We wanted to improve our communications, and figured that understanding the characteristics of messages that had lots of ‘likes’ would help us do that. A request to find the most liked Yammer messages in our network was received, and we and figured out a way to find it…
There’s no direct way to find the most liked messages across the network, so we looked into the five locations where we can access Yammer data:
- The Yammer admin export – this doesn’t include data on likes, but you can export every message in the network from it
- Reports in the Microsoft 365 admin portal – only provides data on likes per group up to the last 180 days – but it doesn’t break down into likes per message unfortunately
- Querying the Yammer API – following the guidance here, I was able to get a breakdown of likes per message for a particular group into Power BI. However, it can only be achieved on a group by group basis, and there are thousands of groups in our network
- Use the Group membership API call to generate the numbers of members in each group
The next step to finding it was to get a list of groups that were most likely to have the ‘most liked’ message in it by getting the IDs of, say, the 20 groups with the most:
- Likes in the last six months from the admin portal
- Messages ever from the admin portal export
- The most members from the group membership API
…respectively. Once you’ve got them, remove any duplicates, and use the API to query each of those groups to find the ‘most liked’ message. Note how many likes it’s had in a spreadsheet, and sort by likes. Bingo! Your network’s most liked Yammer messages.
Figuring this out was a fun exercise, though it might take around five hours to find out using this method. Can you think of a better way of finding out?
Before repeating this, you might want to ask what the business value is of this popularity contest? Who will be paying to find out a photo of Peter’s poodle puppy is the most liked message on Yammer?