3 non-trends in performance marketing for 2021

Digital Marketing, Media & Advertising

It’s always the same story: as soon as the new year starts, we find ourselves immersed in a string of publications trying to predict the latest news and trends that will mark the following year, bombarding us with “7 things that will completely the change the way you work”, “5 changes that will revolutionise the market” or “13 social media trends you cannot miss”, generating an endless list of trends that are in fact, not trendy at all. 

Let’s go back to 2020 and analyse all the trends that were predicted at the beginning of the year, and we can count on one hand the ones that have actually transformed our understanding and approach to marketing. 

It’s true that COVID-19 has accelerated the superdigitalisation process of the already-changing digital marketing environment, bringing new hot topics into the table that marketers need to address urgently. Most probably, a list of predictable solutions and advice would help some of them calm down, especially in the current scenario of constant uncertainty. But, if the pandemic has taught us anything, it’s that we can’t take those lists as gospel for our new year challenges. 

This wave of change, which we at Good Rebels have been predicting since the beginning of the health crisis, is also sweeping through performance marketing. However, the truth is many of these changes have been “a thing” for a while now, such as the real application of artificial intelligence in campaign optimisation, the disappearance of third-party cookies in Google Chrome (the world’s most widely used browser) or the implementation of attribution models based on mathematical models.

Other relevant issues are the rise of ad fraud, the concerns about the quality of digital impressions, the intrusion of advertising into users’ private spheres and the increased relevance of first party cookies in terms of message personalisation and tracking.

These are not just additional bullet points on an ethereal list of trends that will never make it to our strategies. In fact, it’s quite the opposite: they are realities that any professional performance marketer will need to face in the short term (if they aren’t already doing so.) Marketers who address the above points will see the value of their digital business models increase. Do you dare to ride this wave of change?

How does artificial intelligence influence campaign optimisation?

When it comes to optimising paid media campaigns, we are constantly wondering what parameters to modify in order to reduce costs and, therefore, maximise results: A/B testing for creatives or call-to-action? Dynamic ads? Changing the structure of ad sets? Setting automatic signals to raise or lower bids? 

In the last years, there have been several experiments and findings that tried to answer all these questions. The Hagakure and Royal Bert methodologies, for example, seek to optimise Google Ads campaigns through a new way of organising keywords and ad groups. At Good Rebels, we work with the GFC method and the 3 I’s framework on a daily basis, trying to optimise social paid media campaigns through hypothesis models aimed at improving results. 

performance marketing

However, all these methods are time-consuming due to the complexity of the optimisation process itself, which will not get any easier in the near future, and the high volume of options and platforms available. If, on top of that, we also consider the high level of digital competition, the aggressiveness of supply and the low user loyalty of certain markets, the picture is definitely a complex one. 

But, actually, it can get even more complicated. And that’s where artificial intelligence comes in. Usually, we analyse the results of our campaigns through the advertising platforms themselves, or through the tools we use to manage them. But what if we extracted all the data and crossed-checked the different variables, instead of doing it marginally?

This is one of the many solutions that artificial intelligence brings to the table: being able to observe how each and every one of the variables established in an advertising platform affect the costs of our campaign. Furthermore, through mathematical models, we could predict the best possible combination of these parameters, in order to the highest volume of results at the lowest price. 

This non-trend is probably old news, but 2021 is the year in which we will assist to the realisation of its true potential. 

How will the market adapt to the disappearance of Google Chrome’s third party cookies?

Back in 2020, Google announced its intention to follow in the footsteps of other browsers, such as Safari and Firefox, by removing third-party cookies from Google Chrome. This may seem like a mere formality in terms of privacy, but it actually is one of the biggest paradigm shifts in digital advertising since the creation of the ad pixel. 

We are so sure of this because, a priori, this decision allows Google to prevent advertisers’ from “chasing” individual users when they browse the web. Consequently, it would limit brands’ ability to show users personalised ads in the name of internet privacy and anonymity.

