Some years ago, I thought I wanted to scale from a solo freelancer into an agency. I had two reasons: 1) money and 2) accomplishment. It's now my belief that building an agency is not a good way to achieve either of those two goals.
I went a small ways down the path of building an agency. I worked on a project where I hired two other freelance developers and one designer to help me. It was a lot of mental overhead and at the end of it, I didn't make that much money. Taking into account the stress of managing other people and the risk of being responsible for other people's work (including mistakes), I ended up wishing I hadn't gotten any other people involved at all.
One of the least attractive parts of running an agency I can think of is the boom-and-bust cycle. Almost every freelancer/consultant has a boom-bust cycle. With an agency the symptoms of the boom-bust cycle are amplified and made much more painful and serious. I've seen agencies have to lay people off when times are lean. Then when times are "good", you have to scramble to hire more people to do all the work - or make your employees work overtime, which I think is pretty unfair (I've been in this position at agencies).
So at some point I decided I didn't want to build an agency after all. Instead I decided to make as much money as possible (in as few hours as possible) as a solo consultant. Based on the fact that [Alan Weiss](https://www.amazon.com/Million-Dollar-Consulting-Alan-Weiss/...) claims (in such a way that I believe) to earn over a million dollars a year as a solo consultant, I understand it's possible to earn quite a lot as just a single individual.
My biggest success in increasing my earnings has been to move from development to training. I haven't succeeded in charging more than $100/hr as a developer (except some tiny projects and emergencies) but I've charged as much as $13,000 a week as a trainer. Some trainers I know charge $20,000 a week or more.
I've also had some small successes selling products. I wrote an ebook in 2016 that sold about $8,000 worth. Now I'm working on the next product.
So if someone came to me with this question of how to scale into an agency, my answer would be that I think there are way better ways to get what you think an agency will get you, unless for some reason you really want to start an agency for the sake of starting an agency.
If you're looking to up your game from Freelancer/Contractor, you'll do well to read Alan Weiss, The Guru of building a successful consulting business > http://www.amazon.com/Million-Dollar-Consulting-Alan-Weiss/d...
I don't agree that you should focus on anything less than the "Give me results" clients. First off, you know what they say is the problem with goals? You'll probably reach them. Meaning, you're setting your ceiling. I've found that if you hand pick your clients, you can make certain that you have clients who focus on results. Most consultants talk about word of mouth as the main way that they get new clients, but I dislike that approach. The reason is that you're letting clients choose you. My best client is a client that I picked and cold called. I knew they were making lots of money and I knew they needed what I was selling. Selling something as a consultant is about specialization. Specialization doesn't necessarily mean that your experience is focused in one area. It means that you can present yourself as an expert in one area.
Here's something counterintuitive that I've found that goes along with this article: Clients who pay the least are usually the most demanding. I used to lower my price when people complained, but I quickly realized that my price was a filter blocking bad clients. Plus accepting a lower price really led to likely bad outcomes because when the going got tough, the voice in the back of my head said, "These guys are paying you less than your other jobs", then I suddenly felt completely unmotivated to work hard for them.
Here's a really good book about consulting that helped me. Not focused on software consulting, but a lot of the concepts are the same:
His take is that consultant is someone who shares their knowledge of process and works strategically with the client's decision makers... not someone who is knee deep in doing the work of implementing a project. If you're talking about writing PHP code, you may be more setting yourself up as a one man staffing agency, not as a consultant. I'd suggest reading Weiss, think hard about what you really want to do, and go from there.