Claiming “Facebook is failing marketers,” a report by research firm Forrester unleashed a social media firestorm this week. The report documented the results of a survey of 395 marketers in the United States, the United Kingdom, and Canada. The marketers were asked to rank the business value derived from digital marketing opportunities from Facebook, Twitter, and YouTube to onsite ratings and reviews to branded communities and blogs.
They rated Facebook dead last.
In an open letter to Mark Zuckerberg published on the Forrester blog, Forrester analyst Nate Elliot explained the two main reasons behind dissatisfaction with Facebook:
- The lack of engagement brands see via their Facebook pages
- The fact Facebook’s is not “good enough” at pure advertising, even though Facebook is trying to shift its business in this direction
Based on my knowledge of big data and data mining, I have to wonder how the marketers surveyed measure digital marketing success. Unfortunately, we do not know which companies the 395 marketers represent. So we can only assume Forrester’s research expertise would lead it to select subjects with opinions based on objective quantitative data versus subjective guesses and opinions. Without knowing the basis of their assessments, I am unsure what to make of the Forrester report and the resulting controversy (I did not pay $499 for the full report).
What I do know is that many social media experts seem to measure Facebook marketing success using vanity metrics such as likes, fans, or views. But they do not seem to know much about their clients’ fans and followers. Are they customers or donors? Are they potential customers or donors? Or are they bots? Many avoid answering whether digital marketing activity results in sales, donations, or other measurable actions. While you hear a little about the importance of measurement (often using spreadsheets) or data mapping, rarely do you hear any discussion about data mapping in the context that matters for data analytics (no, I am not talking about data visualization mapping, such as Ushahidi).
Data mapping, the step before data mining, involves identifying all possible structured and unstructured data elements collected or available on potential customers, donors, or other study subjects (e.g., digital information from web sites, mobile devices, software logs; verbal conversations; hard copy records, etc.). Then you find ways to link different data sets to find classes, clusters, associations, outliers, sequential patterns, etc. Just as successful social media implementation requires the attention of senior leadership across “silos,” successful data analytics requires the collection, analysis, and sharing of relevant data across “silos.”
Many firms analyze big data with propriety software, such as SAS, and some do this with free open source software, such as Hadoop, MATLAB, or the R programming language. Data analytics is how Amazon knows to recommend items that would make perfect presents for your relatives. Due to Amazon’s obvious data analytics expertise, I am positive an Amazon marketer could analyze quantitatively why or why not Facebook and other digital marketing opportunities are working for Amazon and its commercial customers.
Take a look at all the big data offerings available on Amazon Web Services. In comparison, Facebook Page Insights analytics can accurately, to quote Forrester, be labeled “simplistic.” You don’t have to understand how the marketers Forrester surveyed measure digital marketing success—or even agree with their opinions—to see that.