What an SEO Retainer Should Actually Buy You, Month by Month

An SEO retainer is a standing monthly fee that reserves a defined, continuing block of search work, and the thing a proposal has to define, beyond the deliverable counts, is the order that work happens in. This is written for a founder, marketing lead or growth head holding a proposal or facing a renewal, who can read the deliverable counts but cannot tell whether the sequence behind them is sensible. It covers what has to be in a scope before two retainers can be compared at all, what each of the first six months should produce, what should not be happening yet in each of them, and what you take with you if you stop. It does not cover the fee itself, because what SEO actually costs is a separate question with its own published data behind it.

The short version

• Month one should look like almost nothing shipping. A proposal that promises published articles in week two is describing a worse engagement, not a faster one.
• A retainer buys a sequence, not a quantity of hours. Two proposals at the same fee with the same deliverable counts can still be years apart in quality, and the difference sits in the order.
• Five items have to be in the scope or the comparison is not possible. Missing any one of them turns the proposal into a list of activities.
• A retainer is the wrong instrument whenever you can name the day the work is finished.
• What you keep if you stop belongs in the scope, not in a legal appendix nobody reads until the relationship is already over.

What is an SEO retainer, and what is it not?

A standing monthly fee that reserves a defined, continuing workload, and the only common pricing model built on the assumption that the work has no natural end. That assumption is the whole argument for it. Search results move because competitors publish, because your product changes, and because Google changes, so a site that was correct in March is slightly wrong by September and nobody has done anything wrong.

It is not a maintenance contract, which keeps a thing in the state it was handed over in, and it is not a block of hours, which prices effort and leaves output undefined, and the retainer is only the first of the four ways SEO services are packaged and sold. The useful definition is narrower than either: a retainer is a standing commitment to a sequence of work whose next step depends on what the last step returned. If the scope you are shown would read identically in month one and month six, it is not describing a retainer. It is describing a subscription.

What must be in the scope before you can compare two retainers?

Five items, and the test for each one is whether you could hold someone to it in month four, when the enthusiasm of the kickoff call has worn off and nobody remembers what was said out loud. Each has a weak version that passes unchallenged because it sounds like an answer.

What the scope must name	The weak version to watch for	The question that settles it
A named strategy owner	"Your account will be managed by a dedicated team."	Who decides what gets worked on next month, and will that person be on the monthly call?
Monthly deliverables with counts	"Ongoing content optimization and technical support."	How many of what ships in a normal month, and what happens to the count in a month with a launch or a holiday?
Technical implementation or an explicit handoff	"We will work closely with your development team."	Do you deploy changes or do we? If we hand them over, in what format does a ticket arrive and who writes the acceptance criteria?
Reporting against agreed indicators	"Monthly reporting and analytics."	Which indicators are we judged on, and are they fixed now or chosen later, once we can all see which ones happen to look good?
A named review date	"Rolling monthly, cancel any time."	On what date do both sides look at the evidence together, and what is allowed to change in the scope on that date?

Cancel any time is not a review point. It is the absence of one, dressed as flexibility. A cancellation right tells you how to leave; a review point forces a conversation about whether you should, and the second is worth considerably more than the first.

One thing in a proposal cancels the other five. Google's documentation on whether you need an SEO names three pitches to walk away from: a guaranteed ranking, a special relationship with Google, and an advertised priority submit. A scope that promises positions has stopped describing work and started describing an outcome it cannot control, and no amount of detail elsewhere in the document repairs that.

What should each month actually produce?

Different things, and the fact that they differ is the point. The ladder below is the scope shape worth arguing for in a proposal, month by month. Read the third column as carefully as the second, because a month running ahead of it is usually a sign that something else was skipped, and worth asking about.

