One of the questions I am asked most frequently as Social Media Director at a full-service creative advertising agency, is what is the ROI of social or of a particular social initiative.

Generally, my response is that it depends. On the brief, on the objectives, on the way the client currently measures financial success, and on a number of other factors.

Generally, I am an advocate of bespoke measurement frameworks for each client and project, but I'm interested to hear other people's views on how they answer this question of the ROI of social.

Appreciate it's a broad question, so am happy to delve into specifics once we have opened the debate up a little.

ROI should really be about the $$. Will you create more revenue or reduce particular costs?

If your brief doesn't provide this focus, then it's a different return your client is looking for. And if you can agree that up front, then you don't get into trouble later when the project is live and questions about financial impact arise.

However, you can also be clever about articulating benefits. For example, say you have a self-service community as part of your social media strategy. If you know that each customer service help desk call costs $50 and the number of calls drop off after go live, then there's a measurable return.

I always think ROI is about asking the right questions first.

Totally agree Gavin. I'm usually looking to articulate the types of social behaviours we can expect as a result of the activity, and from there how those behaviours can be tied back to the bottom line - just like the customer service example you gave.

I'd say things get trickier when clients are asking for returns that are hard to tie back to their bottom line in a measurable way. The problem of attribution of financial gains to social activity becomes even harder in multi-channel campaigns.

If ROI is about asking the right questions first, have you found that clients are often asking the wrong question?

It's always interesting to ask what what's valued.
I think it was Eisenstein that said "Not everything that can be counted counts, and not everything that counts can be counted"
That being the case I think it's important to create a value equation for all of the participants in a social system.
It's always good to understand what a high value behaviour looks like to an organization, what an incremental shift costs the business and how much that ongoing change delivers to the business over a period of time.
You're really just after a line in the sand here.

An example of this (and in addition to the case put forward by Gav above) is that a number of organizations are placing significant business value in the behaviour of 'promoters'. The cost to business to implement a Net Promoter System is significant- and there should be some very robust bottom line accounting on where the behaviours of Detractors, Passives and Promoters sit for both internal and external audiences.