(I actually disagree quite a bit regarding Marshall’s view of enterprise monitoring tools unable to create valid data sampling and categorization. I actually believe sophisticated language processing tools can achieve high levels of data accuracy. But, I also recognize, mine is not the only opinion, which is why I published his interview in full. But look for my response later in the week.)
Marshall Sponder, an independent Web Analytics and SEO/SEM specialist working in the field of market research, social media, networking and PR, took some time from promoting his new book Social Media Analytics to talk with me about social media analytics and the changing landscape of social media monitoring technology.
Jennifer Roberts (JR): Tell me a little about yourself and the type or work you do for clients.
Marshall Sponder (MS): I do a lot of social listening work using platforms such as Radian6, Sysomos, BrandWatch and similar platforms as well as search engine consulting, which I have done for several years. Another aspect of my work is keyed to choosing the right platforms (or designing requirements for new platforms). I also consult with clients to uncover what I have termed “ultraviolet” data as part of my social enablement methodology. Finally, I also work in the agency space with integrated communications practice WCG.
JR: You’ve just recently published a new book called “Social Media Analytics”. Can you tell me a little about what prompted you to write a book and what you are hoping your audience learns?
MS: I wrote Social Media Analytics to create a more educated buyer of these types of social listening systems and services.
More Read
Also, I wanted to place analysts on the top of the food chain at various types of agencies and internal corporate positions. In a sense, Social Media Analytics continues the work I started as a board member of the Web Analytics Association a few years back; to help Analysts of all kinds have a seat at the corporate table.
My aspiration here, was echoed by none other than Paul Holmes, founder of the well known Holmes Report, after a debate we had in Davos earlier this year where Holmes (a PR Guru) said most PR and COMMs agencies need a Chief Analytics Officer, and modeled that position after me (or someone like me).
JR: What are the biggest points you were trying to make in the book?
MS: Main points I wanted to make were:
1. In a world where there is no common lexicon for Social Analytics deliverables between clients and stakeholders, platform vendors and developers , and various types of agencies, we have to be much more introspective, descriptive, patient to get good results from the analytics offerings currently on the market or under development.
2. The platform tools we use to report on data, actually shape it and change it into something else, much different than what we’re often used to, and we need to take that in account when we size, price and staff analytics deliverables.
3. Social data is very fluid and malleable, often in social media collection, it is impossible to get a good, valid sample of data which can be analyzed, and even when this might be possible, there are many sampling errors introduced by uneven methodologies that make the kinds of insights you can get out of social analytics very dependent on who does the data collection and research. This is a reality that most people still cannot grasp, and I maintain, many agencies are trying to fit into their existing offerings, with great difficulties due to the above. Most of what does exist is not scalable or reproducible in any meaningful way, and often, therefore, not profitable to undertake.
4. International and multi-lingual social analytics is a very hard nut to crack and full service platforms (as I describe in my book) are often the best choice here. In general, in the world we live in today, with no common lexicon, as mentioned earlier, the best results for this kind of work tend to come out of full service platforms for three primary reasons:
- Custom crawling of selected data sources along with custom data extraction (get rid of headers, footers, ads, etc in a way that data aggregators often are not able to do as well, or at all).
- Custom platform built around the offering
- Analysts paired with the platform
By eliminating uncertainty (3 points above) full service platforms deliver consistent results in a world where standards and interoperability between disparate platforms are more hype than reality.
JR: You touch a little on the evolution of social media and business. What do you think are some emerging trends?
MS: Well, I think we’ll see much of the Social Analytics landscape change in the next year or two by these trends:
- Big Data
- Data Warehousing (changing many of the platforms we see today into data providers or apps, taking their data and adding them to data repositories where they can me meshed up with other forms to data such as Web Analytics and House Database lists.
- Social CRM becomes built into the collection platforms (this is already starting to happen)
- Ascendency of Analysts as shaper of data – rather than the “data clerks” and “data monkeys” they are often considered to be. The Analysts will add measureable value to reporting for high end deliverables.
- Rise of mobile devices with more and more power will mean we will all be on Social Media the large part of our waking life – right now it’s at 25%, but in a few years it could be 50%. As more of the world gets on line and as devices become more powerful, Social Media will continue to mushroom and multiply, eventually engulfing mainstream media, and will eventually merge with it, as technologies continue to evolve.
All of this suggests a world in 5 years, from a media and behavioral perspective that is as yet different from what we’re used to – and evolution is continuing to speed up.
JR: What are some common misconceptions surrounding social media monitoring?
MS: I think there are several misconceptions and really, to get them all readers should read my book. Here’s a few that stick out in my mind:
– Next to impossible to get a valid sample of data (and no one has yet figured out what a valid sample is)
– Data from social media is shaped by how and who is collecting it, therefore, choices of platform, process and people is even more important than in most other discrete marketing disciplines such as web analytics, search analytics and email marketing and analytics (where offerings are more standardized).
– Business requirements gathering for Social Listening is difficult to do well because most stakeholders and clients do not as yet understand the uses of the data or how it needs to be collected and tagged (metadata).
– As social data is being pulled from sources of information that are considered to be “free” and a human right (access the internet itself is considered to a “human right” by many all over the world – generation of content by individuals on the internet via peer to peer channels – social media – is also considered to be a “human right”, as a result), it is difficult to distinguish platforms by the data they pull, because most are pulling this data from the same place, the world wide web. Therefore, premium offerings are placed next to free offerings, making it very difficult for end users to decide what they need and what is worth paying for (this was unlike web analytics or email marketing analytics where data was private and owned by the customer, along with the platform monitoring implementation).
– As ownership and web standards issues have not been hashed out yet, and no common agreement has been reached, nor any certification of platforms, platform users, procurement professionals, vendors and agencies must learn to be much more explicit and open with their offerings than have hitherto been; many are working very hard to avoid just that.
JR: Who do you think does social media monitoring well?
MS: I am not sure anyone does, so far, to be honest with you.
I think this is an evolving medium that needs a bit more time to mature. Right now, there may be instances of some good work being done by platforms and agencies, or within the corporate umbrella, but it’s probably not reproducible or scalable, and therefore, not universally usable.
JR: That’s a bit harsh.
MS: I call the cards as I see them; if people don’t like it, then don’t read my book or my WebMetricsGuru.com blog.
The biggest problem isn’t that the work being done isn’t worth the effort – as that is clearly not the case, here. The real problem is that no one is quite sure what they’re getting, what they should pay for it and how long it should take to perform, or even who should be doing the work in the first place.
