What’s actually wrong with a one-off build, commercially?
Nothing is wrong with the work. What’s wrong is the economics once you zoom out from a single project to a client base. A website build is a transaction: scope it, ship it, invoice it. The relationship’s default state afterward is silence, punctuated by a support ticket, or a full rebuild in three to five years when the platform is finally outgrown. Every new client starts the acquisition cost from zero, because the last one stopped paying the day the build shipped.
A marketing retainer is a subscription to an outcome (rankings, spend efficiency, a growing list of pages that earn their keep), and outcomes have to be maintained, which is the entire commercial difference. The client keeps paying because stopping means the thing they’re paying for stops improving. That isn’t a trick; it’s the honest shape of the work. SEO and paid media are not one-time fixes the way a launched site technically is.
Run the two side by side over three years and the gap is structural, not incremental: the build client generates one invoice and, if you’re lucky, a referral. The retainer client generates thirty-six invoices and a compounding case study you can point the next prospect at. The team, the sales process and the overhead are comparable. The lifetime value isn’t.
What does a retainer actually buy a client that a build doesn’t?
Attention that doesn’t expire at launch. A build’s incentives point toward finishing: every day past the go-live date is a day of unbilled work for whoever built it, so the natural pull is toward closing the ticket. A retainer’s incentives point the other way: the engagement only continues if the number it’s measured against keeps moving, so the natural pull is toward staying in the account and finding the next improvement.
It also buys a team that’s still there when the market changes. A site built two years ago doesn’t know that a Google update just reshuffled its category, or that a competitor started running a campaign that’s eating its branded search. A retainer engagement does, because someone is watching the account weekly, in a workspace you can log into, rather than reading about it as a support request eighteen months later.
Does this mean PixelCrayons is leaving website development behind?
No. It would be a strange thing to abandon, given development is where marketing programmes eventually have to land. The fastest SEO strategy in the world still needs a site that can serve it, and a chatbot needs somewhere to live. What’s changing is the default shape of a new relationship, not the service list. Builds remain the right answer when a client genuinely needs one: a platform migration, a new product line, a rebuild after years of neglect.
What we’re actively steering away from is the build that has no second act by design: a five-page brochure site for a client with no acquisition plan behind it, sold and shipped as a one-time transaction with nothing to build toward. That work is fine in isolation. It’s just the wrong shape for a business trying to grow the value of each relationship rather than the count of them.
What does this change if you’re evaluating us for a website?
Mostly the first conversation. Where a build-only shop asks what pages you want, we’re more likely to ask what the site needs to do six months after launch. If the honest answer is “nothing, it’s a static reference site,” we’ll build exactly that and say so, rather than trying to attach a retainer where none is warranted. Fixed-scope engagements remain a real, first-class option, not a downsell, and when a project is the right shape we say so.
Where it does change something real is sequencing. A build proposal increasingly comes with a plain-language note on what happens after handover: who owns SEO, who owns paid media, whether the answer is genuinely “nobody, and that’s fine.” That note used to be an afterthought. Now it’s part of the scope conversation from day one.
What’s the honest downside of making this bet?
Retainers are harder to sell and slower to fill a pipeline than builds. A build has a clear price and a clear finish line, which makes it an easy yes for a buyer comparing three quotes. A retainer asks someone to commit to a monthly number before the first month’s results exist. That’s a genuinely bigger ask, and one that rightly takes longer to earn.
The other honest cost is concentration risk in the other direction: a retainer client who leaves takes a recurring line with them, not a one-time invoice. That’s the trade we’re choosing: fewer, deeper relationships over a larger number of shallow ones. It’s a bet, not a certainty. The reasoning above is why we think it’s the right one for where the business is going, not a claim that it’s risk-free.
How can you tell, from the outside, whether a proposal is retainer-shaped or build-shaped?
Look at what the proposal measures success against, not what it’s called. The label is the least reliable signal, because plenty of agencies call a retainer a “package” and a build a “programme” for reasons that have nothing to do with the work’s actual shape. A build-shaped proposal measures success against a delivered artefact: pages shipped, features built, a launch date. Its success criteria are satisfiable once, in full, and don’t change after delivery.
A retainer-shaped proposal measures success against a moving number: rankings held or improved, cost per acquisition, a growth curve. Its success criteria are never fully satisfied: there’s always a next month’s version of the same metric to hit. If a proposal’s success criteria could theoretically read “done” and be checked off forever, it’s a build. If they read as a standard to keep clearing indefinitely, it’s a retainer, whatever the cover page calls it.
The other tell is what happens to the team after delivery. A build-shaped engagement’s team composition is front-loaded: most of the people are busiest in the first half, tapering as the deliverable nears completion. A retainer-shaped engagement’s team involvement stays roughly level for as long as the engagement runs, because the work doesn’t taper toward a finish line that doesn’t exist. Ask directly what the team’s time allocation looks like in month six versus month one. A proposal that can’t answer that question specifically probably hasn’t actually decided which shape it is yet.
