ProjectBlog: Calculating your Data Science and Machine Learning Practice Equity Value

Blog: Calculating your Data Science and Machine Learning Practice Equity Value

How can you find out the real value of your Data, Data Science, Machine Learning and Artificial Intelligence Assets.

This is a simplistic effort to use rough valuations to illustrate the concept of equity value of a Data Science, Machine Learning and Artificial Intelligence practice. This IS NOT a use case ROI calculator which is often a short term, myopic approach to building a business.

This is not meant to be an exact definition but rather to create a framework for valuing data, talent, code, models, frameworks and other intrinsic assets for companies that arise from a Data Science as an asset mentality rather than the oft used “use case”, operations driven mentality.

Each $ of MLV equals a $ added to the valuation of the company and is not necessary related to cash flow. The focus for this exercise is Equity Value which is defined as the intrinsic value or book value of the company.

The basis for this model is the Uber equity value ecosystem where Michelangelo is a significant portion of the company value and the Google ecosystem where Google Brain is a significant asset of the company.


CT = Current Year + 10 years company valuation

Value of the company in 10 years

Similar to NPV and other metrics to measure current value of future assets

MLV = Data Science/ML/AI Equity Valuation. Value of entire Practice/Framework/Data to the company.

(includes all internal data, code, talent and other assets of the practice)

Talent(T) = Data Science Practitioner/Machine Learning Scientist, etc

(codes in Python or R, understands Statistics, Cloud Computing, can engage Domain Experts, deploy models, perform EDA, etc. Can essentially perform all functions from end to end of the Machine Learning deployment process.)

Valuation Factor: x * 10 yr projected annual revenues

Valuation Logic: Talent in this space is extremely scare and a significant driver of future revenues.

Models/Code(MC) = Internal(company owned) models in production, code base other productionized ML elements

Valuation Factor: x * 10 year projected revenues

Valuation Logic: A significant and important factor in valuation but rapidly commoditizing

Outsourced Models/Data(OM) = These are models, code, algorithms, data held in external vendor platforms.

-x * 10 year projected revenues

Valuation Logic: Giving external vendors data and allowing them to ingest and build on top of proprietary data and domain knowledge leaks Equity Value and