A MassTLC Topical Discussion on Social Media Analytics

Alan Belniakbusiness, Marketing, Social Media, technology2 Comments

Mark Jeffrey speaking about #datadriven Marketing (courtesy of @AstonishResults)

On Thursday, April 26, I co-moderated (along with Craig Bloem) a panel for a MassTLC Sales and Marketing half-day summit on Marketing Analytics.  The topic of our session was more focused on social media analytics.  For 45 minutes, we had about 50 or so people in the room ask questions, share experiences, and generally learn from one another about social media analytics.  The questions and skill levels and involvement were wide-ranging, which made it a really fun conversation.  I took a few notes on an easel pad (despite having horrific handwriting).  Below is an expansion of those notes, as a way to summarize the key points of the conversation.



Balance of inbound vs. outbound activities – One participant asked if he should dump all traditional (‘outbound’) marketing and go all inbound.  I suggested that it be a mix of both, at least in the beginning.  If inbound fails (perhaps because one is new to the game), then you’ve got nothing on which to fall back.  This is akin to financial advisors suggesting diversification.  Secondly, one should conduct a listening exercise to determine how much of (or if) their audience is online.


Listening and monitoring as a first ‘way in’ with social media with some companies – Sometimes getting executive or even peer-level buy-in is a challenge.  Showing that the conversations are indeed happening online is a good first step in this process, and listening helps.  Additionally, conducting a short listening exercise around one’s competitors is also a good idea.  This could uncover some potential product/service improvements, as well as help inform a content strategy.


Tools mentioned – Radian 6, Sysomos, Google Analytics, Chartbeat, Sharedcount, Topsy, searchhash, SocialMention


Internal subject matter experts – One of the frustrations with content creation is an apparent lack of content creators.  In many technical organizations, engineers and others involved in a product lifecycle often know an immense amount of information.  These are people that can be leveraged as content producers.  If they are shy or otherwise not particularly good writers, consider having someone with stronger writing skills interview these people, and create content from that.


Bottom up or top down – An audience member quick-polled the room with respect to the impetus for a social media movement in their organizations.  The crowd was mixed.  In my mind, a bottom-up approach works well, because the people who need to product the content are the ones behind the initiative.  Yet the initiative needs to be championed or at least supported from the top down.  Having only one or the other is likely a recipe for failure.


Know your audience – Several questions arose about what kind of content works on which channels, and the like.  My response is ‘know your audience’.  If you’re marketing to teenagers for fast fashion, LinkedIn is probably not a good strategy.  Similarly, if you are shopping around for funeral home preparations, Facebook might not be a first choice.  In these hyperbolic examples, it sounds obvious.  But in the real world, it might not be so apparent.  How can you tell?  Conduct a listening exercise across multiple networks and see where your audience is spending their time.  Look to your Google Analytics (or similar) web logs, and see what people are searching on to arrive at your site, and from what referring sources of traffic.  Moreover, understand what they are saying and what kind/s of media they are using.


What’s your point? – A few questions were asked about which network should be used, whether or not to gate an asset, and things of that nature.  A great way to determine this (and one audience member indicated that she and her company does this) is to determine some objectives: what are you trying to do?  If it’s to increase awareness, then form-gating content and not optimizing for SEO is probably bad.  If it’s lead-generation, then perhaps some content is free, while other content is gated to collect information.  If it’s a real-time thing you’re going after, then Twitter is a good strategy, and long-form content might want to be avoided.  Match the medium to the message .


Use content to create community – Someone asked how to scale up a community on something quickly.  First and foremost, a company or an individual needs to have a good product.  If not, then no amount of marketing can fix that.  As in, the product needs to be remarkable (see item #2 here). Then, use content and all its forms to find people in these social outposts (Twitter, Facebook, and even other places like Yelp and things on Amazon) as a way to connect with them.  Bonus points if you can generally corral them into one area.


Is there value in Google+ ?   – Some people are toying around with Google+ and seeing some limited benefits.  It’s still early days.  One comment was that if one uses Google as a search engine, and one has the Search Plus Your World feature enabled, then Google +1s from your circles and connections now influence search results.  The effect of this cannot be underestimated.


What metrics are people examining? – There are some classic web 1.0 metrics, like visits and clicks.  And those aren’t necessarily bad (go back to your objectives and ask again what it is you are trying to do; clicks and visits may be just perfect).  Some other metrics to consider are reach, engagement, sentiment, source/referral, linking to revenue, time spent on site, and click depth.


How do you attribute media/touch points to revenue? – This is, perhaps, the $64,000,000.00 question.  I think many people are asking this, and there are some ways of doing this, and I think just about all of them are terrible.  Last-touch attribution, equal touch apportion among media types, and the like all are prone to error.  In my mind, the best way (though probably the most expensive) is post-purchase survey of customers, with a large enough and random enough sample size.



This was a good session, and I’m thankful to MassTLC for giving me the opportunity to participate.  If you attended the session, did I miss anything?  Or do you have a comment on anything above?


image source: http://instagr.am/p/J4gq_5vuEF/ , via @AstonishResults