The work after launch is the work nobody has priced

You are already doing this

Look at where the hours in your support retainer actually go. In most agencies I have compared notes with, a real share of it is not maintenance in any technical sense. It is somebody noticing that a page is wrong, or that a client's new product has no page, or that pricing on the site disagrees with what sales is quoting.

That is operations work. It requires judgement about the client's business, it is the part the client values most when they notice it, and it is being paid for at the rate of a plugin update because it is invoiced under the same line item.

So the opportunity is not a new service you have to learn. It is a service you are already delivering, unnamed, unpriced, and invisible in the report.

Why the current shape suppresses the price, and the unit that fixes it

Maintenance retainers are priced as availability. The client is buying the assurance that if something breaks, someone will fix it. That is genuinely valuable and it is a low ceiling, because availability is comparable. Any agency can be available, and the client can price shop it in an afternoon.

Operations is not comparable in the same way, because what is being sold is judgement about which change was worth making. A competing agency cannot demonstrate that they would have picked the same page, which means the comparison that usually happens on day rate stops working.

This is the whole commercial argument and it is worth being clear that it is a repackaging argument as much as a new capability argument. The work is largely the same. What changes is the unit, the report, and therefore the conversation at renewal.

One change to make this real: stop reporting tasks and start reporting changes against named pages, each with the reason it was made.

A task says content updates, various. A change says the pricing FAQ now answers the plan limits question, because it came up in four sales calls last quarter, and the expectation is that it stops coming up. The client can evaluate the second one. They have to take the first on trust.

This is the same argument as the retainer evidence piece and I am repeating it here because it is the precondition for everything below. Without it, an operations retainer is a maintenance retainer with better marketing, and the client will work that out by the second renewal.

Availability is comparable, so it gets price shopped. Judgement about which page mattered is not.

A service model that works

Three tiers, each with a clear unit and a clear boundary. The numbers are deliberately absent because they depend on your market, your costs and your client size, and anybody who supplies figures here is guessing about your business.

The important design constraint is that the first tier has to be small enough that a marketing lead can approve it without a committee. Most agencies pitch the largest tier first, meet a procurement process, and lose six weeks. A tier that fits inside somebody's discretionary spend gets you in, and the record you produce is what sells the tier above it.

How to sell the first tier

The entry tier sells itself if you do the work before the conversation, which is uncomfortable and effective.

Take the client's site, spend two hours, and produce the findings list you would produce in the first month. Name the pages that are wrong. Show the two places pricing disagrees. Point at the question their own sales team answers weekly with no page behind it.

That document is the pitch. It is not a proposal describing a service, it is the service, delivered once, for free, at a cost of two hours. Almost nobody says no to the tier above it, because the alternative to saying yes is knowing about a list of problems and having no route to fixing them.

The version of this that fails is a generic audit produced by a tool. Clients have seen those and they know a report generated in ninety seconds when they see one. What works is specificity that could only come from a person looking at their business.

The conversation that gets you from build to operations

The transition point matters, because pitching this cold to a client who has never bought it is much harder than attaching it to a moment that already exists.

The best moment is the handover at the end of a build, and almost nobody uses it. At handover the client is happy, the site is correct, and the risk of it decaying is completely invisible to them because nothing has decayed yet. That is exactly when the argument is cheapest to make and hardest to sell, which is a genuinely awkward combination.

The framing that works is not a warning about decay. It is a question: who is going to be responsible for this being correct in a year, and what happens when your product team renames something. If the honest answer is nobody, that is not a scare tactic, it is a gap you have both just discovered.

The second best moment is about four months later, when the first thing has gone wrong. Somebody has added a page that does not match, or pricing has changed and three pages disagree. The client feels it. A short findings list at that moment lands considerably better than any proposal written in advance.

The worst moment is when they come back asking for a rebuild. By then the decay is complete, the budget is framed as a project, and operations sounds like an upsell on a purchase they already resent needing.

What not to sell

Four things that look like good offers and are traps. I have watched agencies take all four.

Do not sell guaranteed outcomes. Ranking improvements, conversion lifts, traffic targets. You do not control the market, the competitors, or the product, and the moment you attach a number to a result you have made the renewal a referendum on whether that number was hit. It rarely is, and you will be arguing about attribution instead of about work.

Do not sell unlimited requests. It is attractive at signing and it converts a judgement based service into a queue, which is exactly the repositioning you were trying to escape. It also selects for the clients who will use it most, which is not the same as the clients who value it most.

