First Context and Steps of Data Driven Decision Making. Considering High Standards of Individual Data Privacy — an Introduction

Note: this article was originally published by the author, Tim Hammermann, on Medium, on 26. October 2020.

Online presence cannot be circumvented. Growth with undefined boundaries.

The Internet is still changing the way we all behave and work at breakneck speed. The increase in global internet traffic over the last decades - with Covid as the latest stepping stone - seems to know no limits. Approximately 4.1-4.5 billion people now use the Internet. There are about 360 million "Top Level Domains" and more than 1.74 billion websites . In 2020, it is estimated that about 174 exabytes, i.e. 174 billion gigabytes, will be transferred via the Internet every month. In 2022 it should be 273 exabytes (+ approx. 57%).

The total amount of international online trade is estimated at around 4 trillion in 2020. But the value added by the Internet as a whole is hardly ever really quantifiable. In addition to online shops, there are millions of websites that prove their value in the free presentation of art, knowledge, facts, etc. Associated "offline value creation" would also have to be taken into account.

Statistics. Analytics. Interaction. Know your visitors & numbers.

Web statistics, visitor analytics and visitor interaction should always be of particular importance for every website operator, whether large or small, profit-oriented or not, old or new, when it comes to strategic or fundamental decisions. Without this, you cannot know the behavior of your visitors or, hopefully, at some point, customers well enough to make the conversion experience — however individually defined — as strong as possible.

It is not just a matter of establishing superficial facts such as “I currently have 15% more or less visitors than last month”. It is about continuously and repeatedly answering a whole range of questions, which can vary greatly in their depth of answer:

• where did my visitors come from and why did they visit my site? Has there been any changes compared to last year? If so, why?
• how is my visitor traffic composed or structured?
• how do my visitors interact with my website and all its subpages — possibly divided up by target groups?
• how do my visitors convert to customers? Do you convert at all? Which actions and activities do I consider as “conversion” at all? Are there different conversions on my site and should they be weighted/preferred differently?
• are my visitors at all satisfied with the content they see? Do they find what they are looking for quickly and do they enjoy spending time on my site? If not, why do they end up there at all or how can I noticeably improve the visitor experience?
• To what extent have which marketing campaigns of mine produced which benefits and how can I measure them?
• what tools do I need to use to get all this information?
• and so many other questions.

If you have the right tools at hand, answering the questions, the resulting measures and observing subsequent improvements can even be particularly fun. As serious and important as the topic is for every website operator.

Aggregated data vs. (anonymous) individual data

Of course it is important to know how many (unique) visitors you have received on your website in total in a given period of time, how long their average session duration, bounce rate or number of page visits per visit was. Also, the basic composition of your visitors by country or channel origin, the distribution of operating systems, browsers or devices used, the combined distribution of attention within different pages (heat maps), basic conversion path analysis (conversion funnels) etc. are of course crucial knowledge assets. However, the focus must not be solely on the aggregated figures.

As a website operator, you should also provide (large) space for the interpretation and evaluation of individual “feedback” — whether “passive feedback” through session recordings and courses of action that can be seen to be completely uninfluenced by them. Or “active feedback”, through direct questioning of your website visitors in the form of surveys or votes themselves. In the past, it was often claimed that data is the new oil. This is of course open to dispute. Especially considering the — rightly so — ever-increasing individual data protection measures (and since the global economy is desperately searching for alternative energy sources). However, following the analogy, the oil must first be “found”. It must be correctly “tapped”, extracted and then refined as efficiently and effectively as possible. There must be similar processes for data. Data must be correctly identified, filtered, collected and exploited to generate real benefits. Online data is also particularly often of a human or personal nature. If the data is of personal nature, it is the property of each individual. This has never been any different, even if in the past, it has only been considered to a limited extent or by a few.

When is data personal?

