How to Avoid “Orphaned Analytics”
I’ve heard several clients complain about the curse of “orphaned analytics”; which are one-off analytics developed to address a specific business need but never “operationalized” or packaged for re-use across the organization. Unfortunately, many analytic organizations lack a framework for ensuring that the analytics are not being developed in a void. Organizations lack an overarching model to ensure that the resulting analytics and associated organizational intellectual capital can be captured and re-used across multiple use cases.
Without this over-arching analytics framework, organizations end up playing a game of analytics “whack-a-mole” where the analytics team focuses their precious and valuable resources on those immediate (urgent) problems, short-changing the larger, more strategic (important) analytic opportunities.
Consequently, this game of analytics “whack a mole” results in:
- Inefficient use of data engineering and data science resources
- Analytics projects unattached to high level strategic initiatives
- Limited organizational learning opportunities
- Difficult to gain strategic buy-in for investments in analytic technologies, resources, and skillsets.
- Difficult to build a track record, processes, and credibility as a trusted advisor.
- Doesn’t treat analytics as a discipline, so there are no repeatable processes developed to make the next initiatives more efficient.
How do we address this problem? Enter the Analytic Profile.
Role of Analytic Profiles
Let’s first establish some definitions:
- Analytic Profiles are structures (models) that standardize the collection, application and re-use of analytic insights around the organization’s key business entities.
- Key Business Entities are the physical entities (e.g., customers, products, employees, students, patients, parolees, trucks, wind turbines, jet engines) – sometimes called “strategic nouns” (in this context) – around which organizations seek to uncover or quantify analytic insights.
Analytic Profiles are created by applying data science (predictive and prescriptive analytics) against the organization’s growing wealth of internal and external data sources to uncover behaviors, propensities, tendencies, affinities, usage trends and patterns, interests, passions, affiliations and associations at the level of the individual entity (individual humans, individual products, individual devices). See Figure 1 for an example of an Analytic Profile that might be created for someone named Bill and his infatuation with a certain restaurant chain.
Analytic Profiles provide a framework for ensuring that the organization’s analytic efforts are being coordinated around a larger analytics master plan. Analytic Profiles enforce an organizational discipline in the capture and application of the organization’s analytic efforts to minimize the risk of creating one-off, “orphaned analytics.” Analytic profiles help organizations prioritize where and how to invest their valuable data science and analytics resources by forcing the development of the model for what analytics (e.g., scores, recommendations, rules) are needed to support the organization’s key business decisions (e.g., customer acquisition decisions, customer retention decisions, new product introduction decisions).
For example, some of the more common components of an Analytic Profile are scores and propensities:
- Analytic Scores are aggregations and weightings of different variables and metrics that in combination can indicate quality or performance, as well as predict the likelihood of future events, sentiments, and behaviors. For example, the FICO score is comprised of 35 to 45 different financial and payment metrics across 5 different financial categories that in combination create a single score that is used to help optimize a decision about whether or not to give a loan to a borrower. The FICO score predicts a borrower’s likelihood to repay a loan. Likewise, one could create a performance score that can be used to determine when to service a car or a jet engine.
- Propensities are an inclination or natural tendency to behave in a particular way. For example, a propensity can be created about one’s likelihood to advocate for a particular store, or a propensity can be created about one’s likelihood to react to a particular offer.
Example of a Score
I think most everyone is familiar with the FICO score (anyone who has watched any TV certainly understands a FICO score), which uses 35 to 45 financial and payment metrics across 5 different financial categories to create a single score that measures a borrower’s likelihood to repay a loan. There are some key points in the FICO score example:
- A score is created to support key business decisions. In the case of the FICO score, it is used to determine the likelihood of a borrower to repay a mortgage or car loan. This supports the key business decision of whether or not to offer a loan to potential borrower.
- A score is comprised of numerous metrics and variables that by themselves are not very predictive, but when combined, integrated and assigned the appropriate weights, create predictive insight.
I wrote a blog about how Nate Silver uses the concept of scores to drive his presidential election predictions (“Predicting the Iowa Caucus with Big Data”). Nate Silver builds his election predictions not on actually conducting interviews or surveys, but instead has built a model that weighs and adjusts for the biases of existing polls and surveys. Brilliant!
Another example of a score is the Athletic Intelligence Quotient (AIQ) that measures and grades a specific athlete’s intelligence that is most relevant to athletic and sport performance (see Figure 2).
Athletic intelligence is trying to predict which athletes will be successful and which ones will not, given the specific sport and position that they play. The AIQ score is based upon the concept that the “athletically intelligent” athlete can take in new information quickly, grasp complex situations quickly, and generate multiple solutions to a problem. The components of the AIQ score measure an athlete’s adaptability, flexibility, and ability to adjust successfully to new situations and include variables and metrics such as:
- Visual spatial processing
- Spatial scanning
- Visual memory
- Reaction time
- Processing speed
- Long-term memory retrieval
Remember that data science is “identifying those variables and metrics that might be better predictors of performance”. The AIQ score exemplifies that definition by identifying various variables and metrics that in combination might be better predictors of athlete performance (and ultimately, their value). And in a time when journeymen professional basketball players can sign 4 year contracts for $50M, it’s probably quite useful to have a series of scores to help sporting organizations to optimize their investment in athletes.
Intellectual Capital Capture and Re-use
This blog introduced two important concepts for helping ensure that organizations are not wasting their precious analytics and data science resources on tasks that are urgent but not necessarily important:
- Analytic Profiles provide a framework for collecting the organizations valuable analytic insights and results around the organization’s key business entities.
- Scores are useful for aligning data science and business resources to create actionable insights that can be used to help optimize key business decisions.
So analytic profiles and scores can help you avoid the course of “orphaned analytics” by:
- Leveraging analytic profiles across marketing, operations, product development, sales, etc.
- Developing benchmarks that over time contribute to the optimization of future decisions.
- Developing repeatable analytic processes to accelerate the adoption of analytics within your organization.
- Justifying investment in analytics tools and data scientists to further increase the economic value of your data.
- Extending the value of your analytic efforts by making your analytics consumable to other stakeholders for other business initiatives.
- Gaining a better understand of what data you “don’t have” but “could have.”
These concepts are consistent with my previous blog “Big Data Intellectual Capital Rubik’s Cube” where tomorrow’s winning organizations are going to be those that master the art of collecting, enhancing, sharing and re-using the intellectual capital of the modern organization – data, analytics and business use cases.