A complete audit, to read before you pay for one

The business assessed

Company V. is a fictional business, built for this sample. Every figure below was invented by MWY to illustrate the output. It is not a disguised client.

The digital budget breaks down as follows. This is MWY’s reconstruction from invoices and card statements, not the figure the agencies report — the two differ at most companies, and the difference usually sits in tooling and content production.

Summary for headquarters

Written to be read alone. For most boards this is the only page that will be read, so it has to stand without the rest.

Head office reports sell-in. The Vietnam team reports sell-out. Neither document says which it is, and over the last three quarters the two have drifted 31% apart. Every performance conversation about this market has been a conversation between two different numbers.

This is the largest finding and it is not a data problem in the technical sense — both numbers are correct. It explains three things the business has felt without naming: why a strong quarter in Vietnam is followed by a flat one for no visible reason, why the local team and the regional office disagree about whether marketing is working, and why the inventory conversation and the marketing conversation never quite line up.

The other four findings, ordered by the money exposed:

• The Meta Business Manager belongs to an agency, the marketplace seller centres belong to the distributor, and the analytics property belongs to the regional office. No single party can see all three, and the brand entity owns none of them.
• Vietnamese-language brand searches run 4.3 times the English-language ones, while the search campaigns inherited from the regional template are built on the English brand term.
• The distributor and the brand are bidding against each other. Twenty-three percent of marketplace ad spend goes to keywords both parties are buying, which raises the cost for both without adding a buyer.
• Forty-four percent of measurable revenue sits on one marketplace, and no threshold for that concentration has ever been agreed.

Digital health score: 2.3 out of 5. That is a normal position for a brand twenty months into a market it entered through a partner. Nothing here is broken; several things were inherited rather than decided.

What this report does not conclude: that either agency is underperforming. Both are optimising precisely the metric their contract measures. The metric in those contracts is not the metric head office uses to judge the market — and that is a question for whoever signs, not for whoever executes.

Measure

What is actually happening

The first four weeks make no recommendations. The only job is to rebuild the numbers from source and then measure how far they sit from the numbers being used in meetings.

Three systems are counting what looks like the same thing and producing three answers. The figures below are for the most recent quarter, and the gap has been stable across the six months reviewed rather than being a one-off.

None of these is wrong. The problem is that the group report uses the first, the local marketing report uses the second, and nobody has written on either document which one it is.

On ownership, the position is worth stating plainly because it is the finding with the shortest window to fix:

Two further measurement gaps, both specific to operating this market from outside it:

• Zalo carries roughly a third of inbound customer conversations and is absent from the measurement stack entirely, because the stack was specified in a region where Zalo does not exist.
• Sales recorded in Vietnam reach the group system on a 19-day lag, which is not itself a problem — but the lag is undocumented, so every month-end comparison silently compares two different periods.

Fifty-eight percent of Vietnam revenue reaches shoppers through general trade, where no digital measurement is possible at all. That is a normal structure for this category in this market. What is not normal is that the figure appears nowhere in the marketing reporting, so every efficiency number head office reads is silently calculated on the minority of the business.

Digital health scorecard

Six categories, each scored out of 5, aggregated from the nine areas MWY oversees. Scored first during the audit and rescored each quarter — the value is in the second scoring, when you can see what actually moved.

• Strategy & digital funnel

  The channel mix was inherited from a neighbouring market rather than derived here, and the division of roles between the brand and the distributor has never been written down.

  2/5

• Measurement & data

  Sell-in and sell-out unlabelled, a 19-day undocumented reporting lag, and no measurement of the channel carrying a third of customer conversations.

  2/5

• Paid media & budget

  Accounts are well structured and there is no abandoned spend. The deduction is for search campaigns built on the English brand term while the demand sits in Vietnamese.

  3/5

• Marketplaces

  Reported ROAS does not net out platform co-funded vouchers, and the brand cannot access its own seller centre data without going through the distributor.

  2/5

• Owned channels, content & retention

  The direct website converts below the brand’s own marketplace stores. Creative follows global guidelines that exclude the two formats that convert in this category here.

  2/5

• Martech, AI & agencies

  The toolset is proportionate and not duplicated. Both agency contracts measure platform ROAS while the group judges the market on sell-out revenue.

