Our approach to analysis
We see digital analytics as a broad church encompassing quantitative and qualitative data.
Quantitative data is broadly used to describe a situation or predict an outcome while qualitative data is used to understand the more foundational aspects of why certain behaviour is exhibited.
There are three central pillars:
data analysis: split testing, correlation, metric selection, P values, qualitative vs quantitative...
data collection: server side tagging, privacy, user consent, ITP, cross device, marketing attribution, voice of customer, question framing,...
tech stack: BigQuery, SQL, Python, Looker, GDS, GA4, [Google] Tag Manager,...
There is something else that underpins all of the above, human decision-making bias.
As analysts, data scientists, managers, directors etc, we all exhibit cognitive bias in our decision-making. It's inescapable, and it is the single most likely attribute that can lead us to make poor decisions.
Confirmation bias is common in much of what we do as stakeholders.
These biases are often interlinked. Being aware of them can help us mitigate their impact and lead to better decision making based on currently available evidence.