In my first post in this series, I laid out a diagram that showed our view of the different ways we see our clients using Social Media and the corresponding measurement functions for each:
The perfect symmetry of this chart is deceptive and could easily lead one to believe that these functions are roughly equivalent within the enterprise. No so. At most of our clients, Public Relations and Social Campaigns dominate the “conversation”. There are many reasons for this dominance, but one of the biggest is that for both PR and Campaigns, there is a clear institutional owner with a natural interest in social media. This is particularly true in Public Relations. The line between PR and social media is blurry – particularly when it comes to listening and measurement. Some of the leading listening tools started life as PR clipping services and have simply expanded their scope over time to embrace a wider range of channels. As a consequence, PR functions tend to have a larger voice and a clearer set of imperatives when it comes to Social Media and Social Media measurement.
No matter how blurry that line, however, ownership of Social Media by PR is a poor idea. The multiplicity of function that I’ve shown above should stand as simple proof. Can any internal group or agency be less qualified to deal with Social CRM, Customer Research, Customer Support or Community than PR? It’s hard to imagine. Nor is the blending of PR function and marketing function a happy one. PR and social campaigns are one of those pairings that look close enough to be possible but turn out, nearly always, to be impractical.
It’s really as simple as this – PR is a personal business – it’s ultimately about building trust relationships with influencers. With PR Influencers, your job is to get them information that they, in turn, can use with their audience and to give them reasons for wanting to do so.
Marketing, on the other hand, is talking directly to the customer – trying to sell the brand and products in ways that connect directly with their experience. It’s sound similar, and the growth of social media makes it seem as if “we are all influencers” now.
But that’s not right at all. We all may be influencers, but professional influencers have a fundamentally different set of interests and needs than real customers who may, secondarily, become influencers.
So even if the channel of communications is identical (social forums), the strategies and techniques are completely different. It’s not for nothing that PR Agencies and Marketing Agencies are invariably distinct.
It shouldn’t be a surprise that this carries over into Social Media and measurement as well.
In my post last week, I talked about the somewhat unrecognized existence of sampling at several tiers of social media measurement. The sample has a profound impact on how you can use the data and the type of conclusions you might make.
It is fairly common practice, for example, to sample verbatims for human readership and classification based on Influence. If your focus is on PR measurement, this makes perfect sense. Nearly everyone you care about as a PR person should rank high in any reasonable influence measurement.
If your goal is to measure brand sentiment, however, it’s a very poor sample to subset by influence. By selecting for influencers, you’ve likely eliminated almost your entire “customer” population. If your goal is to evaluate the effectiveness of your marketing campaigns, it’s entirely unrepresentative – not only by individual but by channel. Social marketing campaigns are far more likely to resonate in channels outside of traditional media. And it’s hardly even necessary to point out that if your interest is in Product research or Customer Support, an Influence-based sample is worse than useless.
It’s easy enough to flip this around and point out that a sample optimized for marketing or product research purposes is every bit as misleading when it comes to PR. Indeed, the very nature of PR makes sampling anathema. PR, like Social CRM, should really be a comprehensive capture of social listening from the identified population. An influence-based sample is only an effective solution if it happens that the sample size is equal to or larger than the target population of influencers.
All of this is concerned with sampling at the level you have control over.
For PR measurement the sample at the very top of the funnel that the vendor performs is probably as important. The extent to which traditional media are collected and aggregated by the vendor is critical to successful PR evaluation and varies far more than their collection of twitter feeds. This is an area where you need to go deeper than a check-box in your evaluation of vendors if you’re seriously interested in tracking your PR function.
Differences in measurement don’t end with sampling. The type of reporting and the metrics you use when measuring PR are quite different than in any of the other social arenas. Influence metrics are, of course, of the first importance in PR. We also like to use measures of topicality (the extent to which they are topic appropriate) when reporting on PR targets. Such measures aren’t useful for true customers, since a customer’s mention stream of you and your topics is likely to be incidental. For PR Influencers, though, identifying by topic appropriateness is probably even more important than by influence. Besides, influence can only really be measured by topic – I doubt a Semphonic reference from Lady Gaga would add much to our bottom line!
Probably the biggest difference in measurement is that most PR effectiveness measures should be built from aggregations of specific individuals. This is quite different than other forms of Social Media analysis, where the metrics are derived from a constantly shifting population. In PR, you need to identify who you’re talking to and then track their social postings over time to see if you’re influencing the conversation. In customer support, by contrast, you’re interested in the population that happen to have problems right now. Ditto for campaign marketing. There are times where tracking individuals over time is interesting and necessary outside of PR (I’ll show some examples targeted toward Campaign and Community Acquisition in a future post), but it’s the exception not the norm. For PR, it’s the norm not the exception.
All of this amounts to a simple organizational principle: it’s a huge mistake to house the Social Media function in a single discipline – especially PR. It’s equally bad to let one department choose the tool for Social Media measurement (or to assume that only one tool is required) – especially PR. Finally, it’s erroneous to think that a single set of metrics is appropriate across these functions.
If you concentrate your Social Media effort under a single operational department, you’ll either limit your efforts in Social Media to a single function or cripple the data for every other function. When this happens, simply abandoning functions may not be as bad as using the data. When organizations don’t understand how terribly biased their samples are, they will use them in the most inappropriate ways (such as using data sampled by “Influence” to measure brand sentiment or consumer interests). As I’ve said many times before, when it comes to making decisions, it’s really much better to have no data than bad data.
