To become a consultant you need three things: a problem you have solved repeatedly for someone else, one person willing to pay you to solve it for them, and a price you can explain without apologising. You do not need a company, a website, a niche statement or a specialism (specialization) chosen from a list. Those follow the first client; they rarely produce one. I have done this twice, in 2013 and again in 2024, and the honest version is that the first time we got the work right and the price wrong, and the second time I knew enough to worry about different things.
In 2013, my co-founder Steve and I started the business that became DevOpsGroup, a services company helping other businesses build and run software. We both had day jobs. Eight years later the company was sold for over $30m. In 2024 I incorporated Critical Cloud, a services business that helps companies keep their software and cloud systems running reliably, and launched it in 2025. What follows is drawn from my notes, diaries and recollections of both starts. It is about becoming a consultant in the sense of selling your own expertise for a fee, alone or as the first of a firm; if what you want is the full story of how to start a consulting business with a co-founder, the first post on the newsletter tells it properly.
What the work actually is
The word consultant covers people who write reports and people who rebuild production systems at two in the morning. What they have in common is the thing the client is actually buying: judgement they do not have, applied to a problem they cannot afford to get wrong, with someone else taking responsibility for the answer.
I learned what that meant before we had a company. In 2012 I was working for a software firm on a programme to rebuild a motoring organisation’s website and move it out of its former owner’s data centre. The delivery team was releasing software several times a week into test environments, and proud of it. Then we tried to carry that pace into the client’s production organisation, and it stopped: every change went through a committee built for a world of rare, large releases. At the handover they looked at us as though we were speaking a different language, and they were right to. We had solved the problem we found interesting and handed them the one we had created.
The consultancy we started a year later existed because of that gap. The work was never the code; it was helping organisations rearrange themselves around the code. If you cannot describe the gap you close in a sentence a client would recognise from their own week, you are not ready to sell it, and no amount of setting up will change that.
The first client
We did not have one when we started. We had a blog, a point of view and rather more confidence in our technical abilities than our own website justified; WordPress gave us plenty of practice at the problem we proposed to solve for others. My first commercial idea was to become a reseller for somebody else’s software. The vendor turned us down before we had even incorporated.
What worked was slower and less glamorous. Publishing what we believed, in enough detail that competent people could disagree with it. Talking to every organisation that would talk to us about where their delivery and operations teams collided, until the same shape of problem had come up often enough to be sure it was real. Then asking one of them for a small, well-defined piece of work, priced by the day, with a clear end. A first client arrives because you have described their problem back to them better than they could themselves, and offered a bounded way to start. The pain questionnaire, a playbook in development, is the version of those conversations I would run now.
How to price when you have no history
In February 2013, a question about some possible work led me to ask Steve what our day rate should be, and my note from that time argued that, with no real overheads, we needed to be aggressive to win clients.
I can still understand the reasoning. We wanted a start, and we had less to pay for than an established supplier. But I had started from our costs when I needed to start from the customer’s problem, and I was reaching for a discount before anyone had asked for one. A low rent bill did not make our experience less valuable, or the customer less dependent on the work being done properly. That first rate becomes the number every later negotiation starts from, and moving it up is far harder than setting it well.
The method I would use now is to borrow a ladder rather than invent a number. Find the published or widely known rates for the grades of work you are offering, in your market, and place yourself on the rung your experience honestly earns. Define what a day is, on the card, so that the scope does not creep silently. Put the commercial terms beside the rate. Then hold the rate and flex the effort when a client pushes back, instead of discounting. The arithmetic is in the consultant day rate guide, and the method in How to build a consulting rate card with no billing history, a playbook in development with a rate card builder alongside it.
The day-job overlap
We kept our jobs through the founding period. The company was five days old when I bought access to a desk, on Easter Monday; the business cards ordered in February took months to arrive; we did not take an office until late 2014. For several months we were easier to find on the internet than in anybody’s wallet, and it did us no harm at all.
Two honest cautions. Read your employment contract before you sell anything, because many contain clauses about outside work, clients and intellectual property, and it is your problem if you have missed one; this is not legal advice, and a solicitor’s hour is cheap against the alternative. And notice what the overlap does to your pricing. A salary arriving every month makes it painless to underprice the new work, and the underpricing outlasts the salary. Decide the rate as though the job were already gone.
The trap of contractor augmentation
The easiest consulting revenue to win is not consulting. It is supplying yourself, or your people, into a role the client has already defined, on their site, under their management, at a day rate. It pays on time, it needs no proposal, and a recruitment agency will find it for you.
We did a great deal of it in the early years, and it kept the business alive. It also set expectations we spent years unwinding. A client who has bought a person at a day rate does not easily start buying an outcome at a price. Your team’s utilisation looks superb and your margin is whatever the agency left you. And placed people are priced by the market for placed people, not by the value of the problem you can solve. In 2018 we took the explicit decision to leave that line of work; the change in UK tax rules for contractors the year before was one cause, and the larger one was that it was not the business we had set out to build. If you take augmentation work to fund the start, and many should, name it as that, keep it separate on the page, and set a date by which it stops being most of your revenue.
What changed between 2013 and 2024
The first time, the gap between incorporating and doing paid work was measured in weeks. The second time, nine months passed between forming Critical Cloud and launching it. The 2013 version of me would have had questions about that.
Experience gave me better questions, not the answers. I knew more about how a price becomes a promise, how that promise reaches the people you hire, and how hard it is to change once clients rely on it. I could see far more of what might go wrong, which is useful right up to the point where preparation becomes a comfortable way of postponing the conversation with a customer. In 2013 we could act quickly because there were whole categories of problem we had not met yet, and there was value in the smallness of it: we could test an idea, change our minds and try again without dismantling a department first. Whether each of the 2024 differences proves useful is something I am finding out by running the business, and I make no claims about it yet.
If you are changing career
Some of the people searching this phrase are not practitioners with ten years of scars. They are considering consulting as a new career, sometimes from a very different one. The honest answer is that consulting is not an entry point; it is what a track record becomes when you decide to sell it directly. A client is paying for the problem you have solved before.
That does not mean the door is shut. It means the order is different. Either consult on what your existing career has already taught you, which is usually more saleable than it feels from the inside, or go and acquire the record in the new field first, as an employee or on narrow, well-defined pieces of work priced honestly for where you are. Then price the day properly from the start, using the consultant day rate guide, and read a few of the stories on the newsletter before you resign from anything. Several of them are about decisions that looked like progress at the time.
Where to go next
The rate is the decision that shapes everything after it, so start there: the consultant day rate guide walks from a cost base to a defensible card, and the consulting rate card playbook, in development, gives the full method, with the rate card builder beside it. For an independent consultant working alone, the independent consultant guide covers setting up, pricing and the first client with one person in the arithmetic.