Scaling a Law Firm From Five People to Fifteen

A five-person firm runs on proximity. Everyone hears the same phone calls, everyone knows which matter is on fire this week, and the founder can hold the entire caseload in their head while making coffee. Coordination costs nothing because it happens by accident. If a paralegal is unsure whether a filing went out, she turns around and asks. If a client calls about a settlement figure, whoever picks up already knows the story. That is not a system. It is a small enough room that a system is not required yet.

Fifteen people is not three times five people. The work of coordination scales with the number of pairs of people who need to stay in sync, and that number climbs far faster than headcount. At five, there are ten relationships to maintain. At fifteen, there are more than a hundred. Somewhere in the middle, usually between seven and nine, the room stops being small enough and the firm starts running on assumptions instead of information. Nobody announces this. The first sign is usually a client asking a question that three people answer three different ways.

This post is about the four specific things that break on the way from five to fifteen, and roughly when each one goes. Informal coordination stops working first. Then the founder becomes the bottleneck, usually while insisting they are not. Then permissions and supervision have to stop being cultural and start being structural. Then, last and most painfully, intuition stops being a reliable way to run the business and reporting has to take over. Firms that survive this stretch tend to fix them in that order. Firms that struggle tend to fix them in the order the crisis arrives.

The First Break: Coordination Stops Being Free

At five people, information moves by ambient exposure. Nobody has to be told the status of the Hendricks file because the two people working on it sit within earshot of everyone else. This creates a dangerous illusion, which is that the firm has good communication. It does not. It has no need for communication, which is a different thing, and the difference only becomes visible when you add a sixth and seventh person who sit in a different room, or work a different schedule, or were hired remotely and have never met the paralegal who has been carrying half the institutional knowledge for four years.

The tell is duplicated or dropped work. Two people call the same opposing counsel in the same afternoon. A response deadline gets logged in one person's personal calendar and nowhere else, and when that person takes a week off, the deadline goes with them. A client asks for a document that was produced two months ago and three people search three different folders. None of these are dramatic failures, which is exactly why firms tolerate them for too long. The correct response is to move status out of people's heads and into the matter itself, so that the answer to "where is this" is a place rather than a person. A matter stage tracker does this well precisely because it is boring: a clickable stepper on the matter, configurable per practice area, showing exactly which stage the file sits in. In Casely, the deadline diary works the same way. Deadlines attach to the matter with next-date auto-tracking, so they survive holidays, resignations, and the founder's memory.

The Second Break: The Founder Becomes the Bottleneck

Every growing firm has a moment where the founder is genuinely the constraint on revenue, and almost every founder misdiagnoses it as a capacity problem in the associates. The pattern is familiar. Every engagement letter needs the founder's signature. Every fee decision routes through the founder. Every difficult client call gets escalated to the founder because that is what the clients expect and what the team has been trained to do. The founder is now working sixty hours and doing perhaps twenty hours of work only they can do. The other forty is a queue.

What makes this hard is that the bottleneck usually formed for good reasons. In the early years, the founder was the only person whose judgment was reliable on fee structures, on which matters to decline, on how to talk to a difficult client. Delegating those calls felt reckless because it was. The mistake is failing to notice when that stopped being true. The practical fix is to separate decisions that require the founder's judgment from decisions that merely require somebody's authority, and then push the second category down with written thresholds. A senior associate can approve a write-off below a stated amount. A practice lead can approve a payment plan within stated parameters. The founder reviews the exceptions. Firms that do this well tend to write the thresholds down in a single place and revisit them every quarter, raising them as the team earns it.

The Third Break: Supervision Has to Become Structural

At five people, supervision is a byproduct of proximity. The founder reads almost everything because almost everything crosses the founder's desk anyway. At twelve, that is arithmetically impossible, and the firm quietly shifts to a model where junior work goes out because nobody had a reason to stop it. This is the failure mode that carries the most professional risk, because supervisory obligations do not scale down when you get busy. The exact contours vary: the US model rules on supervisory responsibility, the SRA Codes in England and Wales, the provincial rules in Canada and the state-based rules in Australia all impose supervision duties, but they differ meaningfully in scope, in how they treat non-lawyer staff, and in what documentation regulators expect. Confirm your own jurisdiction's requirements rather than assuming the shape of the rule you learned somewhere else.

