When to hire a web design agency for a project, and when to bring one in long term

Most agency relationships fall into one of two shapes. You hire someone for a specific build, they deliver it, you go your separate ways. Or you bring someone in on an ongoing basis to keep improving things over time.

Both are valid, and both are useful. They solve different problems, and a lot of businesses end up paying twice because they picked the wrong shape for where they actually were.

What each model actually is

A project is finite, with a brief, a scope, a timeline, a price, and a launch date. Once the site is live and the snag list is closed, the engagement ends. Anything you want to do after that is a new conversation.

A growth partner relationship works more like an ongoing seat at the table. There's still a defined scope each month, but it's framed around continuous improvement of the site rather than building a new one. Design, development, content, performance, and sometimes strategy all roll into a monthly retainer with an agreed number of hours or output.

Which one fits depends less on the size of your business and more on what you actually need the website to do for you over the next 12 months.

When a one-off project is the right shape

Project work is the right call when you're starting from a clear gap. You haven't got a website at all, or the one you've got is so far from what you need that incremental change won't get you there.

It's also right when the business is in a stable patch. If your offer isn't shifting much, your audience isn't shifting much, and you mostly need a site that holds the line for two or three years, a project gets you there cleanly.

The third case is budget timing. Some businesses can release £15k for a defined project a lot more easily than they can commit to £1.5k a month for ten months, even though the maths is similar. That's a real constraint, and phased project work lets you spread the build across stages without triggering an ongoing commitment.

When ongoing partnership starts to make sense

A growth partner relationship earns its place when the website is doing real commercial work. If it's central to lead generation, paid campaigns, recruitment, or sales enablement, it stops being a finished object and becomes something that needs to evolve.

It also makes sense when you've got more ideas than your in-house team can ship. A lot of marketing managers we work with have a backlog of "we should really do that" items that never gets touched, because there's always something more urgent. A retainer turns that backlog into a roadmap.

The third case is when you've recently invested in a new site and want to actually capitalise on it. The first six months after launch are when most of the conversion improvements get found, and they almost never get made if there's no one whose job it is to make them.

What you actually pay for in a retainer

The honest answer is hours and headspace. You're paying for a team that holds your project in mind between sessions, that already knows your codebase, your brand, your sector, and your goals, and that can move quickly because none of that needs re-explaining each time.

Most growth partner arrangements include a mix of design and development time, regular check-ins, analytics review, performance and SEO improvements, and small build work as the backlog dictates. Some include content production days for photo or video. Some include strategic input at a roadmap level.

The thing they all share is continuity. The same team picks up the same site month after month, which compounds in a way that one-off projects never can.

The cost of project-only relationships

There's a hidden cost to working only on a project basis, and it usually shows up about eight months after launch. The site is starting to drift. Plugins are out of date, forms are quietly broken on Safari, the blog has three posts in it, and the thing that was meant to drive enquiries is doing 60 percent of what it could be.

At that point you've got two options. Either you pay an agency to run a discovery exercise on your own site to remember what's there, then pay them again to fix it. Or you live with it until the next rebuild conversation, and lose months of compounding gains in the meantime.

Neither is fatal. Both are more expensive than just having someone on retainer doing ten hours a month of attention from the start.

The pattern we see most often

The shape of this conversation repeats itself across most of the projects we take on. A business commissions a new site as a one-off, the launch goes well, and for the first couple of months everything feels finished.

Then small things start surfacing. A form drop-off nobody predicted. A service page that's getting traffic but no enquiries. A piece of new messaging the marketing team wants to test. Each one is too small to scope as its own project, but collectively they're the difference between a site that holds steady and one that compounds.

The teams that handle this well have someone whose job it is to ship those small changes continuously. Sometimes that's an in-house developer, sometimes it's us on a retainer, occasionally it's a freelancer kept on a regular schedule. The model matters less than the rhythm. Small frequent improvements from someone who already knows the site tend to outperform big infrequent rebuilds, and they cost less in total over a two-year window.

The businesses that struggle are usually the ones where nobody owns the site after launch. The backlog grows, momentum fades, and 18 months later the conversation turns to "we probably need a new website" when what they actually needed was ten hours a month of attention from someone who cared.

How to tell which you need right now

Ask yourself two practical questions. First, when the new site goes live, who is responsible for improving it? If the answer is "we'll figure that out after launch," you probably want a growth partner conversation built into the project from the start.

Second, how central is the website to how the business actually grows? If it's a glorified business card, a project is fine. If it's the engine, treat it like one.

If you're not sure which side of that line you sit on, our growth partner page lays out how the model works in practice, and our web design page covers the project side. Our recent post on how much a website actually costs in 2026 goes into the budget side of the same question.