Post Performance Analysis

Overview

Playhouse Communication is one of Nigeria's leading digital marketing agencies. They combine design and media planning with cutting-edge tech solutions to reimagine what marketing is all about.

I aim to decode a social media data treasure trove for one of Playhouse's high-profile clients(StanbicIBTC) and transform it into game-changing insights. StanbicIBTC is one of the great financial institutions in Nigeria and Africa as a whole.

In this post, I will be taking you through the process I took to dig deep, get creative, and strike gold with actionable insights that could redefine the future of digital marketing.

About The Data

Playhouse Communication has been handling StanbicIBTC posts for the past 10 years. The dataset contains LinkedIn, Twitter, Facebook, and Instagram post information in a CSV file.

Process

Data Familiarization

I began by understanding the dataset provided, Exploring the columns, types of data, and any missing values. I familiarized myself with the platforms from which the data originates to understand the context.

Data Cleaning and Pre-Processing

To get an accurate analysis, properly cleaned data is required. After familiarizing myself with the data, below are the steps I took to handle missing values, convert data type, and handle duplicate values.

• Splitting the date column into two columns, the 'Date' column and the 'Time' column

• Assigned the right datatype to each column.

• Some rows were blank in the Sent by column, I went ahead to replace these values with ‘Wasn’t stated’.

• In the Twitter table, 687 rows of data had null values in the impression column, reach column, and all other important columns. So I deleted them since they show no metric values. I was left with 7842 rows.

• There were 1,428 empty rows in the LinkedIn table, I filtered them out leaving me with 6,332 rows

• I deleted 1,484 empty rows from the Instagram table, Leaving me with 8,516 rows left.

• There were 910 empty rows in the Facebook dataset, I filtered them out leaving me with 8,893 rows.

• I append the properly cleaned tables into a single table, which I used for my analysis. I renamed this table 'All post-performance' table.

• In the LinkedIn dataset, two posts with engagement of -5 and -23 respectively are both posts that don’t belong to Stanbic IBTC therefore, I excluded them from the analysis.

• In the Instagram dataset, there’s a large engagement in posts and engagement to impression posts between 2 May 2022 and 2 May 2022 which was because of the #mydreamcanbe challenge, a social media StanbicIBTC children’s competition hosted between 18th May 2022 and 27 May 2022. I deleted the following engagement numbers 33395,30247,29594, 27804, 27520, 27557

• In the Facebook dataset, I took out engagements number 17,453, 72474, and 60004 because they are outliers.

• In the Twitter table, I took out engagement 14,972, and 10,743.

Data Modeling

I created a new calendar table, that contained all the dates that exist in the model. I created a one-to-many relationship between the calendar table and the 'All post Performance' table.

Exploratory Data Analysis

Using data visualization tools I took time to understand the distribution and trends in the data provided. I identified my key metrics like engagement rate, click-through rate, Impression rate, and so on. I checked for patterns in content types, posts, post times, and so on. After exploring the data, I defined my key Questions.

Define Key Questions

Below are a few questions I came up with,

• Which platform yields the highest engagement?

• What is the Performance of Each Platform Over the Years?

• What are the Peak Times for User Engagement?

• What are the Most Engaging Types of Post?

• Whose Content Influences Engagement?

• Which Platform Influences the Increase in Website Traffic?

Insight and Recommendation Generation

After thorough analysis and finding answers to my key questions. I came up with the following recommendations that can help shape the future of digital marketing.

In summary, each platform has its contribution to the improvement of the company. Facebook, Instagram, and Twitter is the best platform to improve engagement and awareness. LinkedIn is the best platform to improve website traffic. Below are recommendations that will help improve the overall Awareness, Engagement, and Website traffic of the company.

- Create Value-Added content: Maintain a healthy balance between having fun and being productive. All work and no play make an uninteresting presence. Posting more interactive content like the social media StanbicIBTC children’s competition, #EasterLoveGeng challenge, and so on, will lead to more engaging and relatable content that encourages the audience to interact with it, share it, and comment on it. More specifically to improve the click-through rate, make sure these challenges involve clicking on the web links.

- LinkedIn audience is typically more professionally oriented, so content posted on this channel should be informative, insightful, and relevant to the financial sector. Share the company’s value and community involvements like the #Together4Alimb that was held in Abuja. Collaborate and network with influencers and other professionals in the financial industry. Engage with people and share articles.

- Timely and Consistent Posting: Posting regularly will improve your reach and posting on active time will increase the speed at which users see and interact with your content. It’s best to post or schedule your post when the time is perfect o’clock. Below is a list of active times to post on each social media

• LinkedIn: 9 am, 10 am, 12 pm, and 5 pm

• Facebook: 10 am, 12 pm, 5 pm and 9 pm

• Instagram: post 10 am, 11 am, 1 pm and 5 pm

• Twitter: 9 am, 10 am, 12 pm, 5 pm and 6 pm

- Including Hashtags boost post reach

- Timely response: Contents that receive more immediate interaction and engagement will generally be boosted by an algorithm to create wider reach Including interactive hashtags boosts the reach of the post.

- Avoid posting too many times a day: Posting too many times in a single day can decrease your impressions and throttle your content as possible spam. Therefore, posting once daily is suggested.

Report Compilation

My findings, Insight, and recommendations have been organized in a structured report. Click here