Inside databases

I recently attended a conference organised by Teradata, a company that has pioneered the rapid analysis of vast quantities of data since the early 1980s.

The name Teradata is based, of course, on the word Terabyte – in those days an unimaginably massive amount of data.

To put into perspective how massive, in 1985 a terabyte of storage would have cost $300million.

The reason Teradata isn’t a household name is that it has always focused on customers with really serious quantities of data – we are talking major governments, large banks, Wal-Mart, Coca Cola and so on.

By far the majority of its customers are in the Fortune 1000. So, oddly, while most people have never heard of the company, it stands at the forefront of data warehousing. In my opinion the company is well worth watching because what Teradata does today, Microsoft and Oracle may be doing tomorrow.

One of the recurring themes during this year’s conference was social networking. Not doing it, but recognising that a social site like Facebook or Bebo is merely a large database and is therefore perfectly susceptible to analysis. But why would anyone want to analyse social-networking data?

It has become common for businesses to collect data about their customers’ behaviour and to analyse it, sometimes in combination with demographic data.

This kind of analysis allows us to cluster customers according to their behaviour and demographic status.

Thus young, single, high-income males and older, high-income, married males might be grouped together as ‘gadget-buyers’. In turn this information can be used to predict future behaviour and also to direct advertising spend.

One problem with advertising is that we perceive it as such. Grouped or not, we humans are smart enough to know that advertising is designed to influence us and treat it accordingly.

This doesn’t mean that it is ineffective – whether we like it or not, it doe s influence us but it is still a very blunt tool. We are far more likely to be influenced by our peers than by a 20-second television advert.

My best mate gets a new phone and raves about it so I get one too; everyone at work starts eulogising a new programme on TV, so I start watching it.

We have always been aware of the power of peer group pressure but, until now, it has defied analysis because it could not be documented.

But now we have social networking. Personae on, say, Facebook, collect friends and influence them (as has always happened) – but this time every sheep thrown, every refused contact, every shared recommendation is documented.

Personae can be grouped and, inevitably, some groups have influence on others. And to make matters perfect, this data is already in a database – analysis is not just easy, it’s inevitable.

We can start to understand not just how customers interact with any given business but how they interact with each other. The implications are huge for the commercial world.