Creating an effective social media strategy with AI-driven insights

The challenge

Connecting with diverse audiences on social media

Despite this brand’s strong reputation, their communications team struggled to reach and connect with their audiences on social media. Like most corporate businesses, they had diverse target audiences, ranging from jobseekers and journalists to investors and analysts, with equally diverse interests.

With social media being a key channel in the wider communications ecosystem, the team also had various communications objectives and struggled to hit the right balance between different topics.

We were tasked with conducting a comprehensive channel audit of their LinkedIn and Instagram presence to develop an effective social media strategy to grow their following and drive engagement with diverse stakeholders. This included defining:

• The optimal content mix - what topics to post about and with what frequency
• Posting schedule - how often and what day and time to post
• Key performance indicators (KPIs) based on current performance and objectives
• A high-level editorial plan with key dates and brand moments

The solution

Leveraging AI to accelerate data collection and analysis

In order to accelerate data analysis and generate strategic insights, we developed a streamlined process to leverage AI. The process included:

1.    Data collection

We started with a comprehensive channel and competitor audit of LinkedIn and Instagram, leveraging an AI-powered tool to extract post texts, hashtags, media and performance metrics. This laid the foundation for our analysis.   

2.    Content analysis

To analyse their current content mix (what topics they posted on and how often), we used ChatGPT-4 Turbo. This allowed us to automate what would otherwise be a time-consuming and costly process, quickly reviewing all data and creating a granular tagging taxonomy to get an overview of topics posted about.

3.    Performance data analysis

To understand what content performed well, we leveraged ChatGPT-4 Advanced Data Analysis. This allowed us not only to analyse performance metrics for different posts but also to create data visualisation and get deeper insights into why certain content performed better than others.

4.    Strategic recommendations

We held a collaborative workshop where we learned more about the client’s communication objectives and priorities. With these learnings, together with our audit and analysis learnings, we were then able to develop a comprehensive social media strategy and editorial plan based on our insights.

The impact

A data-driven strategy at a lower cost

Our use of AI significantly enhanced the efficiency and depth of the social media analysis, resulting in effective channel strategies informed by big data. Key outcomes included:

• Enhanced efficiency: We estimated that we were able to process data 81% faster, saving over three days in tagging and analysis.
• Scale: We could analyse a larger dataset, extending further back in time and including more competitors.
• Greater clarity: Big data was seamlessly translated into a structured tagging taxonomy.
• Benchmarking: Thanks to our comprehensive data analysis, we could easily set realistic KPIs informed by both previous performance and future objectives.
• A more data-driven strategy: With data-driven insights, we could establish the optimal content mix, posting schedule, engagement-driving hashtags, effective hooks, and content promotion strategies.

This project highlighted AI's potential to enhance communication strategies by efficiently processing and analysing data, allowing marketers to focus on strategic decision-making. AI's predictive capabilities and data analysis strengths support smarter business decisions, empowering marketers to excel in their core area — communication.

This case is one of many examples of how Comprend is incorporating AI into our daily processes with similar approaches able to provide valuable insights into market research, brand positioning and channel strategy for SEO and advert­ising.