Which digital channels matter in Vietnam

A Vietnam channel plan usually arrives as a translation. The structure that worked in another market gets adapted, the platform names are adjusted, budgets are scaled to market size, and the work begins.

The translation is usually competent. The problem is that a plan can be translated perfectly and still be aimed at the wrong step of the buying journey.

Two questions before any platform is named

Where does discovery happen?

In many Western markets, a considerable share of buying starts with a search: the customer knows roughly what they want and goes looking for it.

In Vietnam, for a large part of consumer goods, discovery starts in a feed — short video, social posts, live selling — or inside a marketplace’s own browse and recommendation surface. The customer was not looking for your product. The product reached them.

This changes what the first touch has to do. A plan built around capturing existing demand will find less existing demand than expected, and will conclude that the market is smaller than it is.

The conclusion is the expensive part. A first quarter spent on search for a category nobody searches for produces a thin result, the thin result gets read as weak demand, and the entry budget is cut before the discovery channels have been tried at all. The plan was never tested; only one assumption inside it was.

Where does the purchase complete?

This determines what you can measure, which in turn determines what will look successful.

The last two rows are where most channel comparisons in Vietnam quietly break. A channel whose job is to start a conversation records a click and nothing else, so it loses every comparison against a channel that closes on a website — not because it performs worse, but because it is measured worse.

What tends to surprise a team arriving from elsewhere

Messaging does the work email does elsewhere. Zalo is where a large share of conversation, follow-up and repeat contact happens. A considered purchase with a three-week decision needs something holding contact in between, and in Vietnam that is usually a message thread rather than an email sequence.

A practical consequence: the head-office metric for this channel is usually the cost of opening a conversation, because that is what the platform reports. The number that matters is the cost of a conversation that ends in an order, and producing it requires the sales side to record outcomes against the source that started the thread.

Live and creator selling are ordinary. Live commerce and creator-driven affiliate sales are established purchase paths, not experiments to pilot. Their costs sit outside the advertising account, which makes them easy to leave out of efficiency calculations and easy to underinvest in for the wrong reason.

The marketplace is a media channel, not just a shelf. Each platform runs its own advertising system with its own rules. Treating marketplaces purely as distribution, with media planned separately, misses that the two are the same decision — and the economics of that decision are set by fees and promotions, covered under marketplace advertising.

Adapted global assets are a channel decision, not a cost saving

The line in the budget that says “localisation” usually means subtitling and a language pass. That is enough for a brand film and not enough for the formats that carry most of the volume here.

Feed and short-video placements are made natively or they are skipped. A horizontal brand asset cropped to vertical, with a voiceover added, reads as an advertisement from somewhere else — and the cost of that shows up as a lower conversion rate, not as a rejected ad. Live selling and creator formats go further still: the creator’s own manner is the format, which means the brand supplies boundaries and product facts rather than a script.

This has a budget consequence worth stating plainly. Production for these channels is smaller, faster and more frequent than a global calendar assumes: many low-cost assets tested continuously, rather than three expensive assets running for a quarter. A plan that funds media generously and production annually will starve exactly the channels that were supposed to drive discovery.

It also has a governance consequence. The approval chain that works for a campaign film — legal, brand, region, market — cannot process a weekly volume of short assets. Either the chain is shortened for this class of content, with clear boundaries agreed once, or the channel quietly stops being used.

Neither of these is a reason to lower brand standards. It is a reason to decide, before launch, which standards are absolute and which were simply the habits of a market where the formats were different.

The channel list, and what each part is for

Naming platforms is the least useful way to plan, because the names change faster than the structure underneath. The structure is stable, and it has six parts.

Two observations follow from the right-hand column. First, three of the six cost centres sit outside the advertising account, so an efficiency calculation built only from that account is measuring less than half the spend. Second, no two groups report into the same place, which is why a single blended number has to be assembled by hand rather than read off a dashboard.

That assembly is a small amount of work done once and repeated monthly. It is also the difference between a channel plan and a list of channels.

Who will run it

Foreign entrants tend to settle the channel mix first and the operating model second. In practice the second decision constrains the first, because some channels are very difficult to run from outside the country.

Most entrants end up with a combination, and the combination is where the coordination cost appears. Three practical points, all easy to settle early and expensive to fix later:

• Own the accounts from day one. Advertising accounts, the business manager, the pixel and the customer lists should sit under the company’s own entity, with partners granted access. Choosing how to select and oversee a local agency is a separate decision from choosing channels, and it should not be made by the same conversation.
• Decide the language of record. Reporting can be bilingual, but one language has to be the one decisions are made in, or nuance gets lost twice.
• Set the chat response standard before launch. A messaging channel with a four-hour response time is a different channel from one with a ten-minute response time, and the difference shows up in the conversion rate, not in the media plan.

Whichever combination is chosen, name one person who reads all of it together. Each partner will report its own part well. The questions that cost money sit between the parts, such as a marketplace campaign competing with a feed campaign for the same buyers, or a creator brief that contradicts the product page, and they are visible only to someone who sees every report side by side.

Concentration beats coverage in the first year

The instinct on entry is to be present everywhere, so that nothing is missed. It is the most expensive instinct in the plan.

Every channel carries a fixed cost that has nothing to do with media: a learning period, a creative format, someone who checks it, and a line in the report that someone has to read. Four channels aimed at the same audience is one channel divided by four, plus four times the overhead.

A more useful rule: fund one discovery channel and one closing channel properly, until the discovery channel stops responding to additional budget. That point — where more money stops producing more result — is real information, and it cannot be found by a channel that was never funded past its learning period.

That point needs to be defined before launch, or it will be found by argument. A workable version: raise one channel’s budget in steps of a fixed size, hold each step for at least two weeks, and record what each additional order cost at each step. When the cost of the extra orders from a step rises well above the channel’s average and stays there, the channel has stopped responding, and the next dollar belongs somewhere else.

Add the third channel when there is a reason expressed as a sentence about customers, not as a gap in a coverage chart.

A sequence that avoids the common errors

• Establish where your category is discovered, separately for each product line rather than for the company.
• Establish where the purchase completes, and accept that this decides what is measurable.
• Fix the measurement gaps that matter — usually chat-to-order, and off-platform traffic into marketplaces.
• Only then allocate budget, and review attribution before using the results to cut anything.

The order matters. Allocating first and measuring later produces a plan that looks data-driven and is driven by the data that happened to be easy to collect.

Across a first year, that sequence tends to fall into three phases:

Five questions for the head office

Before a channel plan is signed, five questions usually expose whatever is still an assumption. None of them requires local knowledge to ask.

• Which line of this plan depends on the customer already looking for us, and is that true here?
• Which channels in this plan cannot currently be measured, and what does it cost to fix that?
• Which costs sit outside the advertising account, and who is tracking them?
• Who owns the advertising accounts and the customer data on the day this contract ends?
• What result would make us stop a channel, and was that written before or after we saw the numbers?

The fifth question is the one that changes behaviour. A stopping rule written in advance is a decision; the same rule written afterwards is an explanation.

Where MWY sits

MWY runs no campaigns, produces no content, and takes no commission from any platform or agency.

The work is helping decide the channel mix against how your category is actually bought, getting the measurement right before spending scales, and then reading the results back each month from the company’s side of the table.

Settling the sequence before launch is what Go-to-Market Strategy covers. Reading the numbers back every month once spending is live, and challenging the plans that produce them, is Vietnam Marketing Advisory & Oversight.

What MWY does not do is forecast what a channel will return. Anyone quoting a return before the measurement exists is quoting a number that cannot be checked, and a number that cannot be checked is not a forecast.