Month	What the provider should be doing	What you should be able to see	What should not be happening yet
1	Getting access, crawling the site, inventorying which pages already earn impressions and for what, agreeing the target set, writing the roadmap.	A verified access list, a crawl with issues ranked by severity, a written roadmap naming the pages the work will touch, and the indicators the engagement will be judged on.	Published articles. Outreach. Any statement about rankings.
2	Working the top of the roadmap: the highest-severity technical issues, and on-page work on pages that already have impressions and no clicks.	A shipped or handed-over fix list with status per item, the first content briefs, and a baseline report you can compare later months against.	Full content cadence. A second strategy document. Anything presented as a result.
3	Running the cadence: publishing at the agreed count, clearing the technical queue, wiring internal links into the pages shipped in month two.	New pages indexed rather than merely published, and movement in impressions on non-brand queries for the pages touched earliest.	Clicks at volume. Pipeline attribution. A scope change requested without evidence.
4	Reading what months two and three returned, then dropping what is not working and doubling the thing that is.	A written decision to stop doing something, with the reason. Average position improving on the earliest pages.	The program being judged on revenue. Every original bet still running untouched.
5	Depth rather than breadth: finishing clusters that are half built, reworking pages sitting just off page one, starting authority work if it is in scope.	Position gains concentrated on the tracked set, and the first clicks arriving from queries with buying intent rather than from definition lookups.	A scope that still looks identical to month one.
6	Reviewing honestly: what compounded, what did not, what the next two quarters should look like and why.	A written review against the indicators agreed in month one, and a proposed scope for month seven onward that differs from the one you signed.	A renewal decision taken without that document in front of you.

When a month does not look like its row, there is a specific question that is more useful than asking what happened.

• Month 1: Which pages did the inventory say already earn impressions, and which of them are in the roadmap?
• Month 2: Which fixes are blocked on us rather than on you, and what is the blocker?
• Month 3: Of what we published, how much is indexed, and what are we doing about the rest?
• Month 4: What have you stopped doing since we started, and why?
• Month 5: Which pages are closest to breaking through, and what is the work left on them?
• Month 6: Which of the indicators we agreed in month one did we miss, and what does that change?

Two limits on reading the ladder. It describes a site that is broadly intact at the start, so a replatform, a domain migration or a site with no analytics history pushes every row later and that is legitimate rather than a failure. And it describes the work, not the results, which arrive on a different and slower clock than the scope does.

Why does month one look like nothing is happening?

Because month one is access, inventory and a decision, and none of those three produce a page. The work is real and almost none of it is visible from outside, which is an uncomfortable position for a provider who has just been paid.

Access alone is more involved than it sounds. Search Console has to be verified rather than merely shared, and Google's documentation on verifying site ownership sets out the methods available, from an HTML file or tag to a DNS record at your domain provider. Which method was used determines how easy the access is to hand back later, which is why it belongs in month one rather than in a scramble at the end.

Then the inventory. A crawl tells you what exists; the inventory tells you which of those pages already earn impressions, for which queries, and at what position. That is the document the roadmap is derived from, and a provider who skips it is choosing targets from intuition and calling it strategy.

From my experience working on retained engagements, month one is the row that gets renegotiated most often, and almost always in the wrong direction. A buyer who has just signed wants something to show internally, so a couple of articles get pulled forward to fill the month, written before the inventory is finished and aimed at targets that later turn out to be wrong. Those articles occupy the cadence for a quarter, and they are the pages nobody wants to admit should be rewritten.

Discovery is also a word a slow provider can hide behind, so the balancing test is artifacts. A real month one ends with things you can open and read: a ranked issue list, an inventory, a roadmap that names pages, and a set of indicators. A thin one ends with a call and a deck. For the separate question of when results should appear rather than when work should, our guide to how long results actually take sets out the stages and the signal that proves each one finished.

How should the mix of technical, content and link work change over six months?

It should tilt, and the direction of the tilt is from remediation toward compounding. Early on, the cheapest wins are corrections: pages that cannot be crawled properly, titles that describe the wrong query, internal links that never got wired, a template issue repeated across a hundred URLs. That work has a floor. Once it is done it is done, and a provider still billing for it is billing for a queue they have exhausted.

As the corrections run out, the weight should move to net-new work: pages that did not exist, clusters that were half built, and whatever authority work the scope includes. By month six, a scope that has not shifted at all is the clearest available signal that nobody is reading the data.

Proportions are deliberately absent here. Any split stated as a percentage of a budget is a guess about a site the person quoting it has not seen, and the right split for a clean site with no content is the reverse of the right split for a large library on a broken template. Ask for the shape instead: what share of next quarter fixes things that exist, and what share builds things that do not?

When is a retainer the wrong instrument?

Whenever you can name the day the work is finished. That single test resolves almost every case, and when it resolves against the retainer, that is not a conclusion a provider has any incentive to reach for you.

