Meaningless Metrics

It surprises me a bit how many people, particularly people who should know better by virtue of their profession, are still obsessed with essentially meaningless numbers.   How many unique visitors arrive at their Web sites, how many times a given page is viewed, how many people "friend" their brand and "like" their comments, and obsess over numeric measurements of activities that have nothing to do with anything.

Part of it is the ongoing obsession with numerology: human experience is reduced to numbers and tortured with implements of statistical analysis to make it confess to whatever the inquisitor cares to hear.   When a number is assigned to something, there's little question as to whether the number is meaningful - but great interest is given to making it better.   And anything that is not measurable must not matter - or if there is a sense that it does matter, then lets' find a way to force numbers upon it.

No-one has yet been able to tender a satisfactory common-sense explanation of why it is important to have more - it is instead taken for granted that more is always better.   But it's possible, and has often been accomplished at the cost of many thousands of dollars, to have more hits, more visitors, more friends, and more likes without providing one more product to one more person that might benefit from owning it.   It's been shown that dumping millions of unwary users on a site's home page doesn't get an equal proportion of increased revenue.

And yet many continue to behave as it would: if a site has a 1.5% conversion ration and you hoodwink 100,000 people into visiting, that's 1500 new customers.  Funny it never works out that way afterwards.  Funnier that in spite of repeated failures, getting more visitors is still a high priority.

My sense in all of this is that the metrics are leading us in the wrong direction.  We don't need to draw a bigger audience, but a more interested audience, and we don't need hordes to "like" a brand, but a much smaller number of people to buy the product, hopefully more often.   All hits are not created equal, nor is there a strong correlation.

The goal of interaction is not to interact with a mass of steadfastly disinterested individuals, but to engage with those who have some level of interest, and who might be inclined to show greater interest.  It's much more difficult to gauge that, because there are no numbers that can be derived, and even assessing a person's likeliness to purchase based on various parameters in a regression analysis is seldom more than a desperate attempt to quantify something that defies quantification ... but is nonetheless extremely important.