a societal pitfall from over-confidence in data

I think the hard problems of integrating Bayes into business is when mixed methods of qualitative and quantitative analysis are required. Finding an appropriate balance between the two (inside/outside view, Kahneman) might make for a good measure of a group’s ability to avoid fallacies of data use, as well as the group’s ability to model some system. These are not independent traits: fallacies create biases in the model.

The nature of Bayes in business is that observations reduce uncertainties at some economic cost. The power of observational technology is increasing while the economic cost of implementing is decreasing. This is independent of our ability to quantitatively model observational output. Maybe this ability to quantitatively model is what allows the quantitative to consume the qualitative. However, this ‘consumption’ is happening slower than people predict, because (1) a bubble is profitable until it pops, and (2) the substitution bias makes us think reality is far simpler than it actually is. I believe that the inflated expectations are causing pre-mature consumption (we are giving it too much attention), and that its result could be harmful.

An example to try and explain more concretely: I feel strong aversion to an aspect of the quantified self movement. The emotional aversion primarily comes from its toll on my attention, which is my favorite resource. But I am also averse because of my hypothesis that people will turn to it for personal solutions insofar as it provides personal metrics. There is definitely something to be had here, but the warning comes from the fact that the incentives and frameworks for the producers of these personal metrics are radically independent from the incentives of the consumers. The producers are driven by capital and see these self-quantifying tools either as a feeding ground, or as an end in itself.

In contrast, the idea behind user-centered design is focused on need finding: profit is viewed as a consequence. Start-ups may be a place for hope in this respect, since their goal is synonymous with user-centered design. The difference between start-ups and established organizations is that for a start-up to succeed, it has to be valued the most by its users. For an established company to succeed, it only has to drive up revenue. The transition can be seen as scaling becomes relevant: small segment quality gets consumed by quantity. Google makes for a good example of how this inevitably demotes to capitalism (the case is arguably much larger for Google given that the resource they are fighting for is breadth of attention). As capitalism grows, misaligned incentives between consumers and producers becomes problematic.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s