Structural supervision means the review step exists in the workflow, not in someone's good intentions. That means a stage in the matter tracker that a file cannot pass without a named reviewer, a document history that records what changed and why rather than just when, and a clear map of who supervises whom that survives reorganisation. Casely records a comment field on every document capturing what changed and why, which turns version history from a list of timestamps into a readable account of the file's development. That matters twice over: once for actual quality control, and once again eighteen months later when somebody has to reconstruct who decided what. The firms that handle this stretch well treat supervision as a designed path through the work rather than an attitude the partners hold.

The Fourth Break: Permissions Stop Being Cultural

Small firms run on an implicit norm that everybody can see everything and everybody behaves. That norm is doing a lot of unacknowledged work. It falls apart the moment you have a lateral hire whose old firm sits on the other side of a live matter, a contract paralegal who should see three files and not three hundred, or two clients whose interests diverge in a way that requires an actual screen rather than a promise. At that point, "we trust our people" is not an answer, because the question a regulator or an opposing party asks is not whether you trust them. It is whether the restriction was effective.

This is where a lot of practice management software quietly fails firms, because it hides restricted matters in the interface while leaving them reachable through search, a calendar entry, a notification, or a link somebody forwards in good faith. That is a cosmetic wall. Casely enforces ethical walls at the server and data-access layer, which means a walled user genuinely cannot reach a restricted matter by any route, including search, the calendar, or a forwarded link. The distinction sounds like a technical detail until you have to describe your screening arrangement to somebody who is not inclined to take your word for it. Screening rules and the availability of screening to cure imputed conflicts differ significantly between jurisdictions, so check what your own regulator requires before relying on any arrangement.

Trust Accounting Punishes Informality First

If you want an early warning system for whether your firm has outgrown its habits, look at the client account. Trust handling is the one area where informal process converts directly into regulatory exposure, and it is also the area where growth does the most damage fastest. At five people, one person touches the trust ledger and reconciles it monthly and knows every balance by feel. At fifteen, three or four people are requesting disbursements, the reconciliation is somebody's part-time responsibility, and the feel is gone. Overdrafts on a single matter's trust balance are almost never theft. They are almost always someone paying an invoice out of the wrong matter's funds because the balance looked fine at the aggregate level.

The structural fix is per-matter isolation plus a hard block rather than a warning. Casely holds isolated ledgers per matter and blocks any disbursement that exceeds that matter's actual trust balance at the database transaction level, not with a dialog box someone can click through at six in the evening. Corrections are voided and remain visible rather than deleted, which is the behaviour most regulators expect from a proper audit trail. The specifics differ by jurisdiction, with IOLTA and state bar rules in the US, the SRA Accounts Rules in England and Wales, and law society rules across Canadian provinces and Australian states each setting their own reconciliation cadence and reporting duties, so confirm the requirements that actually bind you. What does not vary is the principle: as headcount rises, the number of people capable of making an honest trust mistake rises with it.

Intake and Conflicts Break Quietly Before They Break Loudly

Conflict checking at five people is a memory exercise. The founder recognises the name, or the office manager does, and that recognition has been sufficient for years. It stays sufficient right up until the firm has processed enough matters that no single person has seen all of them, and enough people that intake is happening in more than one place. Then the check quietly degrades into a search of the matter list for an exact name match, which misses the party who appeared as a witness in a closed file, or as a guarantor, or under a slightly different corporate name.

A real conflict check has to search the full contact and matter history, across every role a party played, including closed matters, because closed matters are where the dangerous hits live. Casely does that by design, and contact labels let you tag roles and referral sources so that the search returns something meaningful rather than a wall of names. The other half of the fix is procedural. Intake needs one door. If leads arrive by phone to reception, by web form to marketing, and by direct email to two partners, then three of those four channels are skipping the check, and the firm's actual conflict exposure is whatever the weakest channel allows. Decide who owns intake, and make the conflict check a gate the matter cannot open without passing.

Reporting Has to Replace Intuition

The founder of a five-person firm knows the financial state of the business without looking, and they are usually right. That accuracy is not a skill, it is a consequence of scale: five people generate few enough data points that a person can hold them. Somewhere past ten, the intuition stays confident and stops being accurate, which is the worst possible combination. The firm feels busy, and busy feels like healthy, and it can take two quarters of declining realisation before anyone notices that the extra work is being done at a worse effective rate than the work it displaced.

