Yes, project failure can often be laid at the feet of the client. But, IMO, good professionals contract developers are responsible for pushing back on vague and/or changing requirements. Part of your job is to look underneath what the client says to discern what he actually needs. The goal should be to and deliver something that is both useful and fulfills the client's goals. All of this is, of course, easier said than done.
How do you start from nothing?
Anyway, here's what I would do.
If you have nothing to show I would start with small things that have big ROI on "showing off".
Help some non-profit with their website for a few hundred (or for free) make sure to establish a warm relationship and then you can say "I worked for this Church, the Red Cross, University of X" or something like that, "you can contact X for more details", if you did something that's public then show it off! It's all about that.
If you're after technical jobs, contribute to some open source projects, just close down some tickets get on the AUTHORS file and then you can say "I contributed to Firefox" or something.
If you think a bit about it, you can align this experience towards the job you want, a concrete example:
You want to land an incredibly well paid job on the AI/ML space.
* Check out tinygrad, they have a list of open issues with bounties, meaning that you wouldn't be working for free.
* Solve three or four of them. Do it well, documentation, follow procedures, be autonomous, etc...
(by this point I would offer you a job and I'm sure many others would as well, but that's specific to that particular niche)
You get the idea, adapt to whatever you want to do.
Edit: I just realized you're talking about a firm not yourself as a developer, lol.
Most of it still applies I would only add that many people have told me that expos and industry events are where they get like 80% of their clients, it is really worth investing in them, set up a small booth on a local SMB fair or something.
They know what they need and pay what it costs. They know what they don’t need and avoid it…
In any case, sales/marketing/advertising is manifold. Have a website with a good explanation of your services and examples of projects you've worked on. (No past projects? Create some, even if just demos/proofs-of-concept.) Make it easy for visitors to join your mailing list for updates (maybe offer them a free whitepaper for joining). Find out where potential customers hang out (whether online or meatspace) and hang out there with them. Ask questions. Discover their pain points and address them. Look for sites that list contracts (yes, there are a bunch) and comb through them carefully and frequently. Don't try to be all things to all people, i.e. focus on customers and projects that can use your strengths. Create a free product and release it to get customers, then try to upsell them on extra features.
Knowing what type of software you specialize in would make it easier to give more specific suggestions. Government contracts are way different from websites for mom-and-pop shops which are themselves way different from embedded systems. Each of those spaces has their own challenges and quirks.
Since that’s hard work, you could consider this answer “hard work.”
But only the right kind of hard work makes a difference. If you can’t be directly rejected, it doesn’t count as sales.
You can always hire technical talent, later. Sales is all that matters and is a people business.
Getting started is the hardest part, like many are saying, its simpler when you can rely on your network.
Everything is about trust, by this, start reaching out to people who already trust you, there is a chance that this will get you either a project or a referral. Once you build a portfolio, you will keep getting these referrals.
In my case, I have got most of our contracts by sending cold emails, I have got a few customers through the "Who's looking for freelancers" posts + many others by just talking to people.
Lately, market is flooded with AI-generated stuff and fake job seekers which is causing people to distrust others, it seems harder to get contracts through cold outreach and the referrals we have got have done wonders.
You cannot control when a customer buys and you cannot create a sale.
All you can do is charge enough to stay in business and try to be worth your customers wanting to keep you in business.
Nobody cares if you start a business; you will be competing against companies whose customers pull their bootstraps; and starting out you will mostly deal with the inexperienced and the grifters.
Our contracts are literally all self-referred or reputation-referred. My partner and I each have multiple long-term clients, some of whom are former fulltime employers.
On that latter point, a very common method is to "quit" your employer and then they immediately retain you as your anchor client before you begin the process of shopping for more.
We have bench engineer staff and I am the only one who does sales. I write the contracts, as well as do the design work and engineering alongside my coworkers. I consider us a software development firm, but there are other companies that call themselves "agencies" in the valley.
They do a bit more full-service work, while we focus on custom software specifically.
The answer is that I work alongside a network of developers who have an established history of contacts here in Arizona, and I've also been in the industry for a number of years and previously worked at a large number of companies doing a mix of professional services work, contract, and full-time employment. That's the long way of saying that I obtain business by word-of-mouth.
Separately from local contacts here in the state and abroad based on remote work I and others have done, we publish some notes of design work and software development that others do not which garners some attention from clients who acknowledge us as subject matter experts. This is also partially validated by press publications and industry awards.
Additionally, we review requests for proposals where we think we can provide a valuable service, and we work on some internal products by reinvesting earnings into developing technologies from research and development efforts. Fun novel stuff that doesn't really exist in the open source world. In this wing of the business, we focus on competing technologies to proprietary solutions that exist in the wild.
So, it takes some exceptional work to gain enough traction to stay in business, and I think by extension, that's why you don't see too many of these businesses stay afloat.
We've been in business for maybe, I'm not sure now, 5 or 6 years. Not long. But we do so by doing things others don't, or can't. So in terms of market segmentation, we address solutions from a premium perspective.
You got a great answer from @andrewmcwatters and now you and everyone else reading this knows that they're available for custom software development of some type, as a side effect.