Analytics Is a Profession — The Case for Professionalization

Tristan Handy argues that analytics is irreducibly hard and should follow the path of medicine, law, and engineering: professionalization with formal training and credentialing.

The zoom-level problem

Journalists maintain a consistent zoom level throughout their process — always constructing the narrative. Data professionals constantly shift zoom levels: zoomed out on the business problem, then zoomed way in on individual anomalous records, then back out to integrate into a model. This constant alteration is disorienting and demands rare cognitive flexibility.

The data-stakeholder interface spectrum

What data professionals deliver ranges from least to most judgment-intensive:

1. Datasets for self-serve
2. Interactive data products for self-serve
3. Analytical assets answering a specific question
4. Explanatory narratives
5. Business recommendations

Each step requires increasingly more judgment, context, and experience. We expect analysts to do all five.

Three paths for hard professions

1. Specialize — decompose into component parts, create an assembly line
2. Automate — externalize expert knowledge into technology
3. Professionalize — if irreducibly hard, create supports (training, credentialing) to help humans do the hard job well

Handy’s prediction: analytics will follow path #3 — becoming a credentialed field.

Connects to analytics craft, data team operations, analytics at a crossroads.

Open questions

• What would an analytics credentialing body look like? Who runs it?
• Is there a hybrid path where we specialize some parts and professionalize the irreducible core?