In my view, this question cannot be answered unambiguously — it is in some respects a matter of interpretation. However, aspects such as (first) name, date of birth, place of birth, e-mail address, address, ethnicity, etc. are undeniably personal. For me personally, this also includes areas such as my website visit history or IP address. I don’t want third parties to be able to read out easily which pages I am navigating or have navigated in the past. Nor do I leave a note for every shop visitor, listing where I have been today or the days before. These data sets, of course, could determine how you see content on the pages you visited after or helps others to categorize you in certain interest groups. In general, thanks to current developments, the subject is receiving more attention.

What is aggregated or anonymous data?

Aggregated or anonymous data is used when it is no longer possible to draw conclusions about individuals based on the collection or visualization of the data. Simple examples are the presentation of the number of “visitors” over a certain period of time. Or averages of values such as “session duration”, “bounce rate”, “page visits per visit”, etc. Tools such as “heat maps” or “conversion funnels” are also aggregated visualizations of attention distributions or conversion successes, that do not allow any conclusions to be drawn about personal data or personalized behavior patterns. Even session recordings that reflect videos of individual visitor sessions do not allow the recording of personal data — entries in fields are filtered and other tabs or other screen contents are of course not recorded.

Data protection as one of the top priorities

As already mentioned, attempts are being made to put one’s own data back into the control of the owners, namely each individual (charming approaches can be found in concepts such as Prifina.com). Companies are paying much more attention to data protection issues (also evidenced by the increasing growth of platforms such as Datenschutzexperte.de) — currently still specifically in the EU, but other regions of the world will follow over the coming months and years. Cookies will be a thing of the past, and there already are available charming concepts such as Visitor Analytics, which collect all data exclusively without cookies. There is even a mode where you no longer have to tell your website visitors that you use Visitor Analytics and still collect all the data.

It is indeed astonishing how many tools and sites claim to only set cookies when you give your consent in the infamous and all too familiar “cookie banners”. However, if you actually check in parallel in your browser how many sites and tools set cookies before you have even clicked on any consent, frightening results are revealed. Of course, cookies per se are nothing really life-threatening. They simply install files on the hard disk of a PC without asking the visitor — this should not be possible.

While I am very sympathetic to the protection of personal data, I believe that we must also understand that usage data in anonymous or aggregated form, and of course those that are actively and voluntarily shared by visitors, can be of great benefit to all stakeholders.

When you go into a local shop, employees see your face, often you are still being watched by cameras and novel (i)beacon concepts can send you push messages on your mobile phone based on your position in the shop. As soon as you pay with a card, your name is registered. That’s the way it is and you get what you want in return. If you are still interested, you can ask a (known to you) employee who already knows what your tastes are, or take part in a customer survey.

In the web area, we also have the possibility to record and analyze (the behavior of) visitors — but much more anonymously and without personal data playing a (big) role. Of course, if the visitor explicitly agrees, further data can be collected. For example by actively participating in surveys or voting. From the collection of anonymous data records, website operators can gain decisive insights into the use of their website by second parties, which in turn allows them to significantly improve the visitor experience. Visitors suddenly reach their personal goal much faster (e.g. completing a purchase, adding an item to a wish list, etc.) or find relevant content much faster.

Data-based decision making

Now that a clearer distinction has been made between aggregated or anonymous data, which does not require the use of personal data, and personal data which should only be actively provided by a visitor, it is now possible to start collecting and evaluating the data. A detailed consideration of the procedure in this regard will be dealt with in one of the next blog entries I will make. It would simply go beyond the scope of this entry.

However, it is essential that as much data as possible is available and that it is useful for the “action area” envisaged. For example, other data is usually more relevant when defining measures to increase the number of visitors or the average length of sessions than when looking for ways to increase conversion rates. Emotions and one’s own convictions should be set aside in this process — usually the figures do not lie. Ultimately, the aim is to make the website as attractive as possible for the target groups. Not for yourself. As a rule, a decision based on hard numbers always produces better results than following your own “gut feeling”. Do you think 80% of your public does not convert because there is not enough content in the top sections of your website? Heatmaps and visualised click rates on element level will support or reject your thesis. There are dozens of such analysis approaches. Give it a try. It is worth it.