I was groping around trying to remember Jason Fall’s post on What Social Media Can Do For Your Business when i realized I made a note on it in Google Reader – all  I had to do was look back at my “notes”.

It’s too late, or early, to write a long post about this now, though i do want to say something, briefly, about Jason’s post and another at MarketingProfs that says CMOs Want Measurable Results From Social Media.

The Top Social Media Success Metrics (see below) can be easy or hard to measure, depending on how much enablement work has been done, beforehand, and weather there’s access to the information.  In many cases, the data does exist, but is protected (siloed) and you can’t get it.   In many cases, the real problem isn’t the data, it’s the attitude around it, that prevents gathering it and getting the ROI metrics, intact.

I took a few of these metrics – and my note to myself – was to discuss how each one might be collected.

1. Website Traffic – that’s the most obvious one to pull, you can get the data from Site Analytics.   Problem, is, you can’t get the same data about the competitors because they won’t share it, for obvious reasons – so your stuck with ComScore, Nielsen,  Compete, HitWise, etc – and those platforms aren’t really the same thing as getting Web Analytics traffic .

Hell, you often can’t even get the Web Traffic in your own organization, much less, anyone else’s.     Point being – data exists – and even if it’s measured in different platforms (a visit is a visit is a visit – but they don’t add up across different measurement platforms, a known problem).   Want to get ROI against Web Traffic – try getting Web Traffic (and make sure your analytics is enabled and your site is tagged properly for it – including the Flash part).

2. Engagement with Prospects – tough.  Unless your willing to define it, ahead of time, I’d say …. this might be one of the top metrics, but it’s so subjective – you really need to almost “decode it”.    For example, many companies have terms for what what they want their customers to do … “trade up”, “trade across”, etc – these can be translated into what people are saying around a brand.     A smart way to do this would be to build a simple table that would look something like this:

# of Trade Ups =  ….. (subject mentions our “product x” and “product y” in a Twitter or Facebook conversation)

# Trade Across = ….. (subject mentions our “product x” and then talks about “product z”)

Since a business might define “engagement” as furthering their business goals – you can take those conversations, look at them, via Social Monitoring Platforms, or Manually, if you must, and “decode” them.    But have the MAP, first – or Table, ready.  Do that work, first.  Then … Decode the conversations.

3. Brand Awareness – not too hard – if your just going to count “brand mentions” – it could be the same thing as “frequency”.  In the chart below, “Radian6” brand awareness could be defined as the number of daily searches over time – of course, that presupposes a “searcher” and “search audience”, or “share of voice” that is rarely measured at the same time, but ought to be.

Social Media Monitoring conversations around “Radian6″ – latter part of 2009 – measured in Alterian/SM2/Techrigy.

Getting the “share of voice” is probably the harder part – as you’ll have to know the size of the overall audience.  Compete can do that part, but for whole Internet – but there’s few platforms that can do share of voice, effectively – so brand mentions can doable, depending on how you define it and what your willing to settle for.

4. Engagement with Customers – Doable, but hard to do – usually, your going to end up engaging with your “customers” on your own website – so you need the site analytics, but you also have to define what engaging with your customer means in terms of your business goals – so you need to do, more or less, the same type of table for “engagement with prospects” (which might happen, off your website – and picked up by the likes of Radian6, Techrigy, Sysomos, Google Alerts, etc) and “engagement with customers” – which is happening mostly, on your own site.    Still need to map what that “engagement” is – and make sure you can track every part of it – so that’s a Web Analytics issue … and a business enablement issue.

Doable, but often, expensive to configure and maintain.

5. Revenue -  Should be easy to get – when you can get the numbers.   For example, there are campaigns and pitches I’ve looked at where you can measure the increase in Social Media Buzz – and your making a case that … more buzz around your brand = more sales and money in your corporate pockets, but you can’t get the sales data – it’s in a data warehouse – and for bigger brands, might take a month or two before the report is available, when you can get your hands on it.

In theory, if you can pipe and XML feed with the sales data into the Social Media Measurement Platform, much as it’s sometimes done in Site Analytics platforms, you could map revenue against buzz and brand awareness – and small operations can do it more easily – say, in Google Analytics, but a large corporation – that’s really, really, hard.     If your CMO is asking for this kind of “results driven data” … they better be willing to get the right people in to set it up and …. get the right parts of corporate data flowing into analytics / measurement platform.

Again, it’s doable, it’s just not easy…. the bigger or more silo-ed your company is – the harder it gets.  In some cases, it’s a political and/or a corporate security issue, as well.

6. Brand Reputation – to me, that’s “sentiment analysis” (positive, negative, neutral) tied to specific actions taking place around a brand or corporate website, over time.

We know the accuracy of sentiment analysis is not too good, and,  if you really care about accuracy – you’ll need to do a bit of human sampling.   I dealt with this issue in my presentation last month in London at Monitoring Social Media 09 (below)

I’m in the process of creating a new deck that will go more in depth – but we know sentiment analysis is based on emotional responses (this can be detected by Social Media monitoring platforms)  – the problem is – the topic often can’t be detected.  A human can easily do it, but even the most advanced machines, can’t do it well – not today, they can’t.   Maybe in the future – they will, but today – you have to do a bit of work to make sure what your measuring is even relevant – topically.

7. Prospect Lead Volume – doable.   I’ve touched this type of data at the last few companies I’ve worked at – often, the hardest part is finding out who has the data and asking them for it.    The next part, meshing it up with Social Media Conversations – could take a little work … but if you can get the data AND you have a good collection process in your Social Media measurement platform (your using the right phrases to collect the information), you can get some good charts and ROI out of this – it also helps to overlay the Web Analytics data on top.   What you want to see  is …

1. Increased buzz leads to new visitors on your website (social media measurement platform)

2. New Visitors are segmented in your Web Analytics Platform (and tied to the same buzz that was generated via Social  Media); the segment converts at a higher rate than those who don’t come to the site because of conversations in Social Media.

3. Sales data is overlapped on top and you get the full ROI metrics.

But, in a way, since this is a “lead” it sorta implies – CRM – or Social CRM – and there aren’t too many players out there – SalesForce is one – with the ServiceCloud – but if you really want Lead data volume (that converts) – your going to end up with a Database solution – because you need to store “who” they are.  That’s expensive – and needs enablement work – work that alot of businesses aren’t willing to do.  If you want this kind of metric – be prepared to pay for it – and pay … well.

8. Prospect Lead Quality – if #7 was hard, number 8 – your talking defining the quality of someone your 1) already tracking within a CRM solution  2) or your talking about attribution – someone visiting from site X is more valuable than someone visiting from site Y).

This is going to be a value judgment – it’s going to be defined by what the business thinks it should be and will vary from company to company – keeping historical data is often a prerequisite for this type of metric.

9. Product Feedback – that should be the easiest one – but it’s not – platforms like Crimson Hexagon are good approximations of getting feedback from audiences you never directly asked – see slide 8 of my deck, above, on the Future of Social Media Monitoring.

Else, you could use solutions that are triggered on certain pages or by certain actions a visitor takes – and get feedback tied to that page, or event.  Some companies, like IBM, put “rate this page” widgets on most pages of the site.    Again, this can be a complex or easy metrics, depending on what you want.

Also, some companies have wanted this data to be matched up with site analytics – few vendors handle this now – but it’s doable – but can be expensive – though one of them works for Google Analytics – I just don’t recall the vendor.

Enough for one post.