Three proposals arrive. All three look professional, all three use the same words, and none of them can be compared with the other two, because each describes a different shape of work in a different order. Growth Bully, a Malta performance marketing agency, has written enough of these and reviewed enough of other people's to say what the difference is: a proposal that can be checked and a proposal that can only be admired.
What follows is how to read one. Not how to spot a bad agency, which is harder, but how to force three documents into a form where a decision is actually possible.
What should a marketing proposal tell you first?
One business outcome, stated in your language rather than the channel's. More qualified enquiries, more booked meetings, more online orders, a lower cost per sale. If the opening section talks about visibility, presence, awareness or engagement without connecting them to something that pays, the rest of the document has nothing to be measured against.
Check whether the outcome has a number and a period attached. A proposal that commits to a direction but no quantity has given itself nothing to fail. You do not need a guarantee, but you do need a target, because a target is what makes the monthly review a conversation about results instead of activity.
How do you separate strategy, setup and ongoing delivery?
Ask which parts happen once and which parts repeat every month. Research, audits, tracking installation, account builds and creative production for launch are one-off work. Campaign management, creative refresh, reporting and optimisation repeat. A document that blends them makes month two look identical to month one when it will not be.
This is the single most useful question to ask of any proposal, because it exposes how much of the fee buys work that will never happen again. It also tells you what happens if you stop after the setup phase, which is a reasonable thing to want to know before you start. The broader structural question is covered in retainers against one off projects.
What does a vague deliverable look like?
Anything you cannot count at the end of the month. Ongoing optimisation, regular content, continuous improvement and social media management are categories, not deliverables. The test is whether two reasonable people would agree on whether it was delivered, and with that phrasing they never would.
Rewrite each line as a countable unit before you compare proposals:
- How many of the thing, per month, stated as a number rather than a range that starts at one.
- Who produces it, the agency or you, because client-supplied input is the most common hidden cost in a marketing retainer.
- How many rounds of revision are included before extra work is chargeable.
- What the turnaround is, from brief to live, since a deliverable with no timing can arrive in the last week of the month.
- What happens to unused units, whether they roll over or expire.
Doing this to three proposals usually collapses the field quickly. One of them turns out to contain far less work than its length suggested. Writing the brief yourself first makes this much faster, and how to brief a marketing agency covers what to put in it.
How should a proposal handle measurement?
By naming what will be counted, where it will be counted, and who installs the counting. If measurement appears only as a reporting promise at the end of the document, the first month is likely to be spent discovering that the tracking was never reliable and that the baseline cannot be established retrospectively.
Look for a definition of a qualified lead as well. Without one, both sides will report honestly and disagree completely, because the agency counts form submissions and you count people worth calling. The gap between those two numbers is where most agency relationships quietly sour, and what counts as a qualified lead sets out how to agree it in advance.
Who should own the accounts and the data?
You should, in every case, with the agency holding access rather than title. The proposal should say so explicitly for the ad accounts, the analytics property, the tag container, the domain, the CRM and anything built inside them, including audiences and creative files.
Silence here is the costliest ambiguity in the whole document, because it is only discovered at the end of a relationship when leverage has gone. Ask for the handover position in writing before signing rather than after, and read who owns your ad accounts for what to request specifically.
What does a realistic first ninety days look like?
Slower than most proposals suggest. A credible plan spends the first few weeks on tracking, access, offer and baseline, launches in the first month, and expects the numbers to be unstable until there is enough volume to read. Anything promising transformation in week two is selling comfort.
Judge the plan by whether it tells you what happens if the first launch underperforms. Every campaign has a bad first month at some point, and the proposals worth taking seriously describe the diagnostic sequence rather than implying it will not happen. A month by month view of what normal looks like is in the first ninety days with an agency.
How should the money section be read?
As a question about completeness rather than size. The figure you are comparing has to include everything the work requires, so check whether media budget, tools, subscriptions, stock assets, production, landing pages and third party fees sit inside the fee or arrive separately. A low number with four exclusions is the higher number.
Then check the commitment terms, in this order, because each one changes what the figure actually buys:
- What the initial committed period is, and whether it differs from the billing period.
- What notice applies once that period ends, and in writing to whom.
- Whether anything is billed in advance, and what happens to it if you stop early.
- What happens to work in progress and produced assets on the day you leave.
- Which costs sit outside the fee, listed rather than implied.
None of that is unreasonable to ask, and a proposal that answers it without being asked is usually written by someone who has run the relationship before rather than only sold it.
Which parts of a proposal are red flags?
Four, in order of how much they cost. Guaranteed rankings or guaranteed sales, which nobody controls. Proof with no period, platform or context attached. A deliverable list with no numbers. And a document that never once mentions your own business specifics, which means it was assembled rather than written.
One more worth noticing: a proposal that cannot say what it will not do. Scope is defined by its edges, and a document that implies everything is included has simply moved the argument to month three. If you want a proposal built around one named outcome with the deliverables stated as countable units, ask us for a pipeline review first, see how the delivery method works in the Decision-Maker Pipeline, or compare the wider options in building in house against hiring an agency and on our lead generation service.