The Vietnam spend is not delivering: what to check before cutting it

The request usually arrives as one sentence from head office. Vietnam is not delivering; bring a plan to reduce it by a third before the next quarter.

What follows is familiar. The regional team trims every line by roughly the same percentage, the agency argues that the numbers undercount its work, and the decision is taken on whichever argument sounds more responsible. Sometimes the cut is right. It is rarely the right first decision, because "not delivering" is a verdict, and four quite different problems produce it.

Not delivering against which number?

Before anything about Vietnam is examined, look at the target. Write it on one line, with where it came from.

In most foreign companies it came from one of three places:

• The business case written before entry. That was a forecast, made before a single order existed in this market. Missing a forecast tells you the forecast was wrong; it does not yet tell you the market is.
• Another market's benchmark. A return or cost-per-customer figure that works at home, carried across unchanged. It assumes the same payment habits, the same fee structure and the same share of sales happening where a tracking pixel can see them.
• The agency's proposal. A number set during a pitch, where it also had a second job: winning the account.

None of those was built from local economics. A target is only fair to judge against if it was derived from the margin the company earns in Vietnam, after the fees it pays in Vietnam. If that calculation has never been done, "not delivering" means "below a number nobody built for this market" — and the first job is to build it.

Four problems that look identical from abroad

From a head-office dashboard, all four of these produce the same picture: spend going out, attributed revenue below target, no clear trend. They call for opposite responses.

The order in which you check them matters, because they differ enormously in cost. The first two take days and a few exports. The third is a spreadsheet built on a margin figure the company already has. The fourth takes weeks of channel-by-channel analysis. Checking the expensive one first — which is what a debate about agency performance amounts to — is the most common way to spend a quarter arguing about a reporting gap.

Check one: is the result there but unseen?

Vietnam has more places for a sale to happen out of sight than most markets a foreign team has run. The mechanics of each are well documented; what matters here is the test.

Three routes account for most of the invisible part:

• Chat. A buyer clicks an ad, opens a Zalo or Messenger thread, and orders there. The ad platform records a click and nothing after it.
• Marketplaces. Advertising outside Shopee or TikTok Shop sends people who then buy inside them, where your analytics does not follow — the fee side of that picture is under marketplace advertising.
• Distributors. If a distributor sells to the end customer, the brand sees sell-in, which moves when the distributor restocks. Advertising moves sell-out, which the brand often does not see at all.

The quickest test does not need any tracking to be fixed. It needs history. Look back over the last eighteen months for any period when spend dropped sharply for a reason that had nothing to do with performance: an account suspended for a week, a budget freeze at year end, campaigns paused over Tết. Then look at total sales across every route — not attributed sales — for the six weeks that followed.

If total sales barely moved, the spend was probably doing less than anyone hoped. If they fell in step, with a lag, the spend was doing more than the reports credited, and the "not delivering" verdict is at least partly a reporting problem.

It is an imperfect test. Other things change during a holiday, and a single episode proves little. But it uses evidence the company already owns, it costs an afternoon, and it answers a question that no amount of dashboard reading can.

Check two: was the target built here?

The second check is arithmetic, and it is where most of the gap usually turns out to be. An illustration, with invented numbers — a consumer brand selling partly on its own site with cash on delivery, partly on marketplaces:

Read top to bottom, the story changes twice. The reported figure is worse than it looks, because some orders are never accepted. Then it is better than it looks, because about a third of the sales were closing where the platform could not see them. And the target itself was wrong: the business needed 2.9 to break even here, not 4.0.

The verdict moves from "delivering just over half of target" to "roughly at break-even overall". That is not a comfortable result, and it is not a success. But it is a completely different decision. An overall figure at break-even means some channels are comfortably above it and some are well below — which makes this a question about channels, not about the size of the budget.

Two cautions about the restatement. The margin in the last row must be the figure the company supplies, not one estimated from outside. And the definition of "return" has to match on both sides: a home-market target measured on net revenue after returns cannot be compared with a Vietnam figure measured on orders placed.

Check three: can the offer pay for a customer here?

Once the target has been rebuilt locally and the invisible sales counted, some companies find the gap has closed. Others find it has narrowed and stayed open. That remaining gap is the one worth taking seriously.

The test is channel by channel: cost per accepted order in each channel, against the most the company can afford to pay for one, which follows from the same margin and fee figures as above. Three patterns suggest the problem sits in the economics rather than in the marketing:

• No channel clears the ceiling, including the ones that are well run by every other measure.
• The best channel reaches its ceiling at a very low level of spend. The first small budget works; every increase after that costs more per order than the order is worth.
• Orders only arrive at deep discounts. Sales volume depends on vouchers large enough to consume most of the margin, and falls away as soon as they end.

