Practical Data Analytics Framework

I have been doing Data Analytics for some time now and I have taking notes of how I do it. This article outlines that framework and acts as a practical guide to follow.

I. Tracking Infrastructure

This is about setting up Google Analytics which is the main source of analytics. 

1. Account Setup. Creating accounts for Google Analytics (GA4), Tag Manager, Big Query. Linking GA4 with GBQ. Setting up filters, Cross domain tracking.

2. Basic Tracking. By default, GA4 starts tracking page views. 

3. Measurement Plan. We define the success of all our initiates based on achieving the overall business objectives. To measure such success, a measurement plan allows us to quantify and gauge our progress for each of the step in the funnel.

4. Advance Tracking. This includes Ecommerce Tracking using datalayers, tracking important clicks such as Call-to-Action buttons.

II. Data Warehouse

All data will be stored within a data warehouse where it will be able to interact and connect with each other. This eliminates data silos and helps supplement analysis.

1. Pre-process data. Getting raw data organized into readable and query friendly formats. Scheduled queries for repeated tasks.

2. Data cleaning. Transform data to make it cleaner. Remove missing values, correcting incorrect data.

3. Aggregate data. Extract data from clean data and summarizes metrics then merge with IDs.

4. API Rest. Automated scripts that extracts data from other sources and merge into main GA4 data.

III. Idea Creation

We uncover insights in data to try to understand customers and find opportunities to improve. To do so, we first find observations and formulate hypothesis around that. Then, we think of ways on how to use that information to our advantage by making site recommendations. Finally, we validate our ideas by conducting an AB test.

1. Observations. Seeing data helps us find relationships between variables that allows us to make assumptions and come up with feature recommendations.

2. Formulate Hypothesis. Here we come up with possible reasons or causes of the effects we see on our observations. 

3. Make Recommendation. Here we propose a solution or feature recommendation that addresses a need uncovered by the analysis.

4. Validate through AB Test. We validate our assumptions and observations through an AB test to be able to determine if what we have hypothesis is true and if our recommendation has actual value.

IV. Dashboards

Monitoring performance for marketing and site improvements. Evaluating AB tests results.

1. Summary. Contains important KPIs that summarizes the performance of your initiatives in one page.

2. Acquisition. All actions related to advertising any interactions outside the website. This report allows us to evaluate each marketing channel and initiatives and calculate the return on our investments.

3. Engagement. This report measures all the interaction within our site and how well each customer goes through each of our pages. The goal is to have users go as far as possible through the funnel into the checkout phase.

4. Conversion. This report gauges the effectiveness of the checkout process in converting users.

5. Adhoc Analysis. Exploratory analysis to address questions or investigations.