Sales Conversion Optimization

The purpose of this project is to analyze the performance of Facebook ad campaigns and identify key variables that may have influenced the outcomes.

Data

The data contains Facebook ads data and comes from a business whose origin is anonymous.

Limitations

1. Anonymized Data: The origin of the data is anonymized, and we are not aware of the nature of the company, the product and the timeline of the campaigns.
2. Lack of Key Data: The data related to other ad campaigns and revenue is missing, making it challenging to draw comprehensive conclusions.
3. Data Imbalance: There are 3 campaigns available in the data, however the data is not balanced across them.
   That saying, we cannot meaningfully compare them.

We cannot answer questions such as:

• Was the campaign a success or failure?
• In which campaigns should we invest?
• Who should we target with which campaign? Thus, it is hard to measure the efficacy of the Facebook campaigns.

Methodology

This a pure deep-dive analysis into the data. Considering the limitations in the data
it would not be reasonable to apply any clustering or make predictions and make concrete recommendations.

Namely, we look at how variables interact with each other, such as clicks vs.impressions, gender and age vs. conversions or clicks.
These main questions are raised throughout the analysis:

• How good are the conversions?
• Does gender impact conversion?
• Does age impact conversion?

Key Findings

From the data observed these conclusions can be made:

• Conversion Rates: Conversion rates are higher than the average Facebook conversion rates.
• Purchases: In about 30% of cases people buy the product after enquiring about it.
• Gender Differences: Women tend to click more on ads than men, however, they buy less products across all present age groups, and that results in higher conversion costs.

To gain a deeper understanding of campaign performance and make more informed decisions, it's essential to consider the following data points for future analysis:

• user IDs (to facilitate retargeting and re-engagement strategies, e.g. for reselling or cross-selling)
• click and conversion dates (to assess campaigns performance over time)
• revenue (to derive ROAS and measure campaign efficiency)

Also, would be beneficial to know more about the targeting strategy implemented by Marketing.
This way, assess how much of the observed results are intentional rather than random, for example, whether specific age groups were intentionally targeted.