When those patterns appear, cutting spend is correct: there is no point buying customers at a loss while the question is open. But the question itself — price point, pack size, the margin left after a distributor's share — is a commercial one. Marketing analysis can show precisely where the gap is and how wide. Closing it sits with the people who set the price and the commercial model.

Check four: is it the execution?

Only after the first three checks does it make sense to look hard at the work itself. By then the question is narrow: given a local target and complete sales figures, which parts are underperforming?

Weak execution has a recognisable signature. The channels differ from one another far more than the market would explain. One channel was never funded past its learning period and was judged anyway. Creative was translated from another market rather than rebuilt for this one. A single agency contract covers marketplace operations, creator programmes and platform buying, and some of those disciplines are quietly not being done. Reporting is built around a metric the business does not decide on.

Each of those has a specific remedy, and none of the remedies is a smaller budget spread the same way. Where the partner is part of the problem, the questions that separate a fixable relationship from a drifting one are covered under choosing and overseeing an agency.

What a flat cut does in this market

A cut applied evenly across every line looks neutral. In Vietnam it rarely is, for four reasons.

It hits the invisible channels hardest. On attributed data, the channels that start conversations, feed marketplace searches or support distributor sales look like the worst performers, so they are the first to be trimmed beyond the flat rate once anyone looks at the numbers. They are also the ones whose effect was least counted.

The effect arrives late, and somewhere else. Chat volume falls over a few weeks. Marketplace sales soften as search interest drops. Distributor orders slow a month or two after sell-out does. The month after the cut looks fine, and the decline that follows is usually read as the market slowing down.

Marketplace placement is not independent of paid spend. Sales history feeds how a listing is ranked, so cutting advertising inside a marketplace also weakens the organic sales that the same listing was earning.

Restarting costs more than continuing. Automated bidding systems have to relearn after a large change, audiences go stale, and creators who were working with the brand move on. Restoring a budget a quarter later does not restore the position.

None of this means cuts are wrong. It means a cut should be designed, not applied.

If you do cut, cut so that you learn something

A well-designed reduction produces evidence. A flat one produces a smaller version of the same argument. Four rules make the difference:

• Cut in order, not in proportion. First the channels that sit below their own break-even after restatement. Then any test spend that has passed its deadline without a result. Then the portion of spend in working channels that sits above the point where more money stops producing more orders. The core of the working channels comes last.
• Move one thing at a time. Pause or reduce one channel, hold the rest steady, and watch total sales across every route for a fixed period. Two changes at once teach you nothing about either.
• Write down in advance what would reverse the decision. A specific fall in total sales or in chat volume, over a specific number of weeks. Without it, any result can be read as confirming the cut.
• Keep the test clear of the calendar. A reduction in the weeks after Tết, or around the large marketplace sale days, will be compared with a period that was never normal.

Done this way, a cut becomes the cheapest experiment the company will run all year: it saves money while it measures something the reports cannot.

What to bring to the budget meeting

Five items settle most of these conversations, and all of them already exist somewhere inside the business:

• The target, with its source and its definition — which revenue it counts, and where the number came from.
• Total weekly sales across every route for the last eighteen months, set beside weekly spend.
• Every period when spend dropped sharply, and what total sales did in the six weeks after.
• Cost per accepted order by channel, set against the local ceiling.
• The share of sales the reporting cannot see — chat, marketplace and distributor — even as a rough estimate.

If any one of the five cannot be produced, that is a finding in its own right, and a more useful one than another month of debate over whether the agency's numbers are right.

Where this work stops

The check should be run by someone who gains nothing from either answer. The agency has a reason to find a measurement explanation; the team proposing the cut has a reason to confirm it. That is a statement about incentives, not about honesty, and it is why an independent read of the same data settles the question faster than either side can.

MWY runs these four checks as part of a Digital Marketing Audit and does not run media, produce content or take commission from any platform or agency. The recommendation may be to cut. It may be to cut one channel and fund another. It is never shaped by who benefits from the budget staying the same. What that looks like on paper is in the sample audit report, written for a fictional foreign brand selling through a distributor.

Two limits are stated precisely. MWY reconstructs profit after advertising cost; it does not reconstruct true profit, which needs cost of goods and operating costs outside the scope of marketing oversight. And every margin figure in the analysis is one the company supplies.

Keeping the local target, the complete sales picture and the channel ceilings current each month — so that the next budget conversation starts from evidence rather than from a verdict — is Digital Marketing Advisory & Oversight.