I attended a conference last week announcing the new Adobe Social Analytics product. It was claimed to be a way to measure Social Media and put convincing metrics across to business owners for the first time. Adobe is in a good position to do this because of their pre-existing Omniture implementations.

The problem they’re trying to solve is a valid one. I believe that Social media has not only customer service benefits but wider benefits when it comes to sales, that it improves customer loyalty and consequently the bottom line.

The metrics used are currently so fluffy as not to stand up to any worthwhile statistical analysis. The paid search department can say they brought in £1m in revenue last week. Social media is reduced to saying how many people “liked” a post or how many clicks there were in links. There is no real way to draw causation between a “like”, a “comment” and a transaction.

The Adobe product was a great disappointment. It is a solid product, no worse than the other social media monitoring tools already available on the market. However, I expected better.

The Product Manager on stage kept on claiming not just correlation, but causation. They merely put the increase in mentions, comments etc alongside revenue stats in the same tables or graphs, down to a business unit level.

It is easy to think of many examples of how flawed this is. When BP had an oilspill, their mentions would have gone through the roof at a time when their revenue was going down – so the less mentions, the more revenue.

Or a TV campaign is launched, mentions go up and revenue goes up. However, the TV campaign is the root cause, the mentions are as a result of it, not the other way around.

So, we are no further.

We still need advanced attribution models to be able to provide insight into where in the purchase funnel Social Media appears and then try to attribute weight to that “touch”.

Or Facebook starts providing the ability to match up customer data at a user level with brands’ websites, something with so many privacy implications that it would be surprising if it ever happens.