Lets Talk About Data, Data Sources, How We Look at Data.

"30% of kids finish high-school in the USA"

(Bill Gates, TED 2009)"

For a long time, this figure was hidden because the system was only measuring the "drop out rate" from the start of final year to end of final year, not from start of high school to finishing high school. Oh yeah, if you are poor, and a minority, it ain' looking too good for you. You have a higher chance of going to jail than finishing a four year college degree. Turns out the big difference is good teachers, getting them, rewarding them, celebrating them, keeping them. A good teacher will increase the performance of a class by 10% right away. It would remove the difference between the performance of USA vs Asia in Education. Seniority, having a masters degree; no effect on being a good teacher. It just seems that some people are great teachers and we have not the first clue why. The only thing we know is that past performance is a great predictor of future performance.

In the Q&A at Ted with Bill Gates he goes on speak about the problem of Malaria and African poverty illustrating that as you improve health rates, adults need less children to be born, because there is a higher chance of having children that survive into adulthood that can in turn look after you (the oldest pension scheme in the world). As family size decreases, the average wealth per family increases. And now the link. Another one of the Ted Talks investigates the predictive power of interventions with regards Aids in Africa: how about this: your likelihood of engaging in safe practices and heath enhancing activities is directly related to your already existing expectations of longevity. If you expect to die of malaria in the next 10 years, then you are more likely to engage in risky practices in the near term. In other presentations, free trade and exporting are also shown as being directly correlated with higher income levels, and longer life, but what if doubling free trade actually quadrupled the incidence of aids in a particular area? The more physical movement there is, the more contagions spread. The take away being that spending $50bn on education as an intervention for Aids, might be misplaced. Its not more awareness that a condom might save your life, but that you will live longer because the system is going to ensure that you don't die in childbirth, that you have nets to fight malaria, that you have access to micro-capital to build a small business.

What I like about these accounts is that they demonstrate that sometimes what you think you know is wrong. The data is "showing you to increase spending in education to impact aids, as education changes behaviour" (sic). We know that good teachers and good teaching drives student performance, yet we have no granularity, DATA or predictability that helps this to be managed. In other words, we have no causality. In (oh dear, yes 'my masters programme') we came across concepts such as "faulty causality", and "statistical artifacts". I wonder how many of these "facts" surround us every day in the assumptions that underpin how we manage our businesses and relationships? Facts such as "this is an unprofitable customer"?

IT Investment As Barrier To Entry

Facts such as "IT helps smaller players compete globally",(Andrew McAfee)  and the rise of Free IT such as Google Mail, Docs, and eePC's, lower the barriers to entry associated with starting a business. We hear this a lot. We hear that Enterprise 2.0 will empower the end user, re-shape the corporation, perhaps even redefine what it is to be a corporation. But perhaps it also helps larger organisations overcome the drawbacks associated with scale (over formalisation, speed, flexibility) and actually makes larger companies more efficient and effective than smaller competitors. One of the reasons being pointed to is that technology (and its cost) is one thing, but getting more and better data, and making better decisions based on that data, is another. And the problem with that is that people make decisions, and very often, we make decisions based upon incorrect assumptions, faulty causality, and statistical artifacts with no true predictive power. 

Information as As Asset and 'Competence'

Bruce MacVarish has a great take on what this means for enterprise systems. With relatively abundantly available "Technology" (i.e. Technology is not the scarcity asset), the ability to master concepts such as sensing, flow, collective intelligence, and collaboration become key. It is no surprise that these are "soft competencies", "tacit knowledge based", and "culturally embedded". All incredibly hard to develop and incredibly hard to emulate. What McAfee's points out is that the knowledge and data accumulated will have disproportionate effects on market concentration ratios (i.e. there can only be one market maker such as Google, this is the nature of platform economics).  So you ring up two credit card companies and ask for another credit card: the first one knows who you are, who your friends are, what their credit ratings are, and asks a friend of yours how good a credit risk you are (via sms/ voice call), and at the end of a two minute call tells you that you can either pick up your card at the local shop (where a card is being encoded, and which is now expecting you). Oh, you didn't even call the second company silly..... the killer point is that they knew a friend of yours that would vouch for you, which had a higher predictive value than any geo-targeting scheme. And thus, the mass of relationship data has a steadily increasing marginally impact.

Processes Attract Conversations, and Vice Versa

Social Computing Magazine has a great example of this kind of thinking in progress and it deservedly received much commentary this week. I believe it contains some "evident truths" which we've held here at VoiceSage for some time: I have messed with the semantics a bit to put our slant on it:

- Processes That Attract Conversations

- Processes In Support of Conversations

"A good example of Conversation -> Process Integration was recent demonstrated, but to elaborate, by pulling Tweets into the SAP Business Suite and applying a sentiment engine to those tweets, a customer service rep can make those conversations actionable by identifying and emerging customer or brand issue. Someone may be complaining about your product or service. With Sentiment Analysis not only can an organization proactively address a looming customer crisis, but they can initiate corporate processes such as raising a Customer Service Ticket to initiate a problem resolution process"

"Going the other way, a super example of Process -> Conversation Integration is the deployment of Marketing Campaigns using Social Channels. Using Business Suite functionality, users can now design and deploy marketing campaigns which can execute over a variety of social environments, including Twitter"

Now I think that this is a pretty "shallow example" of the potentials in the MacVarish model, but it gives you a sense that the big vendors are definitely getting it. They are not "getting it big" yet, to do that, they will have to come along to hear VoiceSage at the eComm09 talk in March, in sunny San Fran. :)