Who should run your marketing in Vietnam: a local agency, a regional one, or someone of your own

A foreign company entering Vietnam usually starts comparing agencies before it has decided what it is hiring an agency for. Proposals arrive, credentials are checked, and the question of who will actually decide things in the market gets answered by default, which normally means by whoever wrote the best proposal.

That question deserves to be settled first. There are four ways to run marketing in a market where you have no one yet, and each one fails in a predictable place.

Four models, and where each one tends to fail

• A local Vietnamese agency runs execution and reports to head office.
• A regional agency or network extends its existing contract to cover Vietnam.
• Someone of your own is placed in the market: a full-time country marketing manager, or a fractional marketing lead who works part of the week.
• Head office runs it directly, with advertising accounts managed from abroad and perhaps a freelancer or two in Vietnam.

Most companies end up with a combination. The combination is where the real choice lies, but it is easier to assemble once each piece is understood on its own.

What has to be covered, whatever the model

Four jobs have to sit somewhere. A model that leaves one of them with nobody is not cheaper; it is incomplete.

• Execution in Vietnamese, on local platforms. Content, creators, marketplace stores, livestream, and answering customer chats at the hours customers write.
• Day-to-day decisions. Which creative stays, where next week's budget moves, which test stops. These happen weekly, not quarterly.
• Carrying context to head office. Not translation, which any agency can do, but explaining why a number differs from other markets and whether that difference matters.
• Checking the numbers. Reading results against the company's own order records, by someone whose own results are not being judged.

No column fills all four rows. That is the practical argument for a combination, and for deciding deliberately which piece covers which row.

A local agency: fast in the market, thin on the reporting line

Local agencies usually move fastest in Vietnamese, understand platform-native creative, and are more likely to run marketplace operations, creator programmes and chat response in-house. For a brand whose sales will depend on those disciplines, which covers most consumer categories, a local agency is usually the right place for execution to sit.

The weak point is the line back to head office. Often a single account director speaks fluent English, so everything the head office learns passes through one person and one monthly deck. Decisions drift towards the agency, not because anyone takes them, but because the client is several time zones away and the campaign needs an answer today.

Neither problem is a reason to avoid a local agency. Both are reasons to put a named decision owner and an independent check alongside it.

A regional agency: consistent reporting, local work subcontracted

A regional network brings one contract, reporting that matches the group template, familiarity with brand standards, and strong buying on the large international platforms. Where Vietnam launches alongside several markets with the same brand and most spend goes to Meta, Google and YouTube, that consistency is worth a great deal.

What it rarely brings is depth in the local disciplines. Marketplace operations, creators, livestream and chat response need local labour and local relationships, and a regional contract often covers them through a partner. Vietnam is also usually a small line in a regional plan, which affects who is assigned to it and how quickly approvals come back.

A regional agency fits best as the owner of international media and group reporting, with local disciplines carved out by name to a local specialist. The more of the plan that runs through marketplaces and creators, the smaller the regional agency's share should be. Where marketplaces are the main sales route, how marketplace economics work matters more than which network holds the media contract.

Someone of your own: a country hire or a fractional lead

A full-time country marketing manager gives the market a decision owner who is present, reads the numbers in context and builds the relationships with agencies and platforms. The risk is timing. Hired before launch, the manager is hired for a job nobody has seen yet: the balance between marketplace work, creator management and media is unknown, and so is the scale. A second risk is quieter. A manager who chose the agency and approves its plans ends up checking their own decisions, and the checking job gets absorbed into the managing one.

A fractional marketing lead is a senior marketer who works for the company for part of the week, owns the plan and manages the agencies. It is a sensible bridge in the first year, when the job is still being defined and the spend does not yet justify a full-time senior salary. Four questions separate a workable arrangement from a nominal one:

• How many days a week, and on which days? A lead with no fixed time becomes one more inbox.
• How many other clients, and do any compete? Capacity is the constraint that fails first.
• Who covers when they are away? Launches and Tết do not wait.
• How are they paid? If any part of their income comes from the agencies or platforms they recommend, they cannot also be the person who checks those suppliers.

A full in-house team is rarely the right starting point for an entrant. It makes sense once volume is steady and the channels that work are known. The first roles worth bringing in-house are usually the ones closest to the customer: answering chats and running marketplace stores.

Running it from head office: why it looks cheapest and costs most

Running Vietnam directly from head office looks attractive: no new contracts, full control, a team that already knows the brand. It is usually the weakest of the four models, for reasons that compound.

The execution happens in a language the team does not work in, so creative arrives looking imported. Nobody answers customer chats at nine in the evening local time. Marketplace stores need local operation. And head office sees only the platform dashboards, which in this market miss a large share of what happens.

The largest cost is less visible: how slowly decisions move. An illustrative calculation shows why. Suppose a creative test needs four decisions: approve the brief, approve the creative, approve the launch, and decide what to keep. Made in Vietnam, each decision takes a day at most. Made from a head office five or six hours behind, each waits for an overnight round trip and then for the right person to be free.

The head-office version learns half as much in the same quarter, on the same budget. Over a first year that is the difference between a plan that has been corrected a dozen times and one that has been corrected six.

There is one case where the model works: business-to-business companies selling to a small number of accounts, where buyers read English, and where the plan runs on search and LinkedIn rather than marketplaces and chat.

Choosing by stage, not by preference

The right combination changes as the market matures, and the most common mistake is choosing the year-three model in year one.

At every stage the checking job stays separate from the people whose work is being checked. That is the one constant across the table.

The first decisions of an entry plan, including account ownership and how results will be recorded, are covered under market entry. The operating model belongs in the same conversation, because it decides who will carry those decisions out.

One first-year arrangement, written down

For most entrants in the testing stage, the combination fits on one page. The names will differ from company to company; the division of jobs should not.

Two details make the arrangement hold.

The threshold. The decision owner needs to be able to move a defined share of the budget, and stop a failing test, without asking head office. Without that, every decision joins the approval queue described above, and the model quietly turns back into head office running it directly.

The change log. Every budget move and every test decision gets a date and a one-line reason. It costs a few minutes a week. Without it, the monthly check can only compare results with people's recollections of what changed, and recollections always favour whoever is telling the story.

Five questions that settle most of it

Each of these has a short answer, and together they usually point to one combination:

• How much of the first-year spend will go to marketplaces and creators, and how much to international ad platforms? The first favours a local agency, the second a regional one.
• Who will answer customer messages in Vietnamese at nine in the evening? If the answer is nobody, the plan has a gap no agency choice will close.
• Who at head office approves, and how fast can they do it? If the answer is a weekly meeting, the approval loop needs to move closer to the market.
• Who will check the numbers, and is anyone paying them whose numbers they check?
• What must the group report contain, and can a local partner produce it? If not, someone has to translate formats every month, and that someone should be named now.

Where this work stops

MWY is not an agency and not a fractional marketing lead. It does not run campaigns, manage advertising accounts, make decisions on the company's behalf or recruit staff, and it takes no commission from agencies or platforms.

What MWY holds is the checking job, which every model in this article leaves open: setting the criteria before the model is chosen, shortlisting three vetted agencies when one is needed, reading the proposals, reviewing the contract, then reading the numbers each month against the company's own records and reporting to head office in English. That is Vietnam Digital Marketing Advisory & Oversight.

Choosing the operating model as part of a wider entry plan, alongside channels, sequence and budget, is part of Go-to-Market Strategy.