
You made the leap and the firm is real. You survived the launch, which the first two posts in this series walked through, Breaking Away, Part 2, and now you have clients, revenue, and a busy season behind you. Next comes the part nobody warns you about: the transition from a person who does tax work into a firm that runs. Most of that transition is about finding a niche in accounting and building everything else around it.
In the first year or two you took every client who called. That was the right move, and we will start there. But the firms that grow up are the ones that eventually get selective, specialize, systematize, and climb from compliance work into advisory. Here is how that actually happens.
When you are new, taking every client who walks in is not a failure of discipline. It is how you survive. You need the cash flow, you need reps across a range of situations, and every satisfied client is a referral source you cannot afford to turn away yet. Saying yes builds the book and teaches you which work you are fast at and which work drains you.
The problem is that yes has a shelf life. A book built entirely on "whoever called" becomes a pile of mismatched engagements, each with its own software quirks, its own deadlines, and its own learning curve. You are busy, but busy in twenty directions, and your effective rate is a fraction of what it looks like on paper. That is not a client problem. It is a signal that the first phase has done its job and the next one is due.
Pull your client list and be honest about it. In most practices a minority of clients generate the majority of the profit, and a different group generate the majority of the pain: the scope creep, the late payments, the 11 p.m. texts, the returns that pay like a 1040 but bleed like an audit. That is the Pareto split, and once you see it you cannot unsee it.
Growing up starts with acting on that list. For the clients who cost more than they are worth, you have three moves: reprice them to what the work actually demands, restructure the engagement so it fits your process, or let them go. Firing a paying client feels wrong the first time. It is not. Every hour spent on a client you should have released is an hour stolen from the clients and the work you want more of. Cull once a year, right after busy season, while the memory is fresh.
Once you have shed the dead weight, a pattern usually remains: the clients you serve well, profitably, and without dread. That pattern is the start of a niche, and leaning into it is the single highest-leverage move a small firm can make.
The economics are not subtle. Specialized firms consistently report materially higher revenue per partner than comparable generalists, 20 to 40 percent higher by the Rosenberg MAP Survey, because expertise that is hard to replace is hard to shop on price. A generalist billing $200 an hour competes with every other generalist in town. A CPA who knows construction job costing and bonding, or the tax quirks of dental practices, or crypto cost basis, competes with almost no one.
Pick the niche the way it picks you. Look for the intersection of three things: you already have clients there, the work carries real complexity that generalists get wrong, and the market is large enough to feed you. Chasing a niche you read was "hot" but have no foothold in is how you become a generalist with a fancier website. Think of medicine: the primary-care generalist is paid the least and refers out to specialists who command more for a narrower range of work.
Niching pays a second dividend: it makes the work repeatable, which is what lets you build systems. When most of your clients look alike, you can standardize onboarding, checklists, workpapers, and client communications instead of reinventing every engagement. Standardized processes mean more consistent deliverables, faster onboarding, lower training cost, and margins that do not hinge on you personally grinding through the night.
This is the line between owning a job and owning a firm. A job needs you in every seat. A firm has documented processes, a workflow system tracking every client and due date, and a standard way things get done, so the practice can absorb more volume, survive your vacation, and eventually run on other people's hands. Build the systems while the firm is small and simple. Retrofitting them onto chaos later is much harder.
Eventually you hit capacity, and the instinct is to hire. Do it deliberately. The CPA pipeline is still thin, as we covered earlier in this series, Breaking Away, Part 1, so talent is expensive and slow to find, and a bad first hire can set you back a year.
Before you add headcount, add leverage. Automation and AI now absorb a real share of the routine load: Thomson Reuters' 2025 Future of Professionals research has practitioners expecting AI to save around five hours a week, and roughly a third of tax firms already use generative AI in their work. Squeeze that out first. When you do hire, hire against your specific constraint, whether that is a preparer to clear volume or an offshore or fractional resource for the production layer, and keep your own hours on the judgment work only you can do.
The last move up is the biggest. Compliance work, the returns and the bookkeeping, is exactly the work software and AI are commoditizing fastest, which means its price only heads one direction. Advisory work runs the opposite way.
Client advisory services are the fastest-growing area in public accounting, and firms that lean into higher-level advisory earn more than 30 percent higher monthly recurring revenue than compliance-focused peers. Advisory is stickier, too: a client you talk to monthly about cash flow and strategy does not leave over a fee the way a once-a-year tax client will. Layer advisory on top of a niche and the two compound, because the specialist who understands an industry cold is exactly who that industry wants strategic advice from. That combination, a defined niche delivered through systems and priced as advice, is what a grown-up firm looks like.
A lean, specialized firm doing more advisory work still has to answer hard, often niche-specific questions fast, and you no longer have a research department down the hall. This is the workflow tools like Marble's Intelligence agent are built for: ask a federal or state tax question in plain English, get an instant citation-backed answer tied to the controlling authority, and turn it into a client memo or explanation ready for your review. For a small firm, that is how one person delivers work that used to take a team. Sign up for Marble.
Not on day one. Early on you need volume and reps more than focus, and you cannot know which clients suit you until you have served a range of them. The right time is usually after a year or two, once you can pull your client list and see which work is profitable and repeatable. Let the niche emerge from evidence rather than a guess.
Look for the overlap of three things: industries where you already have clients and relationships, work that carries enough complexity that generalists struggle with it, and a market large enough to sustain your revenue goals. The complexity is what supports premium pricing, because if any generalist can do the work, there is no premium in it. Start with your own book, since your existing clients are the best evidence of where you already have a foothold.
No. Specializing does not mean firing everyone outside the niche. It means concentrating your growth, marketing, and systems around it while keeping the good clients you already have. Many firms stack a couple of complementary niches or keep a general base alongside a specialty. What you shed is not the off-niche work, it is the unprofitable and draining work the 80/20 sort flagged.
Yes, and arguably more so for a small firm than a large one. Advisory carries higher margins, builds recurring revenue that smooths the feast-or-famine of a deadline-driven book, and creates the year-round relationships that make clients stay. A solo or small firm can start simply, with cash-flow visibility or strategic planning for clients you already know well, and grow the advisory line from there.