Building My Own Time Tracking Platform, and what it taught me about Agency Profitability
- Operations
Most agencies have a time tracking problem. Not in a sense that nobody is tracking time; most agencies have something in place. The problem is deeper, the data exists but nobody is really using it. Hours get logged after the fact, budgets overrun before anyone notices. A project that was supopost to be profitable turns out not ot be, and by the time that becomes clear that additional work has been done, the invoice has already gone out.
I’ve seen this pattern enough times to know it was a systems problem and not a people problem. Before I joined HC Media Group as COO I decided to build somethign that actually solves this universal problem.
Why existing tools weren’t enough
The busiess was already using Monday.com, and its a solid tool for task management. Theres no compelling reasomn to replace the system, but Monday.com isn’t build for the specific kind of visibility a digital agency needs, which is real-time budget burn against a project, retaner hours consumed this month or year, a clear view of wheather a client is actually profitable or whether the team has been quietly over-servicing clients for years.
The tools that do offer this visibility tend to be either too generic, too expensive for a small agency or so complex that adoption becomes its own project. I’d evaluated and used enough of these systems to know that the gap between what the promise and what the team would acutally use was too wide.
So, I built Pathfinder Analytics. Not because I enjoy unnecessary complexity, but because the problems are too specific that nothing off the shelf was going to solve it cleanly.
What it needed to do
The core requirements were straightforward in principle.
Every team member needed to be able to log time against a project, and a specific service within that project. Development, design, content, project management, seo.. without it feeling like a difficault task. If logging time is harder than doing the work the adoption rate would be minimal and the idea would fail. The interface had to be fast, simple and frinctionless enough to become a natural part of the working day rather than a 5pm afterthought.
Project managers needed a scheduling view. A picture of capacity with available hours per week, commited hours, utilisation warnings when someone approaches their limit. The goal was to make over-commitment visible before it became a delivery problem.
Directors and operations needed financial visibility. Every project carries a budget, and every service within that project carriers an allocation. As hours are logged, the system needed to show real time percentages of each budget consumed, whats remaining and whether the trajectory suggests the project will land within or outside of budget.
The alert layer was also implemented. Knowing a project is at 80% budget with 40% of the work remaining is only useful if someone is told about itat the right moment. I built threshold alerts so that project managers are notified automatically when a budget hits a defined percentange creating an opportunity to intervene early in the project.
Building It
Pathfinder is built in Laravel, Intertia.js, React and typescript. I built the system specifically around my experience rather than as a generic product, which ment every design decision could be made in service to actual problems rather than hypothetical ones.
I tested it at scale before deployment. 2,000 clients, 32 million time entries on a single account, stress testing the maths and performance under conditions that significantly exceed what the business would generate.
The development process was honest about what it was. I used AI tools heavily throughout the build, GitHib Copilot, GPT, Claude as a development accelerator rather than a replacement for my own judgement. The AI wrote a lot of the scaffholding; I made the arcitectural decisions, built the data model, and tested everything against real scenarios. A system that handles financial data fora. business needs to be correct, not just functional.
What it taught me about profitability
One of the most surprising thing about building a time trackign system (and i’d built multiple up to this point) isn’t the technology. It’s what the data reveals once people actually start using it.
Within weeks of Pathfinder going live, patterns emerged that had been invisible before. Certain service types were conistently over-running budgets, others where being quoted at a rate that made them unprofitable regardless of how well the team performed. Retainer clients were being serviced at levels that had a disconnect as to what the client was actually paying.
This wasn’t the fault of anyone, but without visibility, nobody knew. The team was working hard and doing good work. The business was just leaking money through gaps that nobody could see because nobody had the tools or systems to make that visible.
This is the commercial argument for operational infrastructure that most people miss… it’s not about control or accountability but rather giving the people running the business the information they need to make good decisions. Pricing decisions, resourcing decisions and client decisions.
Whehn you can see what the cost of services actually is, vs the time it takes to deliver, you can start making smarter decisions on the reality of the situation rather than intuition or guessing.
The adoption problem
Building the system was the straightforward part. Getting people to use it was harder.
Timetracking carriers cultural baggage in most agencies, and can feel like surveillance and mistrust. People who have never had to log their hours before will find reasons not to.. especially when the system is new and the habit isn’t established.
I handled this by being direct about its purpose. Pathfinder isn’t there to monitor individuals – its there to give the business the data it needs to price projects accurately, protect the team from being overloaded and to intervene when something is wrong. Timetracking is built to benefit everyone, not just directors.
I also gave people a runway, two weeks of operational adoption before the mandatory date. No pressure and no conseuquences during that period; it was about allowing employees to get comforable with the tool and understand how it worked. By the time it was a requirement, everyone already knew how to use it.
Profitability as a leadership responsibility
Theres a version of this where Pathfinder Analytics is just a piece of software and a technical solution to an operational problem…
What i learned from building and deploying Pathfinder is that profitability isn’t a finance problem. It’s a leadership problem. The data that tells you whether your business is making money on its work exists somewhere in almost every agency and the larger question is whether the people making the decisions can see it, and if those systems alert or inform at the right moment.
A business that doesn’t know what its work actually costs to deliver isn’t in control of its future and is just reacting when taking on work, hoping it comes in under budget.
Pathfinder was my answer to a universal problem, but the principle allies everywhere. Build visibility, make data accessible and give people the information they need to make good decisions.