Key Drivers for Social Customer Care
By Edwin Margulies
Those of you who are setting up a practice for social customer care have a lot to think about. There's tuning-in the the right social channels, developing proper search criteria for relevant posts, and getting your clusters and trending topics to work well with your business rules. The list is daunting. But one of the fundamentals you need to weave in to your strategy - regardless of platform - are the key drivers for your customer outreach. Here is advice on balancing sentiment, influence, and loyalty.
You've heard the adage: "The squeaky wheel gets the grease." A version of this was first published in a poem over a century ago by an American humorist. To this day, the idea translates easily into a customer service context. A fundamental decision you will need to make is whether or not sentiment will be the key driver for your social customer care team. A few things to consider:
Does the squeaky wheel adage resonate inside of your company? Or do you have an "everyone is equal" approach to customer service? Once you have established the weight sentiment has for customer service, you can manage sentiment easier by using the following tips:
Rank the sentiment of posts automatically. It is really time consuming to have to eyeball each post for sentiment and then tag posts manually. Consider using a social engagement platform that automatically does this. Further, the ability to sort or filter posts by sentiment is sublime. For example, all the angry ones at the top, the neutral ones in the middle and the happy ones on the bottom. If your key driver for customer service is sentiment, imagine how easy it will be to triage down a list of social posts if they are ranked in this manner!
Correct sentiment tags if they are incorrect. NLP (Natural Language Processing) technology makes it possible not only to automatically tag posts for sentiment, but to make course corrections on the fly. Computer programs are not always perfect, so they need to be corrected from time to time. In the context of NLP, these corrections are called "training." It is a good idea to use a platform that allows you to re-set the sentiment on a post if it is wrong. This allows for course corrections and makes the NLP engine smarter over time.
Track customer sentiment over time. It's a best practice to review sentiment over time. The trend you want to see of course is that a customer went from angry to neutral to happy over time. Some argue that it is more important to track happy-to-angry with the same diligence. Of course, if you have a serial complainer on your hands a dashboard that tracks sentiment might look a little spiky. This is either the result of someone who likes to complain a lot or it could be you're doing something wrong for this customer. You should be able to stack all of this customer's posts in a simple timeline and flip through to see the trends.
Use conversation threading for faster drill-downs. If you have the ability to tag posts with a specific conversation thread name, you can sort on that conversation thread to see the trends on that conversation alone. This is useful in separating out those issues where a customer's feelings are diverse. For example, there might be an "angry" thread about a broken product that needs to be returned, versus a neutral thread on a technical question, and maybe even a happy one expressing gratitude for a resolved issue. Your ability to track these separately really helps to understand the customer dynamic.
Some social care practitioners focus on influence when ranking the priority of social channel outreach. For example, if someone with an influence score over 50% is saying something detrimental about your brand, you would engage with that person first in favor of someone with a lesser score. The idea with this approach is that someone with a higher influence score can do more damage to your brand quicker than someone with lower influence.
The problem with this approach is it does not take into account undeserved influence scores by "influence gamers" and it does not take into account non-public influence metrics. Consider the following:
Check on the relevance of your brand mention. Just because someone with a high influence score was caught in your social engagement net does not necessarily mean any outreach is necessary. It is a good idea to use clustering, spam control and trending topic analysis to automatically rate the person's post first. For example, if a "high influence" individual made a post that your system ranked as "spam" or in a non-critical cluster, the post may not be actionable anyway. Look for a system that allows you to create priority rules based on the intersection of spam, cluster and influence. This will save you a lot of time.
Take a look at the person's friend/follower ratio. If you are on the fence about whether or not you want to take a person's influence score seriously, it's a good idea to consider the ratios. For example, if someone on twitter is following 2,000 people but only has 100 followers, that may be an indication that they are indiscriminately following and few are reciprocating. It could be that this person is insincere, or that their content is not compelling. Either way, the ratios can often give you a clue if their influence score is accurate.
Try to link in corporate influence too. All too often, public influence gets a lot of weight, but it is not necessarily linked to how influential that person is inside of your own company. For example, you may have a long-time customer who spends a lot of money with your company. This person may have the highest internal influence score. Of course, the way companies measure internal influence is diverse. Some measure it based on how new a customer is (top priority to avoid churn), or how much they spend each year, or how long they've been a customer. Some companies use a "star rating" to articulate a customer's internal influence. The point here it to take advantage of this important information and try to put it next to the public influence score at the same time. This makes for a more balanced outreach decision.
When a customer is loyal, that means they are usually happy with your company as well. Loyal customers may not display all of these traits, but here are some things to take advantage of: a) they organically promote your brand; b) they collaborate with community members and act as an expert; c) they are a source of good feedback. Let's take a look at how you can incorporate these traits as key drivers for your customer care practice…
Leveraging Promoters. If a customer is so loyal that they quite often mention your brand to friends and family - and especially on social channels - this is not to be ignored. Consider seeking out posts in which positive mentions of your brand are made. First, recognize the individual for promoting your brand; second, consider re-tweeting or re-posting their kudos; and third, consider some kind of small gift (mug, pen, t-shirt, loyalty points) for the positive mention.
Leveraging Content. I encourage you to set up a peer-to-peer community site so loyal customers can collaborate with one another and also post helpful content for other customers to enjoy and learn from. Companies like Telligent have excellent software for doing this. The useful content that is created by a loyal customer on a community site can be gathered and distributed with people on social channels. This has the effect of not only helping the other customer, but of using third party attribution at the same time.
This can be accomplished with a social engagement platform that has built-in agent assistance in the form of dynamically ranked knowledge base articles. The more advanced systems give agents an automatic pop-up window with ranked knowledge based articles based on the content in the social post. The idea here is that helpful articles posted by loyal customers can be re-distributed. This provides another form of recognition to the loyal customer while at the same time helping a completely different customer.
Sentiment, Influence and Loyalty may all have different weight in your own social customer care practice. Regardless of the rank, it is important to first memorialize your priorities. Second, you should leverage these key drivers by investing in modern infrastructure that automates the heavy lifting for your social care agents. Allow software to tag, prioritize and fire rules that make it easier for agents to concentrate on what's important (outreach on actionable social posts). In social customer care, spam and non-prioritized posts are the enemy. The more time you can claw back for your agents to concentrate on customers in need - the more loyal and happy those customers will be.