Playbook · 7 August 2026

How to Audit a Social Media Agency Proposal in Fifteen Minutes

Two or three quotes, similar numbers, no way to tell them apart. This is the ordered fifteen-minute check to run on all of them, including on ours.

By Launch Media Studios

You are holding two proposals, possibly three. They sit within a few hundred dollars of each other, they all promise somewhere between sixteen and twenty posts a month, and they all use the same vocabulary: strategy, content pillars, engagement, growth. Nothing in them tells you which one contains more actual work.

That is not a failure of your reading. Post count is the number every proposal leads with and the number that means the least. Sixteen posts built from a crewed shoot day and sixteen posts built from your existing photo folder are different jobs, and the gap between them is roughly fourfold in price. Almost every other difference is noise by comparison.

What follows is an ordered fifteen minute procedure. It is arranged so the largest cost line surfaces in the first two minutes and the cheapest omissions surface last. You do not need marketing knowledge to run it. You need a calculator and a willingness to write "not stated" a lot.

Our disclosure, plainly. Launch Media Studios is a Singapore social media and content studio. We sell the thing this article teaches you to scrutinise, so read it accordingly. Run every step below against our proposals too, and if ours cannot survive the check, do not sign it.

Set up: one page, eight rows

Do not read the proposals front to back. Open a blank page and write the eight row labels from the normalisation table further down this article. Your job for the next fifteen minutes is to extract eight facts from each document and put them side by side. Most proposals will not answer all eight. The blanks are the finding, not a gap in your understanding.

Minutes 0 to 2: find the shoot days

Search each document for the word "shoot". Then "filming", "production day", "content day", "video". You are looking for one thing: the number of shoot days per quarter, stated as a number.

Shooting is the largest cost line in a social media retainer and the only one that cannot be made cheaper by someone working faster. A day is a day: crew, location, gear, setup, resets, and someone directing. Quoted standalone in Singapore, a short form batch day runs about S$2,000 to S$5,000 and produces roughly eight to twenty finished vertical cuts. Inside an ongoing retainer the same day costs less, roughly S$1,200 to S$2,500 of the monthly fee, because the format is already established and pre production does not restart every month.

Three answers are possible. A number, which is good. Zero, which is entirely fine if it is stated openly, and which means you are comparing a graphics and captions retainer against a production retainer, and those two prices are not comparable at all. Or a phrase such as "regular content shoots", which you should treat as zero until somebody puts a figure next to it. Write down what you find, or write "not stated", and keep moving.

The substitution to watch for

"Content day" is doing a lot of work in some proposals. Two hours with a phone at your office and one person is a content day. So is a crewed half day with lighting and a director. Both can appear identically in a deliverable list. Ask which one is being quoted, and ask how many people are attending.

Minutes 2 to 4: separate original from repurposed

Go through the deliverable list and sort every item into one of two columns. Original means somebody has to make it from nothing this month: a filmed clip, a photographed product, a graphic designed for this specific message. Repurposed means it already exists in some form: a re-cut of last quarter's footage, a supplier's asset, a stock clip, an older post reposted, a customer photo.

Two counting rules matter here. A carousel is one asset, not ten, however many slides it has. And a re-cut is repurposed no matter how good it looks. Total each column separately. You now have the only deliverable count worth comparing across proposals.

Minutes 4 to 5: check the writing unit

Find the captions. You want to know whether the proposal commits to captions written per platform or one caption reformatted across platforms. Both are legitimate products at different prices. Writing natively per channel typically adds 10% to 20% to a fee, because four native captions is four writing jobs and one caption reformatted four ways is one.

If it is not stated, two follow ups settle it: which languages, and whether a human edits anything that is generated. Generated copy is not the problem. Unedited generated copy is, and your customers tend to recognise it within about three posts.

Minutes 5 to 7: find the community management commitment

This is the quietest line in most proposals and the first one clients notice missing. Search for "community", "comments", "DM", "messages", "engagement".

You are looking for two specifics: hours per week of human attention, and a response window. "Weekdays, within four hours" is a commitment you can hold somebody to. "Community management included" is a word. Watch particularly for "up to" attached to a number of responses, because a deliverable expressed as "up to" can lawfully be delivered as zero.

This one matters commercially rather than cosmetically. Replies in the first hour materially affect how far a post travels on TikTok and Instagram, so this line is distribution, not admin. Expect it to account for something like S$300 to S$900 a month of a fee, depending on the hours covered and the window promised.

Minutes 7 to 9: work out what month four looks like

Read the deliverable list again, and this time ask what happens only once. Strategy deck: once. Brand shoot: usually once. Account setup, template design, content pillar workshop: once. Now picture month four with all of those removed.

A great many proposals are priced around an onboarding burst. If the monthly fee holds while the deliverables quietly thin out after the first quarter, you are paying a retainer for a project. The fix is one question: ask for the month four deliverable list, itemised, at the same fee. An agency that intends to do the work will send it back the same day.

Minutes 9 to 10: find out who does the editing

Ask who edits, where they sit, and how many other accounts they handle. There is no wrong answer here, only an undisclosed one. An offshore edit team is not automatically worse, but it should be priced accordingly and you should know before you sign rather than when the first round of revisions takes four days.

The related check: whether the senior person pitching you is the person who will run the account. A junior running the account with genuine senior oversight is a normal and reasonable structure. A junior running the account while you were sold a strategist is a different purchase from the one described.

