Santa Claus: The Benevolent Data Pirate.
A thought experiment illustrating the importance of aligning Data Science Strategy to the overall Brand Strategy

# of Children Affected this year:

# of Individual COPPA Violations TODAY:

$ Cost of COPPA Violations THIS YEAR:

Santa's a pretty good dude, right?

• He is a globally scaled representation of multiple cultures' willingness to believe that there is some potential for good at the core of humanity.
• He represents a canonical reward-based system where good behavior nets good rewards.
• He represents a flexible system of behavioral reinforcement that allows each society to simply slot in its own moral scorecard.
• He remains consistent through centuries of human social evolution -- he's not really ever undergone a re-branding exercise, though I hear an unrelated re-org might be in order soon as his real estate up north starts to shrink and we explore the possibility that he might be a she and is most probably not caucasian. I'll keep my ear to the ground for any rogue elves that decide to sell the lists, too.

He's a pretty good dude.

But he is simultaneously one of the greatest evils to plague our society... at least through the lens of today's cultural geist.

He's checking his list...

I'm surprised there hasn't been an 8-year-old somewhere in the EU that has figured out his mythological status simply because the governing bodies have not fined him for his blatant surveillance of their online and offline behaviors. Some 'naughty' kid must be especially bent out of shape over that.

That's humorous, sure, but there is a lesson to be learned for how a well-developed, deployed, and communicated Data Science Strategy goes hand-in-hand with mass cultural appeal. And more importantly, I think it illustrates how the organization's Data Science Strategy is now no longer separable from the Brand itself.

So why are Santa's privacy violations OK?

This question extends to virtually any omniscient being -- for example, can I choose to allow a specific subset of deities access to my consolidated human logs? What recourse do I have over a deity gaining access my behavioral data without my consent? Does the reward of a free present once a year, or a painless eternity, justify the cost of being personally targetable by a geriatric toy manufacturer? What about the potential for being served perfectly-placed trials or tribulations?

These questions can lead us to a Data Science strategic framework that, when applied at even the highest levels of the Brand Strategy, allows us to pick out the nuanced human perceptions of data production, data collection, data storage, methodological usage, User vs. Brand benefits, and User vs. Brand costs. All with the goal of ensuring congruency between what a Brand says it is vs. how it uses data in its day-to-day operations.

The Branding of Data Science

Let's define four pillars of Branded Data Science that operate as opposing spectrums:

• What is the Brand's Purpose for Data Science?
  • (D) Decision Making - predictive, future-looking, making decisions moving forward
  vs.
  • (P) Proof - historical, backwards-looking, classifying decisions we've already made
• What level of Societal Focus does the Brand use for its Data Science practice area?
  • (I) Incrementality - individual actions collected and measured in isolation, allocation of synergies
  vs.
  • (O) Overall Outcome - group or societal actions collected and measured as a whole, shared synergies
• What Metrics does the Brand use?
  • (S) Single, Unified Metric - unified measurement, same metric regardless of function
  vs.
  • (J) Job-Specific Metrics - specialized measurement, different functions measured different ways
• How Rigorous is the Brand with its Data Science?
  • (O) Optimization - causal, kitchen-sink, "perfect" collection & measurement
  vs.
  • (H) Heuristic - inference, selective, "good" collection & measurement

Additionally, each of these pillars must be determined for each of the User, the Brand, and the Stakeholder.

Puting it into practice: Santa-Branded Data Science

• Santa's purpose for Data Science is resoundingly PROOF.
  
  The Santa Brand looks at the yearly performance of children to determine whether they were naughty or nice... there is no future implication, prediction, or determination being used in its day-to-day.
• Santa's societal focus for Data Science is INCREMENTALITY.
  
  The Santa Brand makes individual determinations and is concerned with attributing credit to specific children.
• Santa's metrics for Data Science are a mix of both SINGLE and JOB-SPECIFIC.
  
  According to Santa's model, all children are treated equally and are deemed either Naughty or Nice -- that's a single, unified metric. However, each society has its own definition of what Naughty or Nice actually is... that is a culturally-specific metric.
• Santa's Data Science rigor is based on HUERISTICS.
  
  According to Santa's model, he knows of everything that you do, but he uses only a single evaluation (an aggregation) to make a determination of your status. Additionally, Santa only stores your classification at any given point in time, not an exhaustive list of every action. As far as I know, this is a lossy storage algorithm, but then again... they might have a mythological encoding holding the whole thing together.

If Santa changed his Data Science behaviors on any one of those pillars, we would suddenly have a completely different version of the Santa Brand. For example, a DECISION purposed Santa would give presents mid-year based on predicted behaviors or send notifications when your list status changed.

Conclusions

This is not to suggest that the PIS|JH Data Strategy is a perfect fit for Santa, nor is it suggesting that a PIS|JH strategy is the canonical Data Science positioning. Instead, it illustrates the need to form an intentioned Data Science Strategy in conjunction with Brand Strategy -- Branded Data Science


Cheers!