Data Restart 2023 did not disappoint us again, thanks to packed talks, great organisation of the whole event, excellent food and a super venue. The event took place, as usual, at the Dlabačov Cinema, just like the previous Restarts, and it is a very interesting place for conferences like this.
So if you did not attend Data Restart, either offline or online, you can read our short report. We would like to present the experiences, insights and inspiration we are taking away from the event. It is all thanks to experienced professionals who spoke at the event and who gained their experience in practice. If you are interested in the exact structure of the individual talks, you can download them for free on the Data Restart website.
Keynote: It Is Time to Start Using Data – Jan Tichý
The main point of the talk was the planned changes in the area of cookies, including the switching off of third-party cookies. This change has a major impact on online advertising and company strategies. Together with other agencies we agreed that clients often do not understand the importance of cookies, and it is up to us to explain to them why it is key for them. In the future it may also happen that their accounts in advertising networks (FB, Seznam.cz, …) will start being switched off.
Pavel Jašek from Google also revealed that Google wants to look into this more deeply. Google will become more active in checking data flows without consent. What form this will take nobody knows yet, but notifications will be coming. What we can do for now is certainly to keep saying, recommending and reminding clients of the importance of cookies, and have everything prepared in GTM.
How to Run a Company Through Data – Michal Buzek
In his talk Michal Buzek stressed the importance of combining strategy “something must come from above, something from below”. Some things need to be devised under leadership, but you also need active people and teams who will work on specific tasks. That is key to success.
We must also remember to set metrics and key performance indicators (KPIs) so that we can evaluate effectiveness. Joint meetings and shared results and KPIs are also important so that teams know what they want to achieve. Communication between teams is therefore essential, especially when analysing data. If, for example, the cost per click (CPC) in PPC goes up but purchases stay the same, joint communication with the analyst is key.
Sometimes it is necessary to think about the performance of individual channels. You can try cutting the budget of a particular channel, and if the data and purchases do not change, the channel was probably unnecessary.
The idea that data will reduce communication between teams is wrong. The more data, the more communication has to work so that everyone interprets and fully understands the data.
Customer Journey as a Way Out of the Darkness of Bad Company Decisions and Mistakes – Radek Hudák
The talk was a combination of UX and a data-driven approach. The “Out of Dark” tool illuminates the customer journey, including the numbers, which is a challenge because sophisticated tools are missing. They run workshops where they help companies get their bearings on who their customer actually is.
It is key to distinguish the approach to B2B and B2C customers, and the price of PPC advertising has grown by 30% over the last 5 years.
From January 2024 a free two-month trial version of the Out of Dark tool is available.
Looker Studio Pro vs Looker – Jan Šmerhovský
Looker Studio Pro has recently appeared. It is an advanced version of Looker Data Studio intended for enterprise purposes. This version offers organisations ownership, provides support (SLA), allows data to be organised in folders and runs on Google Cloud. Looker Studio Pro is a combination of a data warehouse, artificial intelligence and visualisation, which is ideal for creating a complex data ecosystem in the cloud.
For us, the table from the talk that clarifies the differences between Looker and Looker Studio Pro was very useful. It helps to choose the right tool for specific needs and deals with the question of what each tool is ideal for.
Attribution and Incrementality – Why, How and When? – Adam Šilhan
Adam, who works as a Google trainer, dealt with the topics of attribution and incrementality. Attribution determines the sources of revenue, which can be demanding. The current trend is data-driven attribution, which is hard for a person to understand. Last-click modelling, which is easier to understand, was discussed too.
Adam described incrementality as a phenomenon that tries to statistically describe a change in data depending on, for example, switching a campaign or channel off or on. This does not pay off if we run experiments that are below 10% of turnover.
These topics are current and are also called, for example, Marketing Mix Modelling.
Singularity in the Context of the Customer – Šárka Korychová
The talk focused on tracking customer activities and highlighted the challenge of identifying individual customers. It is common for one account to be shared by several users or to be used to buy for others. So behind one email/account there is not always one and the same customer; there may be, for example, 10 people. The way to solve this problem differs for every company and there is no single general solution. Šárka, however, gave one example of a solution at the company O2: there is a contact person under whom there are another X users/numbers, internet and so on. All of these services are paired with the basic contact number, from which they find out more detailed information about the other users.
Gen AI & ML in BigQuery (and Vice Versa) – Ivan Kutil
Ivan from AppSatori is active in selling Google Cloud and all its other modules. He talked about how Google Cloud has adopted AI and that Duet AI can now be bought as part of Workspace. It helps with text generation in various parts of Workspace. The service costs 30 euros per user.
We found the use of AI in Google Cloud very useful. This function automatically writes SQL queries and is trained on support articles. It can either fix, complete or even create queries based on users’ questions. So you can ask in plain language, for example, for orders by date, and the query will be written. Because it is trained on Google Help, the query will be written correctly.
Ivan also presented Vertex AI, a platform for training models and generating text and images at the Google Cloud level. It makes it possible to create AI models and generate content using advanced technologies.
Debugging Google Analytics 4 in Apps – Stanislav Račinský
Stanislav dealt with the question of debugging Google Analytics 4 in apps outside the browser. The limited options for tracking user interactions in apps lead us to use the Measurement Protocol and the Web&App version. Another option is to examine apps using the Firebase SDK, which also integrates mobile containers of Google Tag Manager. He showed us debugging of specific places in the Firebase/GTM system and examples of errors that can be detected at a given point.
How to Increase Customer Satisfaction and Improve Business Performance Through User Testing – Adam Rakus
Adam focused on the need to do things well and properly so that people recommend us more. Thanks to recommendations it is possible to keep growing and to get more enquiries and good reviews, and people will simply choose you more often.
Working with Data in 2023 – Are You Keeping Up? – Petra Dolejšová
Petra focused on the upcoming events of 2024, above all further changes in the area of cookies and the restriction of US tools in the EU. She discussed gatekeepers and warned against merging data from different sources. She also shared ten tips for checking cookies, which she published on her LinkedIn profile. Petra once again did not disappoint with her entertaining talk and offered an overview of current trends in working with data.
From Theory to Practice: Effective Work with Dataform – Viet Anh Chu and Jakub Kříž
The talk focused on Dataform and its use compared with alternatives such as DBT, which we use. Dataform is part of Google Cloud and is a plugin for GA4 that solves the flattening of datasets. The talk was about how they developed this plugin, why Dataform is good and that they want to keep developing it.
10 Tips for Working with BigQuery – Václav Ráš
Václav is deeply immersed in BigQuery and showed us 10 sensible tips on what to do and what not to do in BigQuery. These 10 tips were very useful for us, and we gladly took inspiration from them and applied it in practice. So we definitely recommend downloading his presentation and going through the points at your leisure.