The work	Why it has an end	What to buy instead
A one-off audit	It is complete when the findings are delivered and explained. Nothing accumulates after that.	A fixed-price project with a defined deliverable and a walkthrough session.
A replatform or migration	It finishes at cutover. The risk is concentrated in a window of weeks, not spread across a year.	A project fee scoped to the cutover, plus a short post-launch monitoring window stated in weeks.
A redirect map	It is a finite mapping of a finite set of URLs. The count is knowable in advance.	A project fee priced against the URL count, with the QA pass included rather than billed after.
A fixed batch of content	Twelve articles is twelve articles. Spreading the invoice across twelve months does not make the work ongoing.	A per-deliverable or per-batch fee, with the brief quality specified as tightly as the word count.

The reason this matters commercially is that a monthly fee applied to finite work keeps charging past delivery, and the only two ways out of that are to stop billing or to invent filler. Filler is how a good engagement turns into a bad one without anybody deciding to make it worse.

What should you be able to see every month?

Evidence you could check yourself, which is a higher bar than a report. Three things belong in the scope, and they are separable: the underlying data, the interpretation, and the record of what changed.

The data: you should hold owner-level access to the properties the work is measured in, not a screenshot of them. Search Console is free, it belongs to you rather than to the provider, and a buyer who can open it is a buyer who cannot be told a comfortable story about impressions.

The interpretation: the monthly report, whose job is to say what changed, why, and what happens next. Its structure is a topic in itself, and our guide to what a monthly report should contain covers the format and the metrics worth including.

The record: a change log. What was shipped, on what date, to which URL. Without it, an engagement cannot answer the simplest diagnostic question in month five: what did we change just before the thing that moved?

When we scope a retainer, the change log is the line we argue hardest for. An engagement that cannot reconstruct its own decisions cannot learn from them, and a provider who cannot show you what they shipped is asking to be graded on effort.

What do you keep if you stop?

Five things, and the scope should list them long before the contract does. This is the part of a retainer that is invisible while everything is going well and is the only part that matters on the day it ends.

• The content: every published page, plus the briefs, drafts and research behind them. Ask for the briefs explicitly, because they are what lets somebody else pick the work up.
• The access: owner-level control of Search Console, Analytics, the CMS and any tool licensed on your behalf, held by you throughout rather than transferred at the end.
• The documentation: the audit, the inventory, the roadmap, and the reasoning for the decisions taken. Without it, the next provider starts by paying for month one again.
• The data: historical exports from the properties, so a later change of tooling does not quietly erase the baseline you were measured against.
• The roadmap: what was planned next and why, in a form somebody else can act on.

The scope names these; the agreement enforces them, and the two documents are not interchangeable. Ownership language, notice periods and what happens to accounts at the end are a genuinely separate layer, and our guide to the clauses worth arguing about works through them properly.

When should you not buy a retainer at all?

Two situations, and neither of them is about budget. Both are about whether a standing monthly commitment is the thing that unblocks you.

You can already execute and you lack direction: if you have writers, developers and a marketer who can run a queue, buying a retainer duplicates capability you already pay for. What is missing is the decision layer, which is bought as advisory rather than as delivery. That is the case direction for teams that can already execute is built for, because what you are buying is judgment rather than hands.

You cannot absorb the output: a retainer produces a queue. Technical fixes need deploying, drafts need subject-matter review, and recommendations need somebody internal to own them. If nobody can spend an hour a week on that, the output accumulates unimplemented and the fee buys a growing document. Fix the absorption problem first, or buy implementation as part of the scope rather than assuming your team will handle it.

Neither case is permanent, and neither means SEO is wrong for you. They mean the retainer is the wrong shape for this quarter.

The habit worth building is narrow: compare sequences, not deliverable counts. Counts are easy to match and easy to inflate, and two proposals offering the same article count and the same number of fixes can differ enormously in whether anybody has thought about what happens once the obvious work runs out. This week, ask your current or prospective provider two questions and write the answers down: what did month one produce, and what does month four look like? Anyone who answers the second one crisply has planned past the honeymoon.

Treating an engagement as an ordered sequence rather than a monthly quota is the whole idea behind SEO built as an acquisition engine, and it is the cheapest thing to check in a proposal that is already on your desk.