Is social the most independent behavioural variable?



When you want to predict how your consumers will behave in near future, you look at behavioural clues. From database marketing perspective the key behavioural variable is recency (mostly transactional recency).  Transactional recency may not be the most independent variable but it does encapsulate a lot of things and gives you a powerful variable to work with. In fact, for offline channels, recency is by and large the only predictive variable available. In online channels, especially your website, many behavioural clues get lost with call-to-actions (natural structure) of your website. Though the functionalism approach from Semphonic solves this problem to some extent, there’s only so much you can do; after all, your website cannot be completely fluid. Social, on the other hand is totally fluid. So, behavioural clues obtained from social media can be immensely powerful. Perhaps they represent the most unadulterated behavioural patterns of your consumers that you can ever lay your hands on. The problem, as you’d have guessed, lies in linking these social clues back to individual transactions; and we are not there yet!! But if we continue to work with social data at an aggregate level, we are certainly leaving at lot on table!!

In any case, I think, marketers searching for independent predictive behavioral variables will find reasonably good answers in social.

Looking at Social data from database marketing perspective!

Gary Angel in one of his blog posts alludes to an interesting aspect of using social data from database marketing perspective and I think it raises many intriguing questions.

First of all, if you have been following web analytics thought leadership for a while, you’d have noticed that there are 2 broad schools of thoughts in this space. On one hand you have people like Gary Angel, Jim Novo and Kevin Hillstrom et al. who come from solid database/direct marketing background and they see very insightful similarities in web analytics and database marketing analytics that others don’t. The other school of thought consists of people like Avinash Kaushik, Eric Peterson et al. who do not see much link between other fields and WA. Of course, the page and visitor level data from web analytics hasn’t traditionally lent itself to customer level data, so, those who’ve pursued WA as an entirely new field weren’t entirely wrong. Personally, after having extensively read Gary, Jim and Kevin, I think, it’d be stupid on the part of a web analyst to ignore the knowledge from established fields of direct/database marketing analytics, and more so when the technology to link visitor and customer level data is becoming a reality.

Measurement is the Strategy

(I know that's a lame title for a blog post but being a huge fan of Marshall McLuhan, I couldn't help it...I'm always looking for "x is y" around me...if you hate this piece, you'll abhor what I'm going to write next...."marketing is the product"..."customer is the business"...eh:-)


“you can't manage what you can't measure” is a bold statement but I think it still undersells “measurement” to people. There are certain aspects of measurement (marketing measurement to be precise)  that are so fundamental to your strategy that they actually decide what your strategy should be; and you can't possibly outsource this part to cheap data crunching analytics companies; but before I talk about that, consider this:
  • Do a Google search for contemporary marketing gurus/strategists and 9 times out of 10, you’ll end up with someone who’s actually a marketing measurement guru. Seriously, try it! I’m not kidding. And while you are at it, see if you can distinguish between a non-measurement-focus-marketing-guru and Deepak ChopraJ


Why measurement is so important for marketing today:
  • There are two parts to any kind of measurement (1) the act of measuring something and (2) deciding what to measure. The “act of measuring” has its own challenges like finding relevant data sources, integrating disparate data types/sources and the availability of tools for processing and visualizing data etc. etc. but these are mostly tactical things (though difficult).  “deciding what to measure” on the other hand is a more strategic/fundamental problem. Internet is full of marketing-measurement discussions about “act of measuring” but not too many people are taking about “what to measure”!
  • “Deciding what to measure” is what forces you to really evolve/refine your strategy. In fact, it could be an acid test that’d decide whether your strategy holds water or not.
  • I’m gonna go out on a limb here and say if you begin by “deciding what to measure” (in the context of your business) and then jump into strategy, you’d save money and have more success.

Deliver insights or Automate insights?

Those are two largely independent pursuits btw! If you want to deliver social insights by capturing social mentions from web and transforming/integrating the captured data with data from other systems, it can be done even without having a hi-fi technology infrastructure in place. In fact, we are so early in the social intelligence delivery game that justifying technology investments for "ERPing of social intelligence" is still a wishful dream; at least for those who want to use a generic off-the-shelf product in that space. We don't even have a decent social CRM product yet!! A better approach (at least for now) is to start small and then build custom technology solutions while leveraging stable off-the-shelf products from market.

Here's our vision for social intelligence program: