How to Evaluate an SEO Agency Proposal
The proposal is the one document you can study at your own pace before committing. Almost nobody knows what to look for in one. Most proposals do their persuading through presentation, so this page is about reading past it.
The Proposal Is The Sample Of The Work
How carefully somebody proposes tells you how carefully they will work. That is the most useful thing to understand about the document in front of you. It changes what you are looking for.
Why it predicts so reliably. Same people, same habits.
The proposal is produced by the firm you would be hiring, under no time pressure, when they are trying hardest. If this is their best effort, the monthly work will not exceed it.
What a generic proposal is. A preview.
A document that could be sent to any business in your trade tells you that your account will be handled the same way. Nothing about the engagement will suddenly become bespoke after signature.
What you are not assessing. Whether you like it.
Presentation quality reflects a firm's design resources rather than its search work. A plain document describing more work beats a handsome one describing less.
The advantage this document gives you. Time.
A conversation is live and hard to assess under pressure. A proposal sits on a desk and can be read twice, which makes it the best opportunity a non-specialist gets.
Does It Describe Your Business
A proposal should demonstrate that somebody looked at your site and understood what you do. Most contain a great deal about the agency and very little about you.
What evidence of looking is. Specifics only they could know.
Named pages, actual problems on your site, the particular searches your customers use, competitors named correctly. Anything that required opening your website rather than reading your industry.
What boilerplate looks like. True of anybody.
Statements about the importance of search, general descriptions of your sector and observations that would apply to every business doing what you do. Remove your name and it still reads perfectly.
The test to apply. Substitute a competitor.
Read it imagining it was sent to your nearest competitor. If it would work unchanged, nobody examined your situation and the plan inside it was not made for you.
Why this matters beyond effort. Aim.
A proposal written without looking cannot be targeted at anything specific. The work that follows will be aimed at your trade in general rather than at your business.
The fair qualification. Depth costs time.
A firm cannot conduct a full audit before being paid. What is reasonable is evidence of an hour's genuine attention, which is a low bar that a surprising number of proposals miss.
Work Or Outcomes
A proposal should commit to work and describe expected outcomes as expectations. One that promises outcomes is either naive about the trade or misleading about it.
Why the distinction is not pedantic. One is deliverable.
A firm controls what it does. It does not control where a search engine places you or what competitors do. Committing to the first is a real promise. Committing to the second is committing to somebody else's decisions.
What committing to work looks like. Checkable.
Stated activity, in quantities or hours, that you could verify happened. That is a promise with consequences attached if it is broken.
What promising outcomes looks like. Unfalsifiable until too late.
Positions, traffic figures or enquiry volumes stated as commitments. By the time you can check, the money has been spent and the answer is always that these things take time.
How expectations should be phrased. As expectations.
A proposal can and should say what the firm anticipates, provided it is presented as an expectation with reasoning rather than as a commitment. That is useful and it is a different sentence.
Where this appears again. Two places.
The contract page applies it to agreements. The red flags page treats a guaranteed position as disqualifying outright.
Can You Tell What You Are Actually Getting
Deliverables should be specific enough that you could confirm afterwards whether they happened. Vagueness here is how underdelivery becomes deniable later.
The test. Could you audit it in twelve months.
Read each item and ask whether you could establish, a year from now, whether it was done. If the answer is no, the item commits to nothing.
Phrases that fail the test. Unbounded effort.
Ongoing optimisation, continuous improvement, regular attention and similar formulations. Each describes a direction rather than an amount, so any quantity of activity satisfies them.
Why they are used. Not always cynically.
Some work genuinely does not divide into countable units, so a firm may be avoiding false precision rather than hiding. The remedy is the same either way: ask what it means in practice this month.
What good specification looks like. Amounts or time.
Either stated quantities of work, else a stated allocation of hours with an explanation of how it will be directed. Both are checkable and the second is frequently more truthful.
The item most often missing. Thinking time.
Reviewing, deciding and adjusting rarely appears as a deliverable because it produces nothing visible. It is also the most valuable part of the arrangement.
Is The Timeline Realistic
You are not checking dates against a standard, since none exists. You are checking whether the shape of what is promised matches how this work actually behaves.
What a realistic shape looks like. Slow then steeper.
Early periods heavy on research and planning with little visible output, then production, then gradual movement that accelerates. A proposal describing that shape has been written by somebody being accurate.
What an unrealistic shape looks like. Immediate results.
Meaningful improvement promised early, else a straight line from signature to success. Either the firm misunderstands the work or is describing something other than what they intend to do.
The other warning. No shape at all.
A proposal with no sense of sequence, where every activity appears to happen continuously from day one. That usually means the plan is a list of services rather than an order of work.
What we will not give you. A period to check against.
No page on this site states when results arrive, because nothing supports a figure and any number would be invented. Judge the shape and the reasoning instead.
What should be dated. The work rather than the outcome.
When the audit finishes, when content starts appearing, when you will receive reporting. Those are within the firm's control and can reasonably be committed to.
What Is Excluded
Exclusions are where the real scope lives. A proposal listing none has not stated them rather than having none. Everybody discovers the omissions eventually.
Why the exclusions define the scope. Inclusions are aspirational.
What is listed as included tells you what the firm wants to talk about. What is excluded tells you where the price stops, which is the boundary that will matter in month seven.
Common exclusions worth finding. Four.
Changes to the website itself, content beyond a stated amount, technical work needing a developer and anything involving your advertising. Each is frequently assumed to be included and frequently is not.
What to ask when none are stated. Directly.
Ask what falls outside the fee and how it would be handled. A firm comfortable with its own pricing answers immediately, because the boundary is one they work with constantly.
Why omission is not usually deliberate. Exclusions do not sell.
A section listing what you will not receive weakens a document, so it gets dropped. That is a sales instinct rather than a trap. It produces the same problem for you.
What to do with the answer. Get it written down.
Whatever they tell you, ask for it in the proposal or the agreement. A verbal exclusion is not one.
Compare Like With Like
Two proposals at similar prices are frequently for very different amounts of work. Comparing the figures before establishing that is how businesses choose badly with complete confidence.
What to establish first. Three things.
Which workstreams each covers, roughly how much time each represents, then what each excludes. Until you have those, the prices are not comparable numbers.
How to build the comparison. One page, side by side.
Write the workstreams down the left and each proposal across the top. The gaps become visible immediately, usually not where you expected.
The commonest difference. Content volume.
Content is the largest cost in most arrangements, so a difference in how much is produced explains most price gaps. Establish that before concluding anything about value.
The second commonest. Who does it.
Senior time and junior time cost differently. Two proposals with identical deliverables may represent very different levels of experience applied to them.
What to do with a cheaper proposal. Find the reason.
There is always one. It is usually visible once you ask. Fewer hours, narrower scope, more junior people or work done elsewhere. Any may be acceptable if you know which.
The Numbers In It
Any traffic or revenue projection in a proposal should be treated as an illustration. A forecast that does not state its assumptions is a sales device rather than an estimate.
What a projection needs to be useful. Its workings.
What it assumes about your current position, the searches involved, how many people are searching and what proportion would arrive. Without those, the figure is a number somebody chose.
Why projections appear at all. They are persuasive.
A chart rising to the right is more convincing than a description of work, which is precisely why it deserves more scepticism rather than less.
What a responsible version looks like. Hedged and shown.
A range rather than a figure, assumptions listed, then a clear statement that it is illustrative. Somebody prepared to show the workings is confident they survive inspection.
The standard we hold ourselves to. The same one.
Nothing on this site states a projection as fact. No client result appears anywhere that could not be evidenced. Applying a standard to others that we did not meet would be visible to you.
What to ask. One question.
Ask what the projection assumes. If the answer is vague, the number was decorative and can be ignored.
Ask Them To Explain One Thing
Pick a single item from the proposal and ask why it is in there. The answer reveals whether the document was considered or assembled from a template.
What to pick. Something specific.
A named deliverable rather than a section heading. Ideally something that would only make sense for a business like yours, so a general answer is visibly a general answer.
What a good answer sounds like. Connected to you.
An explanation referring to something they found on your site or something you told them, plus what they expect it to achieve. Somebody who decided to include it can say why immediately.
What a poor answer sounds like. A description.
Explaining what the item is rather than why it is there. That is the response of somebody who did not choose it, because it was in the template for your kind of business.
Why this works so well. It cannot be prepared for.
They do not know which item you will pick, so it tests the whole document rather than the section they rehearsed.
What comes next. The other checks.
The proposal is one of four vetting steps, alongside the questions and the track record. The full series is on the hiring an SEO agency guide.
Pick one item.
Ask why it
is in there.
Do it to our proposal and to everybody else's. Somebody who chose an item can say why immediately and connect it to something on your site. Somebody working from a template describes what the item is instead.
What we commit to in writing:
Local £350 a month fixed. National £1,550 a month. Published rather than quoted on request.
Every guide.
One practice.
What an agency is, whether to buy at all, which model suits you, how to vet one, what things cost, what should be in a contract and how to work with them afterwards.