The transition here is not about building a dashboard for its own sake. It is about picking three or four numbers that would change a decision if they moved, and putting them in front of the right people on a fixed schedule. Unbilled time sitting in work in progress is usually the first one, because at fifteen people the gap between work performed and work invoiced becomes large enough to constitute a real cash flow problem. Casely's one-click invoicing exists partly to shrink that gap: every unbilled hour on a matter converts into a single itemised draft, which removes the excuse that billing is too tedious to do on time. Hourly, flat-fee, contingency and blended arrangements are all native, and LEDES 1998B export is there for the institutional clients who require it. The point is not the feature. The point is that a monthly billing cycle only works if starting it takes minutes rather than a weekend.

The Middle Layer Nobody Budgets For

Firms scaling from five to fifteen almost always hire producers and almost never hire the middle. They add associates and paralegals because those roles bill, and they defer the operations hire because that role does not. Then somewhere around eleven or twelve people, the founder is spending a third of their week on scheduling, onboarding, software administration, billing chases and vendor decisions, and the firm has effectively hired the world's most expensive office administrator by accident. The economics of that trade are terrible and completely invisible on the income statement, because the cost shows up as founder hours that were never captured anywhere.

The role you need is not a partner and not a receptionist. It is somebody who owns process: who runs intake to completion, who chases the billing cycle, who administers permissions and onboarding, and who is authorised to tell an attorney that a time entry is late. That last part matters more than the job title, because an operations hire with no authority is just an assistant with an inflated description. Give the role a defined scope, a small number of decisions it owns outright, and a direct line to a partner. Most firms who make this hire report that it paid for itself faster than they expected, not because the person generated revenue but because they returned the founder's calendar to work only the founder could do.

Client Experience Has to Be Designed, Not Inherited

At five people, client experience is a personality. Clients are happy because the founder calls them back, because the office manager remembers their daughter's name, and because the file never sits with more than two people. None of that survives contact with a fifteen-person firm unless somebody deliberately rebuilds it as a process. The moment a client's matter is touched by four people instead of two, the experience becomes a function of the weakest handoff rather than the strongest relationship.

The concrete fix is to make the client's view of the matter independent of who is handling it that week. A client portal does this when it is genuinely real-time and privilege-filtered per document, so that what the client sees is always current and never accidentally includes something they should not have. Casely applies that filtering automatically per document, works on mobile because clients read on their phones regardless of what the firm prefers, and keeps e-signature inside the same login rather than bouncing the client to a separate account with a separate password. Every extra login is a place clients drop out, and every dropped client is a chase email somebody has to send. Connected matters, which link related files with the reason stated, matter here too: a client with three related matters should not have to explain the relationship to a new associate every time.

Sequencing the Fixes So They Actually Land

The temptation, once a founder recognises all four failure modes at once, is to fix everything in a quarter. That approach fails reliably, because each of these changes asks people to give up a habit that is currently working well enough for them personally. Attorneys who have kept deadlines in their own calendars for a decade will keep doing it in parallel with the new system, which produces two sources of truth and is worse than either alone. Change one thing, insist on it completely, and only then move to the next.

The sequence that tends to work starts with the matter record, because everything else depends on it. Get status, deadlines, documents and contacts onto the matter itself, so that the firm has one place to look. Then attack the founder bottleneck, because that unlocks the founder's time to do the rest. Then formalise supervision and permissions, which is the highest-risk area and the one you want handled before, not after, the lateral hire arrives with a live conflict. Reporting comes last, not because it matters least, but because reporting on top of unreliable data produces confident wrong answers, which is worse than no answer at all.

What This Actually Buys You

The firms that come through this stretch intact are not the ones with the best lawyers. They are the ones where a partner can be unreachable for a week without a client noticing, where a new hire is productive in their second week rather than their second month, and where the answer to "how are we doing" is a report rather than a feeling. That is the real deliverable of structure. Not control for its own sake, but a firm that keeps working when the people in it are tired, on holiday, or newly arrived.

None of this requires an enterprise platform or a six-figure implementation. It requires deciding that the matter record, not the corridor, is where the firm's memory lives, and then being disciplined about it. Casely is cloud-native with no local install, documents carry AES-256 encryption with a per-firm key, and the Free plan means you can put a real matter through the whole workflow before spending anything. If the growing pain you feel most sharply right now is the founder bottleneck and the loss of a single view of the caseload, matter management is the right place to start, because everything else in this post attaches to it.

Pick the break that is costing you most this month and fix only that one. Write down what the new rule is, tell the whole team in one sentence, and hold the line for a full billing cycle before you touch anything else. Growing from five to fifteen is not a single leap. It is four separate transitions that happen to arrive close together, and treating them separately is the difference between a firm that scales and a firm that just gets busier.