set up a hosting environment for clients, i.e. environments they own and control themselves, but I would never host them myself. If they pay me separately to go check for updates I might do that, but I'd rather just not set up an environment that requires that. Setting up a script to easily do a Docker build based on Alpine Linux can save you a lot of trouble.

Static pages ftw here, if it can be hosted on S3/GitHub Pages/similar there's little need to worry about hosting.