Do not sell a dashboard. The moment the deliverable is a login, the client stops needing the conversation, and the conversation is where the value is legible. Send the record. Let them ask you questions about it.

Do not sell strategy sessions as the whole product. They feel senior and they produce documents rather than changes. Operations is valuable because the site is different afterwards.

The real constraint is on their side

The thing that most often blocks this is not price and it is not appetite. It is that the client has nobody to receive the work.

You need someone at the client who can answer questions about the business, approve a change to a page, and say which of two conflicting statements is correct. In a lot of companies that person does not exist, and the site is a shared responsibility spread across three people who all assume one of the others is watching.

This is worth diagnosing in the first conversation, because it determines whether the engagement can work. Ask who would receive your monthly record and who would act on the two items you cannot decide alone. If nobody can be named, that is your first deliverable rather than a reason to walk away.

In practice you can often be that person for a while, and doing so is a stronger position than being a supplier. But it should be named and priced rather than absorbed, because absorbing it is how an operations retainer quietly becomes a maintenance retainer again.

What to standardise and what to keep bespoke

The margin question. Operations work resists standardisation because the judgement is the product, and agencies that try to templatise the judgement end up delivering generic audits.

Standardise the mechanics: how you review, what the findings list looks like, the change record format, the cadence, the reporting. Those should be identical across every client and they are where your efficiency comes from.

Keep bespoke the part that requires knowing the client: which findings matter, what to do first, which of two contradicting pages is the correct one. That is what you are being paid for and it is not compressible.

The failure mode is inverting this. Agencies build a bespoke process per client, which destroys margin, and then apply generic judgement, which destroys the value.

What this does to how you hire and schedule

There is an operational consequence worth thinking about before selling this, because it changes the shape of your week.

Build work is lumpy and schedulable. Operations work is smooth and interruptible, and the two do not share people well. A developer in the middle of a build who is also the person answering an operations finding will do the finding badly or the build late.

The pattern that works is a fixed block rather than an on demand arrangement. Operations happens on named days, on a cadence, by whoever is on that rotation. Findings that arrive between blocks go on a list unless something is actually broken. This is less responsive than the maintenance retainer it replaces, and clients accept it easily once they see a record of real changes, because they were never buying responsiveness in the first place.

It also changes who is valuable. The best operations person is not your strongest developer. It is whoever is most curious about the client's business, because the judgement about which page matters is the product, and that is a different skill from building the page well.

Where agencies get this wrong

Pitching operations to a client who has just been sold a rebuild by you. It reads as selling the fix for the thing you just delivered. Better to include the first months of operations inside the build proposal, as the thing that protects the investment, and let it continue afterwards.

Naming it something abstract. Website operations is already a slightly abstract phrase. Care plan, growth retainer, partnership tier are worse, because they describe a relationship rather than a deliverable. Describe what happens: we review your site every month, tell you what is wrong, and fix an agreed number of things.

Reporting effort out of habit. The old report is familiar and it is the thing that suppressed the price in the first place. If the change record sits underneath a summary of hours, the client reads the hours.

Waiting for the client to ask. They will not, because they do not have a name for the thing they are missing. The findings list is how you name it for them.

How to evaluate whether it is working

Four signals, over roughly two quarters. None of them is a revenue metric, because revenue is the consequence rather than the evidence.

The last one is the one that tells you the repositioning actually landed rather than just the invoice line changing.

• Can the client name two pages that are better than they were, without checking?
• Does the renewal conversation open with what to do next rather than what was done?
• Has anyone at the client started sending you questions from their sales team unprompted?
• When you propose a change, does the client ask why rather than how much?

Why we care about this

We spent six years delivering Webflow sites, a good share of it white labelled behind other agencies, and the pattern was consistent. The build was the profitable, visible, finite part. The work afterwards was the part clients valued most in the moment and paid least for, because nobody had given it a shape.

Creogen is what we are building for the mechanics of it: findings tied to URLs, changes with reasons, expectations recorded before rather than after. Creobot handles the input side, the questions that should become pages. Both are in private development and open to agency partners first, deliberately, because agencies are already doing this work and are the people who will find out fastest where the tooling is wrong.

None of that is required. The three tier model, the findings list and the change record are a document and a habit. If you run an agency, the cheapest experiment available is two hours on one client's site and one conversation about what you found.