Brand Safety in Vietnam’s Rapidly Expanding Digital Market 

D igital products are always aimed at making the life of marketers a little easier, right? When the first-ever banner ads emerged in the 90s, we were blown away by the ability to link the advert directly to a webpage.

Ad networks in the 00s then made it possible to target the message to unlimited audiences across thousands of websites, and algorithms of the 10s made them targetable and trackable. However, as the amount of content being produced and consumed has grown dramatically, it also became pretty much impossible to maintain 100% control over your ad campaign.

With the rise of social networks and UGC platforms, there is also constant attention on brands’ communications from all kinds of communities and authorities, which has led to multiple PR crises’ across the globe, including , accusing brands of supporting all sorts of illegal or inappropriate activities.

This has mostly come from placing ads in UGC videos. It has happened several times in Vietnam; brands’ adverts have been caught on anti-government video clips and because of these brands, agencies, and publishers have become very concerned about the quality of ad placement that they are buying, planning, and selling to each other.

Given the fact that adtech is an extremely complex industry, it took a while before we understood the most basic definition of what a high-quality digital ad is. The best definition I have seen in Vietnam was elucidated at the recent conference: “An ad should be seen by a human, in the assigned demographic, and in a safe and suitable environment.” Missing even one of those pillars can easily lead to certain financial, legal, and reputational risks, he said.

This definition is perfect in its simplicity, and at the same time it raises a reasonable question – what did we consider to be the definition of a “quality ad” before? What were we thinking? More questions spring to mind if we actually break this down into the keywords: seen, assigned, safe, and suitable. While viewability and fraud remain mostly technical issues, defining safety and suitability is a hard nut to crack.

The approach here, from what we see from both agencies and clients, is a long way off becoming holistic. Some advertisers work by applying a “dirty dozen” filtering system, and some prefer to go the extra mile when defining what a “suitable environment” is.

The goal for both here is to find the golden middle between being completely safe and to perform as the business’ KPI require. On the one side, is the option of not using digital marketing as a channel at all; it is the only surefire way to secure 100% brand safety, no matter how absurd this sounds in the 21st century. On the other – you can choose to let your freak flag fly, with no safety regulations at all. Obviously, not an option for most brands as it contradicts the very nature of branding.

Anything in between is solely based on subjective opinions of certain people, no matter whether they belong to the supply or demand side of the market. Raising awareness of brand safety issues, in reality, has led to multiple misconceptions through the whole client-agency-adtech-publisher chain.

In the era of total AI, automation, and the decentralization of everything, people in the ad industry have had to become more connected than ever before just to ensure all parts of the ecosystem are reading from the same page.

According to IAB India and SEA, global brands in the region apply 4 key tools to secure their safety and suitability needs: blacklists, whitelists, keyword blocking, and 3rd party verifications. From our experience, it is the keyword blocking that raises the biggest concern and is the most controversial part of the package. It’s one thing when you remove “massacre” or “violent death” and it’s a whole other thing when it comes to “girls” or “angry”. Modern tech is pretty advanced and can be very effective in identifying what’s wrong and what ‘s right when the matter is obvious, but sometimes it goes much deeper than that.

Just a month ago, we were discussing the concept of brand safety with some colleagues at Coc Coc, particularly trying to identify if one of the articles on the life of a celebrity that appeared in our browser newsfeed was safe, and if yes, to what extent. Votes we split down the middle, but when we clicked on the tile, we saw an article with a bunch of creatives from Mercedes, Japan Airlines, and some baby food on the page. Score.

In the era of total AI, automation, and the decentralization of everything, people in the ad industry have had to become more connected than ever before just to ensure all parts of the ecosystem are reading from the same page.

According to the tech verifications in place, those companies were clearly using the page and it was therefore completely safe for a premium product, an airline company, and an FMCG brand (and those are the most concerned with safety, by the way). Needless to say, the content was harmless for all of them. But the self-censorship and bias, a natural part of human judgment, led some of us to the conclusion that the article was potentially dangerous.

What can the market do to ensure brands are safe while assuring campaigns remain effective?

The answer is still yet to be found, but what is clear now is that we can’t fully rely only on tech or only on manual filtering. It’s the combination of both that we are seeing to be the most effective in ensuring brand safety.

For example, at Coc Coc we apply a 3-layer safety check:

1. The browser newsfeed is only supplied with content from a little over 300 of the most reliable Vietnamese media outlets. By default, it eliminates the risk of placing an ad near anti-government sentiment, hate speech, and all other forms of illegal content that can compromise the campaign in a UGC-environment.

The articles are suggested to users based on their interests and preferences, which are extracted from the way each of them interacts with our products. Powered by AI from the Russian tech company Yandex, this feed is a source of news millions of Coc Coc users.

2. The next step we take is applying keyword filters and automatic picture analysis. As mentioned below, technologies are a great help in most cases, so here we are already pretty certain that the feed is sustainably safe. Tech is continuously evolving to tackle more and more complicated judgments.

3. Manual filtering for controversial cases is the final stage, where we pick up the last pieces of content that may be considered unsuitable for advertising to be associated with. None of the content pieces on this are specifically unsafe, but they can simply be deemed “unpleasant”.

Simply put, algorithms are the driving force for media consumption, but people trust human opinion more than a machine, so the combination of both computer vision and collaborative filtering is the main goal for publishers in the country now. This requires some serious investment of resources but certainly offers the best possible result for advertisers.

Safety or performance?

With the fast-growing digital ad spend in Vietnam, a global agenda for the digital environment also enters the market. Professionals across the globe have been sold the idea that reaching brand safety parameters always leads to a decrease in reach. of the Sizmek last year’s surveys show a significant amount of marketers, at 64%, saying that achieving brand safety negatively affects campaign reach.

Due to blacklisting or whitelisting, marketers decrease the number of potential domains for their campaigns, which inevitably leads to a loss in the potential reach. However, at the same time, extensive work with multiple inventory suppliers and premium publishers may reduce this risk to a minimum. Another issue is keyword filtering, which was touched on earlier: going too far with excluding words vaguely related to something potentially not good, will definitely lead to the loss of perfectly good inventory.

It requires courage when it comes to reducing self-censorship, but with the list of trusted publishers and safe-proven formats, this is now a much easier thing to do. From my point of view, a high level of self-censorship is one of the things that differs between Vietnam’s marketers and the ones in western or even other Asian countries. Please, don’t get me wrong; this is a natural thing to have a desire to protect the brand, but some keyword lists just make me wonder – who are they intend to protect? The brand or the marketer, and from what exactly?

Nevertheless, the choice between brand safety and reach and performance can also be a false one. As a marketer, you probably know the environments that are safe and those which you consider “profitable/performing”, but you still use both to balance the budget and KPIs regardless of the risks. The takeaway here is to separate the results from those tribes and take a deeper dive into what’s going on with those clicks after they are delivered.

Most likely, safe placements with premium publishers on the desktop will perform better on the sale-through-rate or whatever else is happening in your brand’s digital environment where the audience is directed, while “performing” sources may simply turn up a waste of budget.


Article by channel:

Read more articles tagged: Brand