We attended another MeasureCamp, this time in Vienna. We are bringing back to the Czech Republic new methods, great inspiration and important contacts, and we would like to share them with you.
MeasureCamp Vienna ran smoothly, like every previous MC we have been to. As usual there were several blocks of 30-minute talks in which the speakers touched on current topics in the data and web world. This time we were pleasantly surprised by the larger number of talk blocks, which gave us a wider choice of topics and speakers.
So let us look together at what caught our attention.
Why we track – Steen Rasmussen
Steen’s talk focused on measurement in general. Why do we measure, what is it for, what is happening with our campaigns, how can artificial intelligence help us, and so on. The talk nicely summed up things we already have established in our processes and clarified a few other topics for us.
For example, that when AI models data, it is able to calculate anything. That does not mean it is appropriate to use it every time, though, because in some cases automated systems are out of place. We must keep in mind that we work with people, whether they are clients or developers.
At the same time, you need to work with data as early as possible; we cannot let it lie fallow. We have to work with what is happening now and take into account the factors around us.
People have a natural tendency to reject the truth even when it is backed by data, so the way we word our results is also very important. What matters is to deliver clear outputs, propose a direction, put things into perspective and, as Steen said, “keep it simple”.
SGTM challenges – Michaela Linhart and Christian Kletzander
Michaela and Christian formulated their SGTM challenges talk as a workshop. They dealt with the challenges that can arise when setting up SGTM measurement. We work on server-side analytics in our projects every day, so the topic is no stranger to us.
At the workshop we therefore focused together on questions such as: what are the most common mistakes when setting up SGTM measurement, how to proceed, how to test that the measurement works correctly, where the servers will run, and more. Overall the talk was very detailed and suited to more advanced individuals who already have the basics of SGTM behind them.
Git for data Analyst
The talk was very interesting and focused on using Git in data analytics. Git is also part of our technology stack, so we were very pleased that a talk was dedicated to it.
The speaker framed the topic in terms of general sheets and tables. We, by contrast, use Git to store development versions and for backups, for example for DBT when processing data for BigQuery. Back to the talk. What advantage does Git bring for Sheets? We do not have to store thousands of files or search for them laboriously; instead we store the current versions in Git. The biggest benefit is therefore the overview.
Cookieless is bullsh*t – George
George framed his talk, titled Cookieless is bullsh*t, as a discussion. It all revolved around the idea that cookieless measurement brings no added data, operates on the edge of the law and is generally a pointless method. The people who came to the talk then argued why they think the method is beneficial and why they use it themselves in their work.
Spam traffic in GA4 – Matthias Hausdorf
Matthias’s talk described clearly and concisely how in the Google Analytics 4 settings we can filter out certain addresses, bots, payment gateways and much more, and where all of this is found in the settings. His talk was very detailed and these procedures will certainly come in handy in our practice.
Marketing Mix Modelling in a nutshell – Marek Kobulsky
A very interesting talk focused on Marketing Mix Modelling, the most up-to-date approach to evaluating campaigns. It went from the theoretical foundations of how MMM works and what its main benefit is to how it differs from attribution modelling, with several examples.
From the theoretical basis of how MMM works and how it differs from attribution modelling, Marek moved on to concrete examples. It is a pity there was not more time for such an interesting topic.
GA4 BigQuery sessionisation – Thomas Langnau
The last talk we managed to attend was by Thomas on GA4 BigQuery sessionisation. He dealt with the question of what format to send data to GA4 in, so that everything in BigQuery does not end up scattered. For example, if only numbers are sent to GA4, they are stored in BigQuery in a different format. As a result they appear in a different column in BigQuery, and an automatic query that counts on that one column returns many empty values.
The talk was rather technical. Thomas covered the data types that help us structure and split a query. He also introduced the basic splits and structures for when we need some data from GA4 and BigQuery.