Thursday, September 15, 2011

Social Media Analytics - The net catches the fish that are there...

There is ongoing debate about the value of social media as well as how it can be used to quantify consumer  behavior.

I think we can all agree on the following:

1.  Social media is a huge data set (duh)
2.  Social media is having an impact across the business landscape and has effected companies' business
3.  Social media data is less expensive to harness (and IF you can harness it) because it is freely available
4.  Social media is changing how people communicate and share their thoughts
5.  Social media is causing instantaneous expression of consumer sentiment in real time
6.  Social media in some form is probably here to stay

Although all these things are true, the lack of best practices for utilizing social media in a meaningful way in your business continues to slow its uptake in corporations.  And while this is a culture thing (change I can believe in), it is also about bridging the current methods to the new ones that do not exist.  I would also mention that people are still struggling with how the data is collected in a apples to apples way to give them confidence that they can quantify their data.  And if they cannot quantify or control the variables in their work, it is a struggle to think about this data in a scalable and consistent manner. 

This post is mostly about turning how you think about data collection on its head as well giving some tips to consider when integrating social media into your work flow.

Data Collection:  I have been thinking about the following analogy to help people realize that social media data as a source has some very interesting nuances that make it valid in a different way than traditional data.  For traditional data collection and metric development, most researchers give validity and confidence to these methods (survey research, focus groups, ethnographies) because they control who they are talking to.  I have spoken with researchers and they make a fair point that they can trust the results because through their design they know they are talking to the right people to answer the questions they develop. This is fact...it makes sense...it is right.  I have no argument there.  That being said, I have said here a number a times (and I have worked on and been part of traditional means of market research or consumer interaction) there is tons of bias when directly interacting with consumers.  In addition, sample size is smaller (even though time has shown it can be pretty accurate) and therefore questionable.  And lastly, many of these methods take time and planning to execute and therefore cost more and have much greater opportunity costs because of the time you lose in a world that has changed. 

The fish analogy I like to think about for traditional research is this:  Traditional research is like going to the fish market and handing picking one type of fish, buying a certain number say 12, to see how they differ physically, by smell, by taste when you decide to cook them.

As for social media...I use this analogy.  Instead of picking the best fish and they are all the same, social media is about simply throwing nets into the ocean and seeing how many fish end up in your net.  You don't know how many you will get or what types might be in there, but the real value can be found in by simply catching FRESH fish in a net rather than waiting for them to make their way from the ocean to the fish market. If you can throw 8 nets over the side, you can do it quickly and will be able to directly compare each catch to learn about the difference.  Simply put, you introduce another type of control into your data collection when you use social media.  What this gives you is speed, lower cost (I go one place to catch fish as opposed to walking the whole fish market to find just 8 fish) and potentially the ability to capture knowledge about fish you would ever look at during your study.  If you think about control differently when collecting social media data you suddenly have a new level of control in your data set that was never there before.  The controlled randomness is something that could be considered an advantage.

Some tips to help you see the power of this bridge...

1. Create bridged ratios - Is there a standardized number in your business that you collect that could be merged with buzz, sentiment or passion counts?  Think about it.  If you know that your business grows when your same store sales grow a particular way, you could easily create a ratio of your efforts as it relates to social media to allow you to track if your social media efforts correlate with your traditional and accepted measures

2. Go Broad - Because your have a less control of the demographics of your collected data, you should think about ways to broaden your search to see if there is correlation.  For instance, if your eight nets are looking at your competitors there is real meaning in seeing the buzz or sentiment count differences between you and your competitors.  This difference is important because you method of catching fish is the same.  The differences are real if you create strong definitions when you collect your data.

3.  Trust the lower bias - While it can be hard to know if there is skew in your data online, there is for certain way less bias if you can get pure consumer data into your net.  And because you are simply catching fish that are in your net, you are not looking at them beforehand when you catch them.  Your selection process includes no filter other then the size of the holes you net.  Social media's advantage is lower bias, utilize it to your advantage.

4.  Leverage the speed in your mind - We know the data is instantaneous.  I see it when I give a talk how I am being tweeted about as I do it.  There is no question it is faster, but the you need to trust the value of that to begin thinking about how to bring this into your process.  I trust the speed, I see the advantage of the speed so just go with it.

Just to show you that I don't believe that traditional methods will go by the wayside, I will fast forward in my mind a few years.  In a few years, I see a pipe with social media data being plugged into a company (think of you brand as the data flowing through the pipe).  The end of this pipe has a filter. When the data flows through the pipe it is plugged into a screen.  As the data flows through the filter onto the screen, a company will see numbers instantaneously changing as the data affects the brand.  Essentially, I am talking about real time metrics flowing onto a screen to help you understand what is being said about your brand in the social media sphere.  As we are able on ANY topic to understand instantaneously what is being said and how people feel about it the reversal will start.

What I mean by the reversal is this.  When instant data becomes the norm, the need to slow down will grow again.  At what point when something happens is the change relevant.  Is it 3 hours? 5 hours? 2 days?  At this point the need for older data analysis methodologies will become extremely important and relevant.  We will need to quantify the cause and effect because total optimization while possible in a real time business, will most likely create new business errors that we can't yet see.

The future of how social media will be relevant is exciting, but as with all change will still require further change.  There is no perfect solution on a future state that improves upon the current state.  But that future state eventually is the current state and therefore we must adapt again.

No comments:

Post a Comment