Can you trust marketing data in Vietnam?

A regional head office looks at a Vietnam dashboard and sees numbers in the same shape as every other market: impressions, clicks, conversions, revenue. The shape is familiar, so the reading is assumed to be familiar too.

It usually is not. The numbers are not wrong and nobody is hiding anything. They are answering a narrower question than the person reading them believes — and three local habits make that gap wider in Vietnam than in most markets a foreign team has run before.

An order is not revenue yet

Cash on delivery is still ordinary here, not a fringe payment method. A customer confirms an order online, the parcel goes out, and the transaction only completes when someone accepts it at the door.

Everything upstream of that moment has already counted the sale. The advertising platform counted it. The analytics tool counted it. The agency's report counted it. The bank account has not.

The gap between confirmed orders and accepted deliveries is a real, measurable number. It varies by category, by price point, and by how the customer was acquired — a shopper who came from a heavily discounted ad is not the same risk as one who came from search. Until that number is measured and applied, every efficiency figure in the deck is overstated by an unknown amount.

Marketplaces report inside their own room

Shopee, Lazada and TikTok Shop are where a large part of consumer commerce happens. Each runs its own advertising system, its own reporting, and its own definitions.

What reaches your analytics from those platforms is partial by design. You see what the platform chooses to expose, aggregated the way the platform aggregates it. Commission, platform vouchers, shipping subsidies and seller vouchers all land on the same order, and the report you get rarely separates them.

None of those answers is false. They are simply narrower than the question, and the difference compounds across a quarter.

A lot of buying happens in conversation

A meaningful share of purchases in Vietnam is agreed in Zalo or Messenger. The customer sees a post, asks a question, negotiates, and confirms — all inside a chat thread.

Standard web tracking sees none of that. The last recorded event is often the click that opened the conversation, and the sale that followed is invisible unless someone deliberately connects the chat platform to the order record.

The consequence is predictable and expensive: a channel whose job is to open a conversation records the click and nothing after it, while a channel that closes on a website records a clean conversion.

Two consequences that follow quietly

The three habits above are well known to anyone who has operated here. What is less obvious is what they do to decisions once the numbers reach a head office.

Channels get ranked in the wrong order. A channel whose job is to start conversations shows a click and then nothing. A channel that closes on a website shows a click and a conversion. On the dashboard the second looks twice as good. Budget moves. The move is rational given the data, and the data is the problem.

Efficiency looks better than it is, consistently. Every unaccepted delivery is a sale that was counted upstream and never landed. Because the overstatement is systematic rather than random, it does not average out over a quarter — it compounds, and it compounds most in exactly the campaigns that discount hardest.

A short reconciliation, once

None of this requires a data platform. It requires one table, built once and updated monthly:

Build it by acquisition source rather than as a company total. The point of the table is not precision. It is to make the size of the gap visible to the people deciding the budget, and to keep it visible month over month.

What to establish before scaling spend

• Define the revenue line first. Decide whether the business runs on confirmed orders or on accepted deliveries, write it down, and require every report to use that definition.
• Measure the delivery acceptance rate by acquisition source, not as a single company-wide average. The average hides the source that is generating orders nobody keeps.
• Ask each marketplace for the fee breakdown per order, not the summary. If it cannot be produced, treat the margin figure in that channel as an estimate and say so in the deck.
• Connect the chat platform to the order record before judging any channel that works by starting conversations.
• Fix the definitions before raising budget. Scaling on top of a measurement gap scales the gap too.

Where this work stops

MWY reconstructs profit after advertising cost — recorded revenue, less advertising spend, less the acquisition costs that sit outside the ad accounts such as affiliate and creator commissions.

MWY does not reconstruct true profit. That needs cost of goods, inventory and operating costs, which sit outside the scope of independent marketing oversight. Where a margin figure appears in any MWY analysis, it is a number the company supplied, not one MWY derived.

Saying where the work stops is part of the work. An assessment that claims more than its data supports is an assessment nobody can act on.