Data-driven digital marketing: top down or bottom up approach?

Data-driven digital marketing. Many organizations are working on it. No longer based on assumptions and gut feeling but on hard facts. Real-time insight into the success and grip to be able to adjust in time. But how do you get the right insights? Or more importantly, how do you know which KPIs are important? For the seasoned marketer a no-brainer. But make no mistake, many organizations still struggle with this.

There are roughly two ways you can start with data-driven digital marketing. The first is the top down approach. This one is preferred. The second is the bottom up approach. This one gives less grip and certainty and is more experimental. In many organizations you see that the latter is the default approach. And in both ways there is often an additional challenge: the data gap.

In this article, I explain both and indicate why the top down approach is preferred. I conclude with perhaps the most important challenge: closing the data gap.

The top down apporach

The essence of the top down approach is the traceability of efforts and budgets to top down KPIs. By having insight into these you demonstrate their value in the bigger picture. It provides direction and grip and helps to make choices. The top down approach starts from the strategic ambitions and objectives. Once these are clear, they can be translated into underlying objectives, which may or may not be for each department or team. This is often the tactical level. From there, another translation can be made to the operational level. Let me illustrate this with a simple example:

Imagine a strategic goal in which an organization wants to hire 10 new employees within a year to achieve organic growth. The HR and marketing departments have a role in this. Data shows that of every 100 resumes, an average of 10 interviews take place. And of those 10 interviews, an average of 1 person is eventually hired. This means that for 10 new employees, 100 interviews must take place and therefore 1,000 resumes must be brought in. For 1,000 resumes, ultimately 100,000 visitors are needed. The top down approach then looks like this in KPIs:

This top down approach provides concrete KPIs to steer by. From this, departments can then set out more concrete actions as well as measurement points. Marketing, for example, will be primarily responsible for bringing in 100,000 leads. They can now think about the channels and forms of communication to be used. KPIs can also be linked to this in order to measure success. For example, how many visitors are expected to come in via Google Advertising and at what budget. With this, you take the first steps in conversion attribution, i.e. the assignment of a source to an (online) conversion. This could look like this, for example:

The top down approach helps with structuring and quantifying KPIs and traceability. This allows, among other things, to measure the value of Google Advertising’s share. You can then also ask concrete questions such as “when will we know if someone who was hired came in through Google Advertising?”. And if that question cannot be answered, you can look for a solution to find out. This allows you to make timely adjustments, allocate budgets effectively and more efficiently, and most importantly, learn from them. It gives you real-time insight and grip into the success of a campaign so you can make adjustments. You are working in a goal-oriented and focused way. A characteristic of this approach is that you only measure what you need to be able to adjust.

The bottom up approach

The bottom up approach is the exact opposite. It feels more experimental and is based on assumptions. It lacks focus and purpose. What often characterizes this approach is that departments or teams are not driven by clear KPIs. A marketer who does not have KPIs for a campaign will never know if he is doing it right. If necessary, draw these up yourself. In the example of staff recruitment, you could say that at least half of the recruited staff came in through the online channels. Then you can start asking the questions, which will help set up your platforms and tools. For example “how do we know if someone got in touch with us through the digital channels?”.

In practice, people often look for insights they can measure. But a lot of those measurements do not tell you the things you need to know. Then you keep staring at numbers that do not help you.

One of the biggest risks with this approach is that you start measuring what you can measure instead of what you should measure. With this approach it is all the more important that you keep asking critical questions such as “what does this insight give us?” or more hypothetically “if this KPI develops from the worst case scenario, what would we do?”.

If you do not ask these questions then a proliferation of measuring points will arise. This risk only increases with tools like Google Analytics and platforms like Sitecore. The more you can measure, the more fun and interesting it seems to become. Appearances are deceiving. The more measurement points, the more confusing the whole data becomes. And remember that data analysis takes time and energy, which may defeat any purpose.

If you recognize this, it might be a good idea to have a conversation with management or the board of directors. Most organizations have strategic goals. If they are not clear you should ask yourself why. Ask them and if necessary translate them into your own KPI’s on which you will report pro-actively.

The biggest challenge: Close the data gap

One of the biggest challenges remains closing the data gap. Especially when online interactions turn into offline interactions. What do you need to do to link what happens offline to online behavior? You see this for example in the automotive industry where many people orientate themselves online and eventually buy or lease a car offline at a dealer. In that respect, Tesla did a smart thing by allowing people to buy cars exclusively online, as if you were ordering a book from amazon.com.

The success of campaigns is obviously about much more than measuring, but it does form the basis for making adjustments. So keep asking critical questions. What do we need to know? What insights do we already have and what not? What do we have to do to get them? And what happens if we can’t make adjustments? This will give you much more focus and help you set up measurement points and KPIs.