Retailers love thinking about how they can use IT analytics of social media to get close to their customers. But when a retailer breaks through the invisible social media wall and reacts to an online post with a very personal in-store interaction, it may not reap the desired increased-sales outcome.
Let’s back up for a moment. Because so much of social media happens in the public sphere, retailers have the theoretical ability to capture and analyze millions of customer interactions and make sense of them. But moving from theory to reality is where the fun kicks in.
One example of how sophisticated the analytics is becoming comes from Salesforce.com, which this year tried to quantify the social traffic around Black Friday. Its analysis also illustrates the hurdles that remain before retailers can extract meaningful and useful data from the oceans of conversations made available on social media.
The vendor came up with a lot of hard numbers. For example, Kohl’s led social conversations on Black Friday, with 46,000 posts, a 1,384% year-over-year increase. Some other findings: “Walmart (34K posts, -42 percent YOY), Apple (26K posts, +192 percent YOY), Target (16K posts, +60 percent YOY) and Amazon (15K posts, +134 percent YOY) were among the other most-talked about brands and stores. Conversations peaked around 4am ET (125K posts), 5pm ET (168K posts) and 11pm ET (102K posts), indicating consumers were shopping from the wee hours of the morning until the late evening.”
All quite interesting, but not as valuable as Salesforce’s attempts to figure out what people were saying. The breakdown: 63,000 posts concerned Black Friday brawls, 19,000 chronicled experiences people had waiting in line, and 60,000 were from consumers who were planning to skip the mayhem of Black Friday in favor of Cyber Monday.
Still, the real value for retailers requires an even deeper dive. They want to know just what was being said about them, favorable and unfavorable. Even better would be the ability to differentiate between thoughtful complaints — specific and based on a legitimate grievance — and kneejerk hate posts, like “Walmart sucks.” They’d love to have an analytics system that could recognize sarcasm and therefore move a comment like “Just saw paper napkins at BigBoxStoreX selling for $58. What a great deal!” from the favorable to the unfavorable column.
Best of all would be software that could match social media usernames with real accounts on the chain’s CRM system and could overlay their comments with actual actions. That would tell them how many complaining shoppers ended up returning products and how much time elapsed between the comment and the return. With information like that, the retailer might be able to save the sale. The longer the gap between comment and return, the better the chance of changing the outcome.
All of this becomes especially interesting when the retailer can tie a comment to a specific store. Frighteningly enough, that’s quite easy to do, if a customer is riding that store’s Wi-Fi or has posted a photo of a product — say on Snapchat or Twitter — and didn’t think to remove the location tags. Once that starts happening, the potential for getting close to the customer makes a quantum leap — and with consumers increasingly making their comments on geolocation-broadcasting smartphones, it could be happening soon. While the possibility remains in the offing, retailers should be thinking about ways that this sort of thing could go very wrong for them.
Consider a scenario laid out by Kyle Lacy, the director of global content and research at Salesforce.com: “I tweet and the customer service rep gets the tweet” and contacts the precise store where the complaint originated. “A clerk can then come up to that customer and answer that question. The brand should know every single interaction.”
Let’s see. How could this possibly go wrong, especially during the holidays when store associates are overworked and, every so often, rather grumpy? Associate: “Hey, lady. So you think that store display looks stupid? I worked for six hours on that display. Think you could do any better? Based on the ludicrous outfit you’re wearing, I think not.”
But even if the store’s representative doesn’t confront the unhappy customer with hostility, the customer could well be taken aback at being approached at all. Although the feeling is woefully misplaced, social media users tend to have a sense of anonymity. They don’t expect the target of their complaints to identify them easily, and many of them are likely to be uncomfortable when presented with proof that their supposed anonymity is hollow.
Some of this is a matter of sensitivity and discretion, knowing how to approach the shopper in the nicest, least confrontational way possible. But this kind of interaction is likely to feel jarring and intrusive to the shopper, who may feel embarrassed. The store wants the customer to feel respected and listened to and taken care of. It wants to make this person a loyal customer. There’s an excellent chance, though, that the customer may develop an immense need to leave the store as quickly as possible and never return.
The identical interaction contained within the initial context, though, is very psychologically different. If a customer service rep posts a response post in the original forum — offering an explanation, apology and some large coupons — it could have the desired impact. I applaud the idea of leveraging data from one arena to another, but am cautioning that human beings are rarely as logical as software.
IT leaders are going to be feeling the pressure to provide more and more of this level of analysis. They should not hesitate to warn their bosses that they could end up inadvertently pressing a lot of customers’ buttons in the wrong way.
Evan Schuman has covered IT issues for a lot longer than he'll ever admit. The founding editor of retail technology site StorefrontBacktalk, he's been a columnist for CBSNews.com, RetailWeek and eWeek. Evan can be reached at firstname.lastname@example.org and he can be followed at twitter.com/eschuman. Look for his column every other Tuesday.