However, it is important to clarify that the disappearance of cookies does not imply the end of customised digital advertising. What will disappear (or, at least, change) going forward is the technological process by which this customisation will be carried out. Therefore, we will also have to develop new ways of finding out which ads have led to a conversation.

Since the decision, Google has been working on an alternative solution: Privacy Sandbox. The idea, broadly speaking, is to stop working with individualised data and start targeting clusters of users with similar interests. This will respect the privacy of the individual user while still offering advertisers the ability to segment their campaigns and impact people who might be interested in their products or services.

What happens to web retargeting databases, then? Privacy Sandbox is also making progress in this direction, in order to continue offering advertisers the possibility of retargeting. However, it is not an easy task, as it directly clashes with the concern for privacy that motivated the decision. The truth is, little progress has been made in this area, although the possibility of creating “trusted servers” where data can be anonymised, is under discussion. However, for the moment, all we can do is wait for Google’s updates on their progress.  

Nevertheless, browsers were not the first to make a move in this regard: Apple has also limited the ability to track users, forcing platforms to ask them for permission before collecting and sharing their data for advertising purposes, thereby reducing Facebook’s ability to track users of iOS 14 or above.

More specifically, Facebook Ads’ attribution window has been reduced from 28 to 7 days post-click for iPhone users with this operating system. In other words, the platform now only tracks conversions for users who have purchased a product from the website seven days post clicking on an ad, whereas previously it was able to attribute conversions up to 28 days after the ad was clicked. This means that the conversions that Facebook is able to associate with ad campaigns will decrease significantly. 

This second non-trend is probably the one that worries the advertising community the most. Will users’ privacy finally be respected? Will our ability to customise digital ads decrease?

When will data-driven attribution models be given the importance they deserve?

Besides changes in cookie policy and user tracking, it is also crucial to highlight the importance of establishing an attribution model that shows, as accurately as possible, which digital platform had the most influence on the purchase decision process of a given user.

Lineal, positioning, last click and first click models were useful when we didn’t have the technology, nor the skills, to develop models that estimated the relevance of each platform, but they have now become obsolete. 

For several years now, it no longer makes sense to attribute all the importance of a conversion to the last platform on which an ad was shown before a user clicked on it and made a purchase, particularly not if several advertising tools were involved in the media plan. It would basically not be fair and, if we were to rely on this indicator, it could lead to wrong conclusions and sub-optimal decision making. 

Markov’s chain theory, which was developed in 1906, is one of the foundations of the technology developed for tracking users through the different advertisements they view in the digital environment. By applying this theory to attribution models, we are able to determine how important each of the advertising channels with which a user has interacted were in their final purchase decision.

The aim is to find the platform that had the most influence in the purchase decision, in order to increase its relevance in future media plans (say, on creative or budget terms). At the same time, this approach also allows us to dispute the attribution data displayed by Facebook or Google on their own platforms. 

Although the implementation of data-driven attribution models requires significant financial effort due to its complexity, it offers a straightforward return on investment, and has an impact on decisions has relevant as the selection of advertising platforms for media plans, the budget allocation per platform and target cluster, or the definition of synergies among platforms based on user behaviour analysis.

This non-trend can have the greatest impact on a brand’s return on investment, as it indirectly involves the two previously mentioned trends. 

These issues don’t just affect performance marketers. They also impact the work of a range of data, technology, analytics, creative and strategic experts who must be prepared, if they are not already, to understand and implement them in order to continue yielding the maximum benefit for their businesses, projects and clients. Because one thing’s for sure: these non-trends are here to stay, at least for a few years. 

But… if they are not trends, what do we call them, then? At Good Rebels, where we love to come up with names for things, we are certain that, at this point, we would call them RebelFacts. Because they are not abstract bullet points on a list. They are realities to which we must react as soon as possible. 

The time to act is now. If we don’t, we would have missed out on a huge opportunity to generate added value for our business, our brands and, above all, our clients. 

Menu