Social Media Measurement - An Honest View

We have all come in contact with the maxim that “what you can’t measure you can’t manage”. It’s been at the very core of most marketing and management decisions and is slowly providing the world with an empirical view of reality. Online is notoriously measurable - there is more data online than has existed in the entire lifespan of humanity, yet it’s only useful when it’s available. In almost all cases, websites don’t make their site details public which means we, as the Social Media Measurement industry, are left with an untold monster lurking in the corner.

Over time I believe this will change. Site owners will be less interested in website traffic and more interested in conversions so we may very well see this sort of information being made public in the near future.

As mentioned, the problem comes in when trying to measure the value of something when you are denied access to the statistics (i.e. you don’t have access to usage statistics). The second great problem with online (offsite) measurement is that nobody seems to be able to agree on what to measure and how to go about it. Below, I will present four different models and explain a bit about each. All are focused on assigning a financial value to conversations online.

Ad Value Equivalent is a relatively simple concept which has only ever applied to offline. It denotes how much it would cost to purchase an advert in a newspaper or magazine of an equivalent size. This number is typically multiplied by a “trust” factor - ranging anywhere between 3 and 10. This is done in order to work out how much it is worth, given that people trust PR more than they do advertising (or so the theory goes). The rough formula is as follows: AVE Value = size of advert x number of readers x some financial value per reader. For online it’s slightly more difficult - this is not because of the math, but rather because we don’t have access to all the information required to make the calculation. We know the size of the advert (number of times the brand is mentioned) and the value per reader (Cost Per Mille average) but the reach is often difficult to measure. To solve this, we have used a combination of a wide range of variables including; Compete rank, mozRank, PageRank and a number of on-page factors to determine reach and activity. This is then recommended to the user and they can update the specific credibility of the author to further influence reach and impact.

The output is simple; a mechanism of benchmarking PR success and a metric which is already understood and accepted by the community.

AVE is all well and good but it’s really been a case of “this is the best we can do” for the traditional offline environment. Given the drastically improved measurability of the Internet, it becomes possible to compute daily conversation value based on its impact on the brands reputation. FTI published a paper in 2007 which stated that 20% of a company’s value was based on its reputation. Furthermore, the Edelman Trust Barometer 2010 states that perception is more important that the quality of product and services. It may vary wildly per brand but a percentage of the company’s value is derived from its reputation. As such, changes to reputation can be extrapolated to provide values in gains and losses for the company.

Consider the following – a brand is worth R1 billion. Its reputation is worth 20% (i.e. R200 million). A 1% change in reputation is R2 million. So, if your reputation went up by 1% in the course of the day, it provides an indication of the value of that day’s conversation. Typically, we find that brand’s reputations change by fractions of a percentage a day and only a couple of percentage a month so this model definitely shows promise.

We haven’t formally integrated this into BrandsEye as yet, although it is possible to manually do it from the daily summary emails which provide reputation analysis reports.

The last of the mechanisms, and likely the hardest, is to measure the value of a conversation based on the strength of the relationship. The problem arises when trying to find the means to measure a relationship. As a partner mentioned yesterday “I don’t go around rating my friends and family in terms of relationship and if I did, who’s to say I am right? My view of the relationship may be completely different to theirs” - a pertinent point. It was also noted last week that “it’s not about how many people listen to what you say, it’s about how many people care”. There are however, some mechanisms to determine the strength of a relationship. Looking at genuine engagement rates and determining the amount of investment which individuals have invested in a brand is a good measure of this. For example, a blog post about a brand may take 3 hours to write - that time is opportunity cost which therefore defines the value which the author attributes to the post.

Measures such as word count, conscious investment, research etc all contribute to determining the amount of time somebody is wiling to invest writing about a brand and consequently, the strength of the relationship. Incidentally, this same mechanism applies for negative mentions as a measure of “un-relationship” (or hatred, to use a real word).

Social Media Measurement isn’t going to be fully solved in the near future. That is, until websites, somehow, provide everybody with access to onsite conversions and traffic information to trace the full sales cycle. That said; offline has been “accurately” measuring the value of billboards etc for decades. It’s only now that there is an expectation for the Internet to provide measurability that we come against opposition.

This is a journey of discovery and learning for all of us and I would love to hear your thoughts on this subject.