Analysing LinkedIn Performance
Internship project analysing post-level LinkedIn data to identify which content characteristics drive engagement and who is interacting with the company's posts.
Context
De Innovatiespotter posts regularly on LinkedIn, but like most organisations, content decisions were mostly guided by instinct rather than evidence. During my internship I analysed the company's LinkedIn performance to find out what actually works.
Problem & research question
Which content characteristics generate the strongest engagement, and who is interacting with the company's LinkedIn posts?
Approach
- Collected and structured post-level LinkedIn data, categorising each post by content type.
- Classified the professionals who reacted or commented on posts based on their job roles, to understand who the content was actually reaching.
- Compared engagement across content types to identify patterns rather than one-off outliers.
- Explored whether engagement patterns could be used as a signal to identify potential clients.
Results & insights
Posts featuring maps and comparisons between municipalities on specific innovation themes generated the strongest performance — a clear, actionable pattern for future content planning.
Challenges & limitations
- Classifying the professionals engaging with posts by job role required manual judgement calls where titles were ambiguous or incomplete.
- I also tested whether engagement data could help identify potential clients. The dataset was too limited to provide reliable evidence either way — a fair result to report honestly rather than overstate.
What I learned
This project reinforced that good analytics means reporting what the data actually supports — including when it doesn't support a hoped-for use case, like lead identification. The clearer win was content strategy: knowing that map-based, comparative posts perform best gives the team something concrete to act on.