The Work Nobody Posts About

Key takeaways:

• The difference between agencies that deliver reliably and ones that operate on heroics is almost always invisible infrastructure: databases, templates, and systems that clients never see and rarely think to ask about.
• Investing time in internal tooling — client intelligence, standardised design components, deployment processes — compounds over years. The payoff is consistency, not speed on any single project.
• Most agencies skip this work because it doesn’t win clients, doesn’t appear in proposals, and doesn’t photograph well. That’s exactly why it separates the ones who last from the ones who don’t.

What separates agencies that last from ones that look like they will

I spent the break between Christmas and New Year doing work that will never appear in a case study. No client will reference it in a testimonial. It won’t show up in a proposal or impress anyone on a discovery call. It was internal systems work, and it was some of the most valuable time I’ve spent in years.

That probably sounds backwards. The instinct in most agencies is to fill any downtime with sales activity, content, or at minimum a planning session with optimistic revenue targets written on a whiteboard. I’ve done all of that. But after 25 years of watching agencies operate, including building and running Chillybin since 2009, I’ve come to a fairly firm conclusion: the agencies that deliver consistently are not the ones with the best pitch decks. They’re the ones who did the boring infrastructure work that nobody posts about.

What does “internal infrastructure” actually mean for a web agency?

It means the systems and tooling your team uses every day that clients never interact with directly but absolutely feel the effects of.

For us, that breaks into three areas right now. First, a proper client intelligence database. We’re rolling out a setup with 120+ data points per maintenance client, connected to ClickUp, ActiveCampaign, and our ticketing system. Before this, too much knowledge lived in email threads or in someone’s head. Hosting credentials buried in a chain from 2021. Notes about a client’s preferences scattered across Slack. That’s not a systems problem unique to us — every agency I’ve spoken to has some version of it. The fix isn’t complicated, but it requires someone to actually sit down and build it properly, which means not billing client hours while doing it.

Second, updated design templates in Figma. Standardised component library, consistent naming conventions, structured so new projects can start from a real foundation rather than from last quarter’s client work with the colours changed. Back in 2015 we were doing a version of this with Photoshop documents that nobody could find and even fewer people could open without something going wrong. The tools have improved dramatically. The discipline to actually build and maintain a proper component library — that requires intention, not tools.

Third, Beaver Builder 2.10. We deployed the update across client sites starting in October last year. The Components feature specifically is saving meaningful time per build. When I say meaningful I mean hours, not minutes, across a project. That’s real margin recovered, or real time returned to quality work that used to get squeezed at the end of a build.

Why don’t more agencies do this?

Because it doesn’t win clients and it doesn’t get celebrated.

Nobody is going to post on LinkedIn about the fact that they spent three days rebuilding their internal database schema. No client is going to reference it in a referral. There’s no award category for “most improved ticketing system integration.” The incentives in agency life all point toward client-facing work, visible deliverables, and things that can be put in front of a prospect.

I understand the pressure. In 2009 when I started Chillybin, every hour that wasn’t billable felt like a loss. That framing is almost universal in the early years of running any service business. But it’s the wrong frame, and the longer you hold it the more it costs you.

The agencies I’ve watched fall apart — and I’ve watched a lot of them over the years — don’t usually collapse because of a bad client or a lost pitch. They collapse because the internal scaffolding was never built. Everything depended on one person knowing where things were. Every project required heroics to deliver. The margin got eaten by rework and hunting and miscommunication that a proper system would have prevented.

One client we’ve worked with for several years came to us after leaving another agency. The handover was a disaster. No documentation. Hosting credentials locked to a personal email address. No record of what plugins had been installed or why. The previous agency wasn’t negligent — they were just operating without the boring infrastructure that makes clean handovers possible. That’s not a horror story unique to small operators. I’ve seen it happen at agencies with 30 people and decent revenue. Scale doesn’t fix a systems problem. It amplifies it.

What does it actually cost to skip this work?

The cost is almost never visible until it’s expensive.

A disorganised client database doesn’t cause a crisis on a Tuesday in March. It causes a two-hour delay on a Friday afternoon when someone’s site is down and you’re hunting through email for the hosting login while the client is watching. A missing Figma component library doesn’t fail a project — it just means every project starts slower than it should and the design consistency that clients pay for slowly degrades across a portfolio of work.

This is the insidious thing about infrastructure debt: it doesn’t look like a problem until the cost is real. In services businesses especially, the default is to absorb those costs invisibly — through longer hours, through stress, through senior people doing junior work because the system doesn’t support delegation properly.

At Chillybin we track things like build time per project type and ticket resolution rates across our maintenance clients. When those numbers move in the wrong direction, the cause is almost always a systems gap rather than a people gap. Someone isn’t slow. The process is missing or broken. The distinction matters because fixing people is hard and fixing process is usually just unglamorous work that someone has to prioritise.

Is this only relevant to agencies?

No, but agencies feel it more acutely because the work is relentless and the margin for error is narrow.

Any business that delivers services on repeat — whether that’s web development, accounting, legal work, or consulting — accumulates the same kind of invisible debt. The consultancy that keeps everything in the founder’s head. The law firm that stores matter notes in email folders. The accountant whose client onboarding process is a mental checklist rather than a documented workflow.

The pattern is identical. The business works because the people are good. And then growth, or turnover, or illness, or just volume reveals that the business was actually running on individual heroics with no supporting structure.

I first noticed this clearly in myself back around 2012. Chillybin had grown enough that I couldn’t personally oversee every project. The gaps in our systems became immediately and painfully obvious. Not because anything catastrophic happened, but because things that had worked when I was close to everything stopped working when I wasn’t. That period of rebuilding our internal processes was genuinely uncomfortable. It required admitting that what had worked wasn’t scalable and then doing the work to fix it while still running the business day-to-day.

Most people postpone that reckoning until they’re forced into it. Doing it proactively, when you have the time and the margin to do it properly, is almost always better. Downtime between projects, quiet periods in the calendar, the gap between Christmas and New Year — that’s when the infrastructure work actually gets done properly if it’s ever going to.

How do you know where to start?

Start with whatever caused the most pain in the last six months.

This is the only heuristic I use now. There’s no shortage of systems that could be improved at any given time. If you try to audit everything you end up in a planning loop. But if you ask the team where the friction is — where things slow down, where information gets lost, where someone has to ask the same question twice — you get a specific, honest list.

For us, the client intelligence database came directly from a maintenance review where we realised the information we needed to do good proactive work for clients was scattered across four systems and one person’s memory. The Figma component library came from a retrospective where two designers independently rebuilt similar components for two separate projects in the same month. The Beaver Builder update came from tracking build time data and seeing the efficiency gain clearly enough to prioritise deployment across the portfolio.

None of these were strategic insights. They were responses to observed friction. The strategy is just: fix the friction before it becomes structural.

The agencies and businesses that operate with the most reliability aren’t necessarily the ones with the best talent or the highest-profile clients. They’re the ones that did the unsexy systems work early enough that it stopped being a problem. The infrastructure becomes invisible because it works. That’s the point.

Most of this work will never show up anywhere a client can see. It won’t be mentioned in a sales conversation. The clients who benefit from it won’t know they’re benefiting from it. That’s fine. That’s what good infrastructure is supposed to do.