  3/5

Average 2.3 out of 5. How to read the scale: 1 is absent; 2 is present but not dependable; 3 is working and sufficient; 4 is producing an advantage against the market; 5 is a standard others would copy. Most companies sit at 2 and 3, and that is not a criticism — moving two categories from 2 to 3 is usually worth more than moving one from 3 to 4.

Why

Why the results look like this, and why the money goes where it goes

The slowest part of the four weeks, and the part an audit run by the party spending the budget almost never reaches.

The channel mix was inherited, not chosen. The split between Meta, Google and TikTok matches the template used in a neighbouring market to within two percentage points. That is a sensible way to start in a market you do not know. It stops being sensible at twenty months, by which point there is enough local data to derive a mix from, and nobody has been asked to do it.

The search budget is spelling the brand name wrong for this market. Vietnamese-language queries for the brand run 4.3 times the English-language ones, but the campaign structure came from the regional template and buys the English term. The gap is not a translation error in the ads; it is that the keyword set was never rebuilt for the language the demand is in.

The brand and its own distributor are competing at auction. Twenty-three percent of marketplace ad spend sits on keywords both parties are bidding on. Neither is doing anything wrong: the distributor is paid on sell-through and is behaving accordingly. But no territory was ever agreed, so both sides are paying more to reach a buyer who was going to arrive either way.

Marketplace ROAS is being read before the discount. The platforms report 5.4 on the core range. Forty-one percent of that revenue carries a voucher co-funded with the platform, which the reported figure does not net out. Using the gross margin Company V. supplies, the break-even threshold for this range sits near 3.8 — so the headroom is real but a great deal thinner than 5.4 suggests.

A limit worth stating: MWY can reconstruct profit after advertising cost. It cannot reconstruct true profit, which needs cost of goods, landed cost and operating expense — outside the scope of a digital marketing assessment. The margin used above is the figure Company V. provided.

Global brand guidelines are excluding the two formats that convert here. Price-led creative and bundle offers are both prohibited by the global toolkit. On Vietnamese marketplaces those are the two formats that move this category. Holding the rule may well be correct — a premium position is worth protecting. But the cost of holding it has never been quantified, so the decision is being made by default rather than on purpose.

Nothing in this table is a mistake. The pattern is that five decisions shaping this market were taken somewhere else, and none of them has a scheduled review with Vietnam data in front of it.

Concentration has no agreed ceiling. Forty-four percent of measurable revenue sits on one marketplace. That number is not a problem in itself. It becomes one because nobody has said what level the board is willing to carry, which means there is no trigger at which anyone is required to act.

The agency contracts measure something the group does not use. Both agreements are settled on platform ROAS. Head office judges the market on sell-out revenue. The distance between those two is the same 31% established in the Measure section. A partner will improve quickly at whatever their contract measures, so this belongs to the party that wrote the contract.

Yield

What to do, in what order, and who owns it

Five actions inside 30 days. Each has a named owner and a deadline, and none of them requires additional budget to begin.

Action 4 is the one with a deadline set by something other than convenience: account transfers are straightforward while a relationship is good and become slow the moment it is not. Action 1 is first because until it is done, the other four cannot be judged.

Priority matrix for everything else, impact against effort:

Proposed KPI framework. The right-hand column is not a set of new metrics; it is the same metrics taken from a different source:

The 90-day roadmap runs in three stages: the first 30 days settle what each number means and move the accounts into the brand’s name; the next 30 rebuild search and marketplace reporting on that basis; the last 30 open the two conversations that need evidence in hand — territory with the distributor, and the price of the global creative rules. The order matters. Reopening either of those conversations before the numbers are agreed turns a question of evidence into a question of who is more senior.

What this report does not say

This section appears in every real audit, not only in the sample. Stating the boundary inside the product is the only reliable way to stop it being used for something it cannot do.

• It does not conclude anything about true profit. The scope stops at profit after advertising cost; cost of goods, landed cost, inventory, delivery, returns and after-sales sit outside it.
• It does not assess individuals, the distributor, or either agency as organisations. It describes the metric currently measuring them.
• It is not legal advice. Where advertising regulation is touched, the report flags it for the company’s own counsel rather than ruling on it.
• It does not forecast revenue. MWY commits to delivery dates, not to a level of result.
• Where the data cannot support a conclusion, it says so. In this sample, the contribution of general trade to online demand is one such place: nothing currently measurable connects the two, so the report proposes how to establish the link rather than asserting a number.