Advanced analysis life cycle

2023-06-13 | Article | Insights

This article aims to provide a deeper understanding of the typical steps involved in advanced analysis.

As visualised in the snail-shaped process below, advanced analysis begins with a research question or a hypothesis that drives the analysis. From there, the relevant available data sources and the most effective approach for answering the question are determined.

Then we come to the most challenging part of the cycle, the data preparation. In most cases, the data required to answer a research question comes from multiple sources, such as GA4 and DV360. Integrating these data sources in BigQuery can be challenging, requiring a deep understanding of the data to ensure that the data is accurate and complete. Another challenge of data preparation is cleaning the data itself. Data is often imperfect, with missing values, errors, and inconsistencies that can compromise the quality of the outcome of the analysis. Remember that the more effort and care you put into preparing and cleaning your data, the more accurate and robust your final results will be.

With the clean data in hand, the analysis itself can begin. This step can vary depending on the question in the first step but typically it involves slicing and dicing the data and applying statistical techniques to identify patterns and trends in the data. These findings are evaluated and interpreted to gain insights into the research question or hypothesis. The interpretation of results is a crucial aspect of advanced analysis, as it requires a deep business understanding and the used methodology itself.

Finally, the analysis result must be activated, usually either by creating a dashboard for questions that need to be answered on a regular basis or by deploying a model that will be updated regularly. It is very common that after the last step, follow-up questions are raised and the snail cycle could potentially start again.

At Digitl we are facing the exciting challenge of executing each step of this snail-shaped process in the most optimal way on a daily basis. We have successfully executed numerous advanced analysis projects. As a result, we have been able to enhance the capacity for data-driven decision-making for our clients, which you can learn more about in the following sections. Below, you will find a selection of some of these projects that resulted in significant improvements for our clients.

  • Data-driven strategies for user retention in app: The client offered a wide mixture of content and products to their users. As their App platform was only launched in September 2022, the client was curious to learn about the impact of content consumption within user journeys and its influence on retention. The analysis insights together with identified implementation improvements built the foundation for Highsnobiety’s 2023 strategy to significantly uplift user retention and e-commerce conversions KPIs. Read more here.
  • Data-driven approach to SA360 inventory management campaigns: As the client managed hundreds of products while dealing with a highly competitive landscape, speed was a crucial factor. Thus, the team sought ways to identify top-performer products with fast implementation methods to use in Search Ads 360 campaigns managed by their partner The Boutique Agency. With the new approach, the company was able to identify and start promoting top-performing products in a fast and highly reliable way. Read more here.
  • ROAS optimisation through frequency management in DV360 campaigns: To fully use the programmatic capabilities, the client sought methods to understand user behaviour after campaign exposure to better control contact frequency. The potential cost of user overexposure beyond the efficiency point should be reinvested in further campaign reach, ultimately increasing the campaign's ROAS. As a result, frequency settings were deducted resulting in 23% ROAS uplift for display campaigns. Read more here.

Do you need more Info?