I’m continuously amazed at how many people are willing to accept certain key metrics as important to their business without fully understanding what they mean.  Analytical tools are becoming more advanced and the available metrics are increasing daily.  There are a lot of options and choices available today for a business to measure the impact of their digital initiatives and the available options can be overwhelming.  However, overwhelming or not, businesses should choose metrics that matter to them and metrics that inform action.

Share of Voice is my favorite misunderstood metric in the social arena.  You might think it’s sentiment (and I assure you that article is in the works as well) but it’s Share of Voice, hands down.


Share of Voice in Online Advertising is an ad revenue model that focuses on weight or percentage among other advertisers. For example, if there are four advertisers on a website, each advertiser gets 25 percent of the advertising weight. This method ensures one ad will not be seen any more than the other three ads. And, since there are typically a limited number of advertisers using a Share of Voice model, ad exposure is optimized.

Not what you expected? Confused? The model or term of art grew out of the advertising industry based available advertising space. When you think about it from this perspective, it makes sense.  A large conglomerate like P&G may buy up a certain percentage of the available advertising on television or magazines and it can signify how much influence  or marketing they are trying to exert on the marketplace at any one time.

My guess is that as the internet and social evolved, the term was used to label the “Share of Voice” one brand might have over another in online conversations.  There is just one problem with that giant leap in understanding.  Actually, that’s not true, there are many problems with that giant leap, but let’s start with the first and largest assumption.

Assumption 1: There is a finite amount of conversation

In the advertising world there are variances in the amount of available ad space, but the swings are not near as volatile as they could be in a conversation analysis.  Look at the major sports categories.  Certain times of year people will talk about certain sports or events. Other times of year, the conversation does not exist. Over time you can average the “Share of Voice” but where the metric is completely inappropriate is during the off season.  Huge swings in percentage of share may exist where the volume is so low the metric is practically useless. This is a real brand problem year round for brands that don’t have much chatter at all or exist in an industry that is not focused in digital at the moment.

Assumption 2: You can compare apples and oranges

In the advertising world, someone is making the ad purchase. There is a purchase order, a mailing address, a company or someone that is tied to a brand or company making that purchase. It’s a certainty because you can clearly identify who made the purchase and that number will be clear and accurate.  Comparing conversation is an exercise in the uncertain. For instance, let’s compare apples and oranges.  You can easily measure the “Share of Voice” of Apple or Apples and Orange or Oranges and return a nice, clear metric to report.

I’m sorry, that’s almost correct. The metric will be nice and clean after you remove all mentions of the “Big Apple” from people referring to New York City.  But, be careful because you don’t want to misconstrue those conversations with conversations from Apple advocates who are super excited about a big apple they purchased at the grocery store.

Also,we need to discuss juice.  Does juice count? It’s not the whole fruit but could be a part of the conversation and have an impact on the perception of the fruit as a whole. If we include references to apple or orange juice, does this means we need to measure “juice” as well to capture unknown conversation that could go either way? If we add juice, are we going to measure Tang and other artificial flavors? If we add the artificial world, does this mean we need to think about candy?

One more thing: food.  What about apple sauce, apple fritters or orange marmalade? Are we going to need to think about food products that include the flavor or chunks of fruit as well? Scratch-and-sniff stickers or plastic fruit designed for decor?

You get the point. For online conversations, it’s a process. It requires thoughtful application and planning.  This is not something you can slap a few brand names or product names and assume you are done.  You have to think about what you are trying to measure and what is really important to you. This is what separates good insights teams from those going through the motions.

To be clear, I’m actually a fan of the metric. I just think it’s casually misunderstood and people assume too much with their use of the metric. It takes time and patience but in the end, a thoughtful application of this metric can help offer insights into your programs and deliver value.




related post


Your email address will not be published. Name/Email fields are required *

You may use these HTML tags and attributes:

<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>