The big terms of our time are without a doubt Big Data, AI and Machine Learning. Many would like to get a piece of the big data cake, align their own business data-driven and thereby use the methods of machine learning. They want to process large amounts of data in the cloud and take advantage of artificial intelligence and chatbots to tap new, unknown potential.

However, with all the big words, which are usually associated with few concrete ideas on how to best use them in the business context, the most important basics are often forgotten. The assumption that data automatically adds value with a trusted vendor’s tool solution is unrealistic. In my opinion, three basic things are needed to work successfully with data:

1. The tools of the trade: methodological knowledge about dealing with data
2. Knowledge of the domain: In what context do the data occur and how are they interpreted?
3. Targets: According to which target values do you optimize? What is important to have more success?

Goals are the fundamental basis for everything that follows. Every company sets itself goals such as increasing sales and growth. To ensure that these targets can be reasonably measured, incoming orders, sales and profits are documented and evaluated by Controlling.

In this context, the objective works very well in most companies. In the online sector, positive examples can be found above all in the e-commerce sector. This is where focused, data-driven work takes place, and there is a very precise idea of monetary goals.

However, if we look at digital environments beyond e-commerce, we end up in the area of content or corporate websites. There, websites or apps usually exist as an addition to the well-functioning offline business. Since digital products and services do not yet contribute to the critical part of business success in every company, objectives and a clear orientation are usually neglected here. But without target values, one cannot actively control a change process.

At this point, it should be carefully considered how, for example, a digital presence can expand the offline business in order to create added value. Which currency is valuable to us? These could be for example leads, newsletter registrations, downloads or constantly returning visitors.

It is not an easy task to think about the meaning of the provided platform and to define it in measurable goals. In addition, these measurable objectives must be accompanied by target values in order to provide a meaningful analysis.

Goals: the key to success

Clear goals with key figures and target values form the basis for data-driven work. Let’s take as an example the creation of a new blog post for this blog or Evesham SEO. We must have clearly defined why this blog exists and therefore also which goals it pursues and with which concrete target values we measure which key figures. At best, this should be part of the content strategy. Only with goals can we begin to evaluate target groups, adapt data sources and media budgets, test new formats, text lengths and images in a structured way and thus continuously improve ourselves.

This requires, as described above, objectives, domain knowledge and methodological knowledge of data structuring.

1. Objectives: to know whether the content created contributes to the success of the project
2. knowledge of the domain: to formulate test hypotheses, for example
3. Methodological knowledge about data: to cluster data, for example, and to evaluate these clusters as well.

The bounce rate is not a KPI

Overarching goals should be defined in discussions or workshops with product owners or product managers. The central question in these conversations must be: “Why does the website/app/campaign exist? This question does not sound complicated, but it requires some critical thinking and intensive discussion. Overarching objectives should be SMART.

You should now define more specific goals for each higher-level objective. These goals represent more specific strategies that are used to achieve an overarching goal. An objective can have multiple goals.

The next step is to define Key Performance Indicators (KPIs) for the goals. These key figures help to evaluate performance on the basis of the goals set. It is important to exchange information with the data managers in order to define the KPIs together. The core content of this step is to make the existing goals measurable on an abstract, technical level and to answer the question “What key figures can I use to map my goals? In addition to knowledge of the technical product environment, tool knowledge is also required to determine measurable KPIs. However, not every key figure is suitable as a KPI. The bounce rate is a number that cannot be considered a KPI. It is well suited as a supporting key figure, for example in the area of landing pages, but it is not meaningful enough to evaluate corporate success. This is mainly due to the way in which a bounce is measured. This is measured if the user’s visit includes only one hit, i.e. only one piece of information (e.g. a landing page) is sent to the Analytics server and no further information is transmitted during the user’s session. So a single hit decides whether the visit is counted as a bounce or not. In my opinion, this information content is not sufficient to be considered a KPI.

Target values must be set meaningfully in a context

The next step is to define targets for the individual KPIs. These are very important to determine success or failure. Let’s assume, for example, that a lead campaign generated 80 leads. Is that good or bad? This question can be answered with previously defined target values.

You can derive target values from historical data. (How well have lead campaigns performed with what budget in the past?) If there is no historical data, one should try to define target values using assumptions. An experienced partner can help.

The result should be a table that has a similar form to the following one. This was filled with an example strand for an objective.

This process requires some effort. However, it is essential for successful data-driven work to write down these goals and then optimize them in the ongoing process. Of course, objectives and target values are not set for eternity but should be reviewed at regular intervals and adjusted if necessary.

With this foundation, we can begin to align our tracking specifically to these goals, manage campaigns accordingly, and apply analytical methods of machine learning to optimize metrics.