The unbundled retainer: paying for supervision, not progress

Open your SEO contract and find the line that says what you get each month. On a lot of them there is not one. The monthly fee covers "SEO management" or "ongoing optimisation". Content is quoted per piece. Links are quoted per link. Press releases are extra. So everything that could actually change your position on Google sits outside the fee you already pay, and has to be bought again, one approval and one budget conversation at a time. None of that is hidden. You signed it. The problem is what happens next.

This is failure number twelve in my series on why SEO agencies fail, and it is the one owners push back on hardest, because on paper the arrangement looks fair. You only pay for what you use. I have no quarrel with that in principle. The quarrel is with what it does over a year, when the extra-cost conversations quietly stop happening and the retainer settles into paying someone to watch your rankings rather than change them.

A deliverable and a process word are not the same thing

You need to tell these apart, because a scope document is a mix of both and the mix is the whole story. A deliverable is something you could open, read, count or point at a year from now. A process word describes how someone spent their time.

Deliverables look like this:

• A new page published on your website, live at a URL you can open.
• A rewritten page, where the old and new versions can be put side by side.
• A link on a named third-party site, pointing at a named page of yours.
• A press release written, distributed and live somewhere you can read it.
• A specific technical fix - a redirect map switched on, broken links repaired, page speed work with before and after numbers.
• Schema markup added to named pages, or Google Business Profile posts and review responses with dates.

Process words look like this:

• Management, account management, campaign management.
• Monitoring, tracking, ongoing oversight.
• Optimisation, with nothing after it naming what is being optimised.
• Ongoing improvements, continuous refinement.
• Strategy, consultation, support, analysis and reporting.

To be fair, there is real work behind those words. Diagnosis takes judgement and honest reporting takes time. None of it is fraud. But process words share three properties that matter to the buyer. They cannot be verified from outside. They can never be finished, so they can be billed forever. And they are not what moves you up the results page. Google does not rank you higher because someone monitored you attentively.

A deliverable leaves a trace. A process is a promise about how time was spent. You can only audit one of them.

The honest case for unbundling

Unbundling done properly beats a lot of all-inclusive retainers, and the strongest argument is that you only buy what your site actually needs. A business with twenty solid service pages and no authority behind them needs links and nothing else, and it is wasteful to fund a content quota it never asked for because the retainer happens to include four blog posts a month. Bundled packages produce whatever the package says, constraint or not.

The second argument is that cost tracks work: a month where three new location pages go up genuinely costs more to deliver than a month of maintenance. The third is transparency. When a page and a link each carry a stated price, you can see what your money buys. An all-inclusive fee hides that arithmetic - you have no idea whether most of it goes on production or on the account manager.

Done honestly, unbundling looks like a modest management fee alongside an agreed production budget, set in advance for the quarter, with a plan for what it buys and both numbers visible. If that is what you were offered, you were offered something good. Most of the time it is not what arrives.

Why a low headline retainer wins the pitch

What makes unbundling attractive to the seller has nothing to do with dishonesty. It is arithmetic and human nature. When you collect three quotes, the number you compare is the monthly retainer, and that is what decides who gets shortlisted. A low retainer with a rate card underneath beats a single all-inclusive figure, even when the second delivers more, because the comparison happens on the headline and the rate card reads as optional extras.

Illustrative numbers, made up to show the shape rather than real prices. Provider A quotes a flat two thousand four hundred a month, including two new pages and three links every month. Provider B quotes nine hundred and fifty for management, with content at six hundred a page and links at three hundred and fifty each. On the spreadsheet B is nearly fifteen hundred a month cheaper, so B wins.

Now do the real comparison. Buying from B what A includes as standard costs nine hundred and fifty, plus twelve hundred for the pages, plus one thousand and fifty for the links. Three thousand two hundred a month. B is a third more expensive for the same output, visible only if you do arithmetic the quoting process never invites.

But the third outcome is the one that actually happens. You approve one page in month two, feel the pinch, and approve nothing else. B costs eleven thousand four hundred over the year and delivers one page and twelve reports. You did not overpay against the rate card. You underbought against it, which is worse.

Notice what that does to risk. Under A the provider must produce, or they have breached the arrangement and you can point at it. Under B the retainer arrives whether or not a single deliverable is produced: revenue guaranteed, output optional. I do not think most providers set that up as a trap. They set the retainer low to win the pitch, then live inside the incentive it created.

Why the extra-cost conversations never happen

This is the heart of it. The model depends on a repeating conversation where someone proposes work, you approve budget, and the work gets bought. In a good arrangement that happens monthly. In most it happens twice and stops. Five reasons why.

Approval friction. Every deliverable becomes a decision with a price tag. A quote is prepared, an email is sent, you read it between two other things, someone chases you, you say yes. Five steps of friction on top of work you already pay a retainer for, and in a busy business a deferred decision is a no.

Budget fatigue. You paid the retainer eleven days ago. A second invoice in the same month does not feel like buying more, it feels like being charged twice. That is not rational, but it is how it lands, every month.

