The same four things fail, in the same order, in almost every agency and consulting firm that grows past two dozen people. None of them show up on a P&L until the damage is already done.
Somewhere between hire 18 and hire 26, something in a services firm quietly stops working. Not the client work itself, that usually stays fine for a while longer, but the coordination underneath it. We've sat in enough ops reviews at agencies, consulting shops, and IT services firms to see the pattern repeat: it's rarely one dramatic failure. It's four smaller ones, arriving in a predictable sequence, each one masking the next until someone finally asks why margin is down 6 points and nobody can say exactly why.

At 8 people, the founder or managing partner is in every kickoff call, every scoping conversation, every status check. That's not a process, it's just proximity, and it works fine. The trouble starts when the firm hits three or four concurrent engagements with different account leads, because now the one person who used to hold the full picture in their head is stretched across meetings they can't all attend.
We've watched this happen at a 24-person consulting shop where the managing partner still personally approved every statement of work. Fine at 6 active clients. At 14 active clients, SOWs sat unsigned for a week because he was the only reviewer, and two deals nearly walked because of it. The fix wasn't hiring another approver, it was building a system where scope, pricing, and delivery capacity were visible to more than one person at once, so approval didn't depend on one calendar.
This is exactly the gap that professional services automation software is built to close: a shared source of truth for pipeline, projects, time, and billing that doesn't live in one person's inbox or one person's memory.
Second to break, almost always, is the handoff between sales and delivery. At a small firm, the person who sold the deal is often the person who delivers it, so nothing gets lost in translation because there's no translation happening. Past 20 people, those are different humans, and the notes that mattered during the sales conversation (a client's real deadline, a scope boundary the client pushed back on, a discount tied to a specific deliverable) live in an email thread, a Slack DM, or a salesperson's head.
Delivery teams discover this the hard way, usually in week two of a project, when a client says "but we agreed you'd also handle X" and nobody on the delivery side has any record of that conversation. It's not a trust problem. It's a data problem. A handful of things typically go missing in that handoff:
The fix is boring and it works: one CRM record that turns into one project record, carrying the notes forward automatically instead of relying on someone to copy and paste them. That's the whole point of connecting CRM and project delivery in a single flow, which you can see laid out on the features page.
Third break, and this one costs real money: capacity planning turns into a guess. At 12 people, a team lead can eyeball who's slammed and who has room. At 30 people across four project teams, that same eyeball estimate is wrong often enough to matter. We've seen ops leads say "yeah, we can staff that" in a sales call, only to find out the two senior consultants they were counting on were already double-booked on another project's crunch week.
The result is a firm that either overcommits and burns people out delivering, or undercommits and turns down revenue it could have actually handled. Both are expensive. Overcommitting shows up as attrition six months later. Undercommitting just quietly caps growth, and nobody notices because there's no line item for "deals we didn't pursue."
A capacity guess that's wrong 20% of the time isn't a rounding error, it's a staffing decision made on bad information roughly one out of every five times. At 30 billable people, that's not a minor inefficiency. It's the difference between a 68% utilization rate and a 75% one, which on a typical services firm's numbers is a meaningful chunk of annual margin.
This is where time tracking earns its keep, not as a compliance exercise but as a live feed into staffing decisions. When logged hours, project forecasts, and the pipeline sit in the same system, capacity headroom stops being a guess and becomes a number someone can actually look up before saying yes to a deal.
The fourth break is the one that finally shows up in cash flow, and by the time it does, the first three have usually already been happening for months. Billing slows down because finance now has to chase hours across more people, more projects, and more approval layers than it used to. A firm that invoiced within 3 days of month-end at 15 people is suddenly closing the books 10 or 12 days late at 35 people, not because anyone got lazy, but because reconciling time entries, expense receipts, and project sign-offs across a bigger team just takes longer when it's done by hand.
Days sales outstanding creeps up right alongside it, since a late invoice is a late payment waiting to happen. This is often the moment leadership finally starts comparing tools, and it's worth looking at what a connected platform actually costs against what slow billing is quietly costing already; the pricing page breaks that down by team size rather than making you guess.
Book a walkthrough built around your team size and see the handoffs, capacity view, and invoicing flow in action.
It's not a hard rule, but it's roughly where one founder or ops lead can no longer sit in every client meeting and every internal handoff. Below 20, informal coordination works because one or two people hold the context. Past that, you usually have 3-4 concurrent engagements running with different account leads, and the informal system stops covering every gap.
A PM helps, but a PM without shared, real-time data on capacity, scope, and billing status is just a second person guessing instead of one. The fix isn't more headcount watching the problem, it's giving whoever is watching it (PM, ops lead, founder) a system where sales, delivery, and finance already share the same numbers.
Three signs show up reliably: project kickoffs where delivery discovers details the sales rep never wrote down, a founder or ops lead who can't say within a day who's free next week, and invoices that go out later than they used to because someone has to chase down hours first. Any one of these on its own is manageable. Two or more happening at once means the informal system has already run out of room.