The Brief
Day 3 shifted gears from Tableau to Power BI. We were provided with social media campaign performance data from RWFD, containing metrics designed to measure the effectiveness and impact of marketing campaigns across digital platforms.
The task was to build a KPI dashboard suitable for a senior marketing audience - specifically, a Head of Marketing who requires a yearly, high-level overview of campaign performance to support decision-making and reporting.
Requirements:
- Explore and model the data in Power BI
- Design a dashboard for executive-level stakeholders
- Focus on communicating campaign success clearly and efficiently
- Highlight key performance indicators relevant to marketing performance
- Allow quick assessment of what's working well and where performance is lagging
- Presentation at 3:30 PM
My Approach
Going into Day 3, I felt cautiously optimistic. While I know many people in the company aren't particularly fond of Power BI, I don't mind it at all. I saw this as a good opportunity to see how I really felt about tackling a project in Power BI, especially since we haven't touched it in over a month.
I also felt somewhat at ease with the subject matter. Before joining the Data School, I went through an interview process with a social media marketing start-up that involved becoming familiar with social media marketing terms, so the domain wasn't completely foreign to me.
Planning and Sketching
I spent a significant amount of time planning - probably about 2 hours. I wanted to know exactly what I wanted to build before jumping into Power BI. Yes, I had to sacrifice some building time to do this, but I knew it would help me be more focused during development. I'll get quicker with time and more confidence.
My strategy centered around meeting the three key requirements from the brief, but I was particularly focused on the last one: "allowing the stakeholder to quickly assess what is working well and where performance may be lagging."
Given that the audience is a Head of Marketing, I knew money would be a major focus. This influenced my choice of metrics - ROAS (Return on Ad Spend), conversion value, cost per click, and value per conversion all featured prominently.
I organised the dashboard into four clear sections:
- Campaign Performance - which campaigns drove results
- Platform Performance - where to allocate budget
- Target Audience Performance - which audiences to focus on
- Trends Over Time - how performance changed month by month
Final PowerBI Report
Technical Challenges
The main technical challenges were around formatting and date handling. Finding the specific format codes in Power BI requires searching through menus, and once I applied formatting to one chart, I needed to ensure consistency across all visuals.
I also ran into an issue with sorting months chronologically across years. Initially, my months sorted alphabetically rather than chronologically. I had to create a calculated column that combined year and month into a sortable number (e.g., 202401 for January 2024, 202501 for January 2025).
The DAX measures themselves were straightforward, ROAS, conversion rate, value per conversion, cost per conversion, but getting them formatted correctly for an executive audience took attention to detail.
Feedback
The feedback session went well. The dashboard successfully communicated the key insights, and the structure made it easy to assess performance across campaigns, platforms, and audiences.
Areas for improvement:
- Some axes needed adjustment for better readability
- I didn't need the slicers on the side as the charts could filter each other through cross-filtering
Personal Reflections
I'm happier with today's dashboard than the previous two days. There's always room for improvement, but I felt this was a solid piece of work, especially given it was in Power BI.
What I learned:
The biggest takeaway from today is that I actually enjoy using Power BI. I want to spend more time during bench getting better at it and the broader Power suite. It would be valuable to become known in the company as someone to go to for Power BI help.
I also felt significantly less overwhelmed today compared to Days 1 and 2. The planning time helped, and having some familiarity with the subject matter made a difference. I'm learning what works for my process.
Looking Ahead
Tomorrow begins a two-day project, and I'm excited because the data involves sports. Given my goal of building a portfolio to work in sports analytics (specifically football), I'm excited to see how these next two days will go.