Ask HN: Developer-as-a-Service?
I’m currently testing this. One customer is already paying $2,500/month for development services. I’m wondering if I should provide my freelance services this way.
Like with DesignJoy, only one active task at a time. We use GitHub to manage almost everything. CI/CD is configured so deployment is a breeze. Each issue is a task. We use GitHub projects (kanban-style board) to manage the backlog and task stages (todo, in-progress, done). Big projects are broken into milestones and stages; each milestone or stage corresponds to one month of work.
Another benefit I see is that the costs is clear to the customer; he can budget properly and may pause things as needed. I also don’t need to spend time preparing estimates/quotes (which I’m not good at, and I don’t enjoy doing).
I find this works well for people looking to build MVPs or small to medium web applications. There are no long-term contracts or strict project scopes.
What do you think? Any feedback and discussion is welcome.
Isn't it called consultant/consultancy?
Isn't consultancy more of providing guidance and advice about specific subjects?
Yes, and senior is supposed to be the highest rank of IC. But at some point, people gave this advice to call yourself a consultant instead of a freelancer... and it worked. So we have to live with this new term, and so do actual consultants.
I prefer contractor because it doesn't have the same stigma as freelancer, and is still accurate about their relationship.
My wife is from civil/construction engineering and she thinks terms like developer, engineer, and architect are all BS when software people use them. But idk, eventually we need terms for what we do.
This way wouldn't it be hard for the customers to choose between developers with different skill level and productivity? Some developers get more done in a certain amount of time than others. If the price per time ends up being dependent on the project workloads / productivity etc. it might as well be back to the way things were because you will still need to spend time justifying your price imo (since there are competitors out there).
I don't see why doing DaaS would limit the customer in choosing developers. I mean, I'm not talking about having a SaaS-style platform where the customer registers an account and submits work. I'm talking about freelancer-customer relationships and how the freelancer charges for its services; instead of a fixed, per-project price, the customer pays the freelancer's monthly fee (which could also be seen as a monthly retainer?) and submits any work he'd like.
Modern web design is better commoditized than software development. For example, you need to paint typical pictures for every new blog post, press release etc.
But you're expected to build a blogging software only once, and any fixes/bugs are unexpected in scope and complexity.
Unless you take advantage of your customer (nobody else can fix that complex system that you left him with) - it doesn't look sustainable to me.
Valid point! But don't you think there could be instances where this could be useful for both ends, the freelancer and the customer?
There are instances where the customer already know what he wants, could be typical maintenance, bug fixing, etc... those things can be added to a task queue, details are agreed and work is performed, without preparing quotes every time for each task
For example, I would only use something like this if I can automate as much as possible using CI/CD, etc
But having predictable maintenance and fixing is counter the whole point of software engineering.
If you can predict how much and what issues need to be solved - why haven’t you already automated it??
Sounds like a competitor to existing platforms like Upwork/Fiver etc?
Not really, it's more of the way I would charge my customer as a freelancer. When I mention development-as-a-service, I don't mean to create a platform, its just a way to describe the way I see this (charging a monthly fee in exchange for queued work)
nice try, soham
I see your point, but it's not the intention here :)
It's a joke obviously, but the deeper point is context is extremely important. In the pre-AI days, we'd budget 3 months for someone to get the context of the codebase and the product. These days, it's still 2-4 weeks. You can have someone extremely skilled spread out on multiple projects and they'd do a terrible job. Whatever low hanging fruit is all taken up by AI.
The exception I've seen so far is something like devops or infra. Where the person has done their job whenever nothing goes wrong. People are happy to pay monthly for nothing going wrong.