Large-scale brand WOM monitoring is more than a one-man job
By Molly FlattRichard Stacy is having doubts about the way some companies use sophisticated tools to gather online consumer conversation about brands:
“When I look at all these impressive reports I can’t work out how they help me design and run a social media strategy. They could help me craft a one-to-many message (but that’s called advertising not social media) and the sentiment / volume metrics might help in measurement – we did x and the volume / sentiment needle moved x per cent in this direction. But that’s about it. And in any case all of this intelligence I would be getting through using my homemade tool anyway – albeit the intelligence would be in my head, not in a chart.”
He’s obviously been subjected to some pretty pants reports, but his suggestion that large-scale or more sophisticated social media monitoring is therefore obsolete is a case of throwing the baby out with the bathwater.
Yes, it’s true that many companies won’t have the resources to employ an agency such as ourselves to undertake comprehensive WOM listening work, and in that case it’s a great start to use the free tools out there to keep an eye on what’s being said out there. And he’s spot on when he says that the value of the resulting data lies in the human insight applied to it. But apply his one-man model to a brand with more than a dribble of WOM, and that insight will be severly limited, and possibly dangerously skewed.
Conversation is nothing without context. Volumes and trends are important, but unless you have the ability to not only identify relevant conversation from all the different platforms online, but drill down into what’s going on in those conversations, you can make the pretty patterns mean whatever you like. At 1000heads we use proprietary software that has evolved over ten years to gather seriously comprehensive amounts of WOM, but we also have a specialist team of human researchers trained to examine what that data means. They can rate the polarity of each unit - it’s relative positivity, neutrality or negativity on a scale of -10 to + 10, something no computer or single person could do with any effectiveness - but that’s just the start.
What about emotional engagement? Exactly what sorts of emotions are being expressed in relation to the brand - anticipation, trust, passion etc - and how strongly are they felt? WHO is having these conversations - is it the same few advocates or dectrators ranting in different venues, or a large number of consumers only commenting once? Do the conversations tend to snowball or quickly lose steam? What are the WOM triggers - what issues or emotional connections are inspiring these guys to talk? What are their stated intentions? And what about competitor benchmarking? Sure, a brand might have a low volume of chiefly negative WOM, but how does that compare to others in the industry? Getting a true impression of a brand landscape is much more complex than checking out a few threads and ‘getting a general idea’. And I have doubts that any single person can truly stay on top of these complexities.
Stacy is of course also spot on when he warns against unactionable WOM reports. Every observation in the data needs to be accompanied by a recommendation, shown why it is important and how it feeds into a larger WOM strategy. It is essential not to reactively jump at every fluctuation or negative comment, but also to have some mechanism whereby burgeoning problems can be identified early and pre-empted, and crises dealt with immediately.
Stacy talks a lot of sense. I agree that much of social media monitoring IS snake oil. But that’s exactly why, when done properly, it takes time and resource. Word of mouth research too often suffers from the impression that it is a ‘cheap’ or ‘easy’ option: no other specialist discipline would be assumed to be covered by a few free tools and the (no doubt capacious) brain of one guy. As I said, many companies will only be able to start with that one guy, and that’s great. But it’s important that they are aware that it will give a limited impression of their true conversational landscape.
