Instead of ignoring this new paradigm shift, it is instead crucial that you switch your focus to understanding what is happening in the marketing industry. What effects will it have on your business and, most importantly, how can you reap the rewards of your two years of struggle by quickly switching priorities and gaining a solid competitive advantage?
And, for those of you who have not yet taken any actions to “avoid data silos in marketing” – this is your chance to catch up!
What paradigm shift are we referring to? We’ll do a deep dive soon, but first, let’s conclude that the following tactics, that have likely been a part of your marketing strategy until now, will not be the way to go:
- Creating rule based segments based on basic heuristics rather than robust logic or modelling, and sharing these segments across channels for activation
- Striving to understand how every single customer behaves across every single touchpoint and channel
- Focusing heavily on building and sustaining your own ecosystem around third-party data
- Collecting as much data as possible about the customer without informing them or getting their consent
- Personalising the website experience for every single customer using cookies
- Relying solely on path based attribution using cookies without any additional logic
Back to the paradigm shift. Privacy regulations (GDPR, ePrivacy, CCPA), technology changes (ITP, privacy browsers, ad blockers), and trends in user perception have made it clear that the future for digital marketing will require that we think about data differently. The numbers we used to rely on are not trustworthy anymore. Instead, the questions we should be asking ourselves are:
- How can we drive better results with less data?
- How do browser developments like ITP affect us?
- How will privacy developments affect marketing evaluation?
- And most important, what should you do to turn panic into a competitive advantage?
Let’s dig into a bit more detail. The illustration below (figure 1) shows what one of many paths to purchase for a specific customer could look like today.
Two years ago, if you did a good job of collecting as much data as possible, again often without letting the customer know, your path that served as input data for the attribution algorithm could look something like illustrated in figure 2. And the output (distribution of revenue) could be shared across the channels to assign value to their contribution to the final conversion.
Unfortunately, the figure 3 below, is a better representation of the situation we face today. Starting with the mid to upper funnel, interactions are much more challenging to capture in a large part due to ITP (Intelligent Tracking Prevention) . The reason is that these paths depend heavily on third party cookies which are now completely blocked by Safari. Given that Safari represents 20-30% of all users, this is a red alert.
But that is not enough. GDPR has and will make the situation even more challenging. According to the European court of justice, active consent is needed for tracking. No consent means no tracking, which means that even the last interaction will be impossible to measure. If you agree to last click attribution being a bad oversimplification of reality, how about no attribution at all?
Given the realities that we now face, it’s rather clear why relying solely on cookie-based interaction paths without any additional logic is a very bad idea. We’ll soon get into what you might consider doing instead, but for the sake of clarity, let’s first revisit the other tactics that are not recommended.
1. Creating rule based segments based on basic heuristics rather than robust logic or modelling, and sharing these segments across channels for activation
This is by far the most common tactic for a “unified customer journey project”. The vision of having one single place where people without technical competence can build segments of customers and potential customers, which can be activated in marketing channels, is compelling. However, even before the paradigm shift, this tactic was rather dangerous. The reason is that you are trying to build a solution that fits your company and people working there, rather than the strategy that would benefit your business and customers the most.
So, let’s break it down.
By using a basic interface to build segments, you are encouraging the use of basic heuristics and intuition rather than the type of robust algorithmic analysis that machines can do more effectively than humans. Even before the paradigm shift, too much focus was placed on creating too many segments (and falling into the curse of dimensionality). Instead, time should have been spent on creating the richest and most actionable datasets. One example is the over-dependence of segment building based on pixel data that does not properly represent cross device behaviour. Instead, it would make sense to put more effort into building trust with customers and using 1st party data to enable better cross device insights.
The downside was already a fact before the paradigm shift. But the situation is much worse now. Because the cookie data we collect is much less reliable we need to spend much more time collecting richer data. It is likely that we also need to model the data before we even think about creating segments, to make sure that it is representative. We have a solid base to work from, applying a technique such as a prediction model for churn will be more impactful and provide more long-term value than having humans build rules based on gut-feeling or flawed indicators. Finally, it is crucial to have in mind that the ad platforms are taking a lot of actions to mitigate the impact of the paradigm shift. This means you will have to be an expert in the ad platforms or collaborate with experts to determine the smartest way to segment data and share it with the platform in the most efficient way. This means applying a one size fits all approach when creating segments has become an even worse idea.
2. Striving to understand how every single customer behaves across every single touchpoint and channel
By now, it has become evident that this is a utopia; it is simply not possible to achieve. But it is also essential to question what the purpose of doing so would be. Let’s say you invest a lot of time and resources to succeed. You end up with fancy customer profiles, collecting tons of data on every single customer across every channel. Apart from the fact that your customers most likely will not be impressed by the profligate data collection practices, what will you do with the information? 80% of companies fail at the activation phase If you are one of the few who succeed in activating the data, is it the best outcome to be able to show the customer the same ads across all channels? Even if some have succeeded in this, the adoption of ad blockers shows that customers are not impressed.
3. Focusing heavily on building and sustaining your own ecosystem around third-party data
The third-party data market to create audiences via platforms such as DMPs has almost completely disappeared with the introduction of GDPR. The use of third-party data is now mostly limited to the ad platforms and DSPs. This means that there is a need to accept a more fragmented approach to managing audiences. The companies who succeed here are the ones who find smart ways of utilising their first-party information in combination with the ad giants’ data.
4. Collecting as much data as possible about the customer without informing them or getting their consent
Given current privacy trends, a strategy of unfettered collection of data carries with it a serious risk of losing customers’ trust. Privacy breaches can result in large fines and significant brand damage. On the other hand, allowing legal teams to set the agenda entirely for data collection brings with it the risk that the value of data and the right way to communicate with customers could be missed in the implementation of privacy measures.
5. Personalising the website experience for every single customer using cookies
Web personalisation based solely on cookies is already flawed. One obvious drawback is that, in many cases historical behaviour will be forgotten with the deletion of cookies, making the potential for a personalised experience based on previous behaviour very limited. Another significant consequence is the risk of mixing up users in the test and control groups. In other words, the quality of the experiment results could be very doubtful.
I get it, but what should I do?
If you have read this far, you have probably confirmed what you already knew. Or, you’ve learnt that your marketing is in need of major strategic change. Now it is time to think about what to do instead. Should modern marketing give up? What is data-driven marketing in the 2020’s? In the following blog posts in this series we will go into what you should do in order to stay ahead of the competition. In the second post in the series we are sharing our thoughts on how to build a modern marketing data warehouse.