June 6, 2012 Insights, News & Events

A recent online article in the Harvard Business Review discussed three misperceptions widely held by many brands and businesses when it comes to their social engagement strategy. The article Three Myths About What Customers Want underscores the importance of providing your particular audience with information that is high quality, not high quantity.

The HBR’s article speaks directly to the message I’ve been shouting from the rooftops for quite some time regarding how we understand and use the data from social channels. A number of popular published studies have used averages to draw insight from social data, which are then designated as optimal attributes for the entirety of a social network, flattening it into a simple, highly biased model. This one-size-fits-all model has been used to identify the “best” social media publishing times but it ignores completely the most important variable: humans.

Social is everyday conversation happening online between and among people. It is full of chatter and discussions that flow around myriad topics trickling and gushing as interest peaks and then fades. Human behavior online is not all that different from human behavior in the real world, which is to say it is inherently unpredictable. When entering into the conversational stream in social, marketing managers would do far better to remember that people are not the sum of the data. Each of us has the capability to be a different version of ourselves at any given moment. We might be a movie buff right now but in ten minutes when we start to think about dinner we may become a chef only to get sidetracked moments later by something that turns us into our work self. Human behavior will always surprise us (thank goodness!) and any look at the data gleaned from human conversations on social needs to keep this is mind.

The following visualizations are based on messages posted by the Guardian’s Twitter followers over the period of a day. The fluidity in the shape and movement of conversations is crystalized when we see the data as a visual. The first snapshot was taken at 8am UTC. Amongst the multitude of topics in flux, we can easily identify the “Easter” and the “Bahrain” cluster, topics that resonated with a segment of the Guardian’s audience that morning.

The next snapshot, taken at 2pm, looks completely different. While the morning’s events in Bahrain are still of interest, there’s a substantially denser cluster of conversations happening around the Thames Boatrace.


Finally, when we look at the audience activity that evening, 8pm UTC, the audience is clearly mostly engaged around entertainment and sports related topics.


As you can see, some topics linger on for many hours, while others have a short attention span. Making sense of this ebb and flow of user attention gives us the ability to better predict levels of engagement as well as identify optimal times to be a part of the conversation that is already taking place.

So given this state, what is a better way to think about publishing times on social if there really isn’t a best time as determined by demographics and data?  A better way to approach identifying key times to publish on social is to use data and predictive analytics without removing the underlying humanness.  We must move away from the idea that there is an identifiable constant and instead use real-time data that embraces the unexpectedness of the human conversations flowing through social media channels.  The goal shouldn’t be pushing messages out to as many people as possible at one optimal time.  The key to success in today’s dynamic social channels is to use real-time data to identify and become a part of the conversation that puts you in front of as many interested people as possible.

Optimizing for social, in essence, means learning how to have a conversation with the right audience at the right time in this new forum. And just like face-to- face conversations, social conversations follow the same unwritten rules of etiquette. Most of us wouldn’t think of inserting ourselves in an ongoing conversation until we both understood what was being discussed and assessed whether what we had to say added anything to the conversation. Social follows these same basic rules. In order to be heard effectively by an interested audience, marketers need to be able to identify specific windows of “conversational” opportunity where what they have to say adds something to what is already being discussed. By effectively using real-time data, we can identify these windows. This helps marketers find and enter a conversation at a time when an audience is far more likely to listen and reduces the social “noise” that contributes to the attention deficient problem we hear so much about.

Learning how to “read” the data from social is crucial to social media success. One of the first steps marketers can take is to shake off conventional data analysis measurements and the common myths they spawn and look at the information the way it is generated – in real time, by real people.

Written by SocialFlow Marketing