Minutes 10 to 11: read the reporting definition

Find the reporting section and look for one thing only: does any part of it connect to enquiries, or is it entirely platform metrics? Impressions, reach and follower growth are free to produce and they will not let you judge the work in six months. A report that ties posts to enquiries costs something like S$150 to S$400 a month of analyst time, because it means touching your enquiry data rather than the platform dashboard.

It is the smallest line on a proposal and it decides whether you can evaluate every other line. If reporting is a screenshot, the thinness of everything above it will never become visible in a document.

Minutes 11 to 13: read the exit terms

Skip to the contract terms and find four things: the minimum term, the notice period, who owns the platform accounts, and what happens to raw footage and working files when the engagement ends.

The answers should be immediate and boring. Accounts in your name. Raw footage, working files and the content calendar handed over on request at any time, not only at the end. A notice period you can live with. Be careful with a long lock paired with a long notice: a twelve month minimum term with a sixty day notice period is worth considerably more to an agency than a 5% to 10% annual prepayment discount is worth to you, and long lock ins correlate with agencies whose retention depends on the contract rather than on the work.

Minutes 13 to 15: do the arithmetic

Two sums, on a calculator, about ninety seconds.

  1. Fee per genuinely original asset. Take the monthly fee and divide it by the original column from minutes 2 to 4. Ignore the repurposed column entirely. That is the number to put side by side across proposals, and it is crude but far better than cost per post because it measures the thing that actually costs money.
  2. What is left after the shooting. Take the monthly fee and subtract S$1,200 to S$2,500 for each shoot day the proposal includes in that month. Whatever remains has to pay for planning, writing, publishing, replying, reporting, account management and margin.

The second sum settles most arguments quickly. A proposal promising twenty original videos a month for S$1,200 all in has spent its whole budget on the shooting before a single hour of anything else. There are only three explanations: the videos are not original, the day is not a day, or the fee rises once you are onboarded. Ask which one it is, and watch how specific the answer is.

The normalisation table

Fill this in for every proposal before you look at price again. Two quotes at the same monthly figure routinely contain very different amounts of work, and the ranking often changes once the table is complete.

What to extractWhat counts as a real answerWhat counts as "not stated"
Shoot days per quarterA number, plus crew size and duration"Regular shoots", "as needed", an unqualified "content day"
Original assets per monthA count, with carousels counted as oneA single total post count with no split
Repurposed assets per monthA count, stated openlySilence, or originals and repurposed merged into one figure
Writing unitWritten per platform, with the languages named"Engaging captions", "copywriting included"
Community hours and response windowHours per week plus a window, for example weekdays within four hours"Community management included", "up to X responses"
Month four deliverablesAn itemised list, the same in substance as month oneA month one list only
Who edits, and their other accountsNamed team, location, workloadThe pitching senior's name and nothing else
Exit: term, notice, accounts, filesNumbers, plus "your name" and "on request"Anything you would need a lawyer to interpret

A scoring rubric you can finish in two minutes

Score each of the eight rows: 2 if the proposal answers it specifically, 1 if it gestures at it, 0 if it is absent or hedged with "up to", "as needed", "where applicable" or "best effort". Sixteen points are available.

  • 13 to 16. A scope you can hold someone to. Now compare on the two arithmetic numbers, and expect the cheaper monthly figure to lose sometimes.
  • 8 to 12. Normal, and fixable by email. Send the specific gaps back and re-score. Most proposals in this market start here.
  • Below 8. Not a scope, a range with the bottom hidden. Do not negotiate the price yet; there is nothing stable to negotiate against.

One caution about the rubric. It measures the clarity of a document, not the quality of the work. A small studio that writes plainly will outscore a larger one with a cautious legal team, and that is a real signal but it is not proof. Use the score to decide what to ask next, not who to hire.

When the honest answer is not to hire an agency at all

This is the section that costs us money to write, so treat it as the test of whether the rest is worth anything.

Below roughly S$1,500 a month, an agency retainer is usually worse value than paying once for a system and training your own team to run it. The arithmetic explains itself: at that level, after tools, account management and one round of revisions, a team has perhaps a day and a half of blended working time in the month, and a day and a half cannot contain a shoot day. A thin retainer buys thin work, and thin work is indistinguishable from no work in the reporting.

A freelancer is frequently the right call as well, particularly at the lower bands: roughly S$800 to S$2,000 a month for management plus S$800 to S$1,500 per shoot day, provided you have a clear brief, one channel, and somebody internally who can art direct. What a freelancer cannot absorb is a bad week. There is no second person, so a delayed edit becomes a missed posting slot and a resignation becomes a stopped channel.

And an in house hire beats both when the content genuinely depends on deep product knowledge and daily access to your team. Count it properly, though. A capable social media executive here runs S$3,500 to S$5,000 a month in salary, plus employer CPF at 17% for staff aged 55 and below, plus leave and software, which puts the real run rate at roughly S$4,200 to S$6,200 for one person with one skill set.

What to do with the fifteen minutes of work

Send one short email to each agency listing the gaps you found, phrased as a request for minimums rather than as a complaint: shoot days per quarter as a number, community management as hours per week and a response window, and the month four deliverable list itemised. Give them a working day.

The replies sort themselves. Some come back the same day with the numbers filled in, which is the strongest signal available to you at this stage. Some explain why minimums are not how they work. And occasionally the honest outcome is that the quote you suspected was expensive turns out to include twice the shooting, which makes it the cheaper one per unit of work that will still exist in six months.

For the market context behind what you are reading, our guide to what social media management costs in Singapore sets out the four price bands and what each one genuinely buys, and our social media management page states our own scope in the same terms this checklist asks for. Both are written to be checked against each other, which is the only reason to trust either.