If you listened in on the Webinar we participated in last week (listen to a recap here), you may have seen Mark Hatch’s slide on measuring influence (that slide is also floating around on Twitter…). We call our impact (or – if you must – influence) scoring methodology STARR. I thought I’d spend a few blog posts discussing each of the elements of STARR and some use cases where they may come to the fore – starting today with Sentiment.
After analyzing heaps and heaps of social data over the past several years, we’ve come to realize two big truths: First, that there’s not a single element that defines web presence– it’s not just popularity or topical relevance or authority (which is what one of our clients told me this week is what defines “influence” for him); instead, you need to look at a number of metrics together to really understand an individual’s impact (hence STARR!). Second, we realized that what’s important in web presence analytics is different for everyone. In other words, different use cases place different emphasis on different metrics.
In some cases all you really need is something simple – like a popularity (or Reach, in our lingo) measurement. Most of the time you’ll want to evaluate several metrics at once, but depending on what you’re doing, you’ll want to place more emphasis on some metrics and less on others.
For social customer service/customer care applications, we’ve found that all five of the STARR metrics can be important. For example, the Reach and Authority metrics are really handy when looking to identify potentially high-impact customers with complaints, so they can be queued for higher-priority or higher-tier service.
But what one of our customer service partners has found to be really helpful is sentiment analysis. By looking at a customer’s social activity – either as part of a proactive outreach to someone who’s complaining online or delivered as part of a social profile screen pop when someone makes an inbound request to the customer service team – and in particular looking at the sentiment of that activity, a customer service rep can be forewarned and forearmed when the time comes to being the engagement. Indeed, some organizations may even use sentiment data as part of their ticket routing protocol – by itself or blended with other metrics to identify those people who should be queued to the best and most experienced agents.
Working with our customer service partners we’ve found that it’s not at all necessary to get too granular with sentiment analysis either. Something as simple as happy/mad/neutral is enough. One of our partners – shown in the screen cap below – simply uses cartoon smiley faces in the queue, so the agent knows with an instant glance what to expect when they pick up the phone or engage a user via Twitter.
In the end there are a lot of ways that understanding sentiment can be useful to a customer service organization. Here it’s on an individual basis, but understanding the overall sentiment trends of customers can be important, as can slicing and dicing it into demographic or regional subgroups.
Next up: TopicNo Comment