You do not know which one to say yes to. Handed a menu - a press release at five hundred, three service pages at eighteen hundred, five links at seventeen fifty - nothing in your experience says which of those moves your business. The only person who knows is the one selling. So the safe answer is not this month, twelve times over.

Nobody owns the ask. An account manager who raises budget every month feels like they are upselling a client who keeps saying no. It is uncomfortable and it is easier to stop. So they stop, the account settles at the retainer floor, everyone is comfortable, and nothing gets built.

Drift becomes normal. Once three months pass without production, the arrangement has silently redefined itself. The provider delivers reports. You expect reports. Neither party revisits the point of the engagement. This is a very common reason a site sits still for a year. If your pages have parked just outside the positions that pay - and I have written separately about being stuck at number 3 - the missing ingredient is nearly always authority, and authority is exactly the line item that lives outside the fee and never gets bought.

The ten-minute self-check

You need two things: the scope document and six months of invoices. No tools, no SEO knowledge.

1. Print the scope of work - the schedule attached to the contract, not the proposal deck, which often promises what the contract does not.
2. Take a highlighter and mark every line that is a deliverable: a page, a link, a press release, a fixed piece of work with a countable output.
3. Leave every process word unmarked: management, monitoring, optimisation, reporting, strategy, support.
4. Check the marked lines for quantities. A deliverable with no number - "content creation" rather than "two service pages per month, published" - goes back in the unmarked column, because it commits nobody to anything.
5. Add up six months of invoices, including every extra you approved along the way.
6. Work out what share of that total put something new on your website, or a new link pointing at it. Not what was worked on - what exists now that did not exist six months ago.
7. Open your site and count the pages genuinely new or rewritten. Then ask for every link built in the same period as a plain list of URLs. If it takes two weeks, or comes back as a count rather than URLs, treat it as a zero.
8. Compare the two columns. If the deliverable column is thin and the process column is thick, you have been buying supervision rather than progress.

Reading a scope document line by line

The highlighter pass gives you the shape. A closer read shows where the wording is doing work. Four things I look for.

• Quantities, or their absence. "Ongoing content optimisation" commits to nothing. "Two existing pages rewritten per month" commits to something you can check. Any deliverable without a number is a process word wearing a deliverable's coat.
• Ceilings dressed as floors. "Includes up to four hours of technical work" sounds generous and is a cap - up to four hours is satisfied by ten minutes.
• Whether the rate card is in the contract. If content and links sit outside the fee, their prices belong in writing at signing, not quoted on request. Unpriced extras postpone the budget conversation indefinitely, because every proposal needs a quote before it is even a proposal.
• Who is obliged to initiate. Does the scope say the provider will present a production plan every month? Almost none do. If nobody is on the hook to propose the work, the default state of the account is no work, and defaults win.

That last point is worth raising at quote stage, alongside the other questions in how to choose an SEO provider. Ask who brings the plan, and what happens in a month when nobody does.

How to restructure it so the fee buys output

None of this requires you to leave. Most of the time the fix is one conversation, and a decent provider will meet you on it because it makes their account healthier too.

• Rebundle. Ask for one fee with a monthly production quota written into it: for this amount, two pages published and three links built, every month. Simplest fix, and it moves the delivery risk back where it belongs.
• Or keep the unbundling and fix the buying. Agree a production budget for the whole quarter in one conversation rather than twelve, with a minimum monthly spend on deliverables inside it. You approve the budget, the provider proposes the plan, and no individual page needs its own email chain.
• Change what the report opens with. The first section should be what was produced this month: pages live, links built, with URLs. That makes a quiet month impossible to hide.
• Ask for the split. Of this monthly fee, how much is production and how much is management? A provider who has thought about it answers in seconds. As a rough illustrative benchmark, if management eats more than about a third of total search spend, I would want a very good explanation.

One caution on the menu. When owners do start buying from a rate card, the instinct is to take the cheapest line, and the cheapest line is almost always links. That is the wrong place to economise, because cheap SEO can damage your domain in ways that cost far more to undo than the links cost to buy. Buy the expensive items and buy fewer of them.

The point of all this

Some of the best arrangements I know of are unbundled. A small, honest management fee alongside a real production budget is a clean way to work, and it beats a fat all-inclusive retainer that quietly delivers four generic blog posts nobody reads. The structure itself is not the failure. The failure is the version where nobody is obliged to produce anything, where the low headline number won the pitch, and where the friction of buying is high enough that the buying stops.

It ends with a business twelve months in, thousands of dollars down, holding a folder of reports and a website identical to the one it started with. That is one of the more common routes to having paid for SEO and gone backwards, and it is preventable with a highlighter. So go and get the scope. Mark the deliverables. Count what you actually bought. If the answer is supervision, the fix is not necessarily a new provider - it is a fee tied to output, so that a month where nothing is produced is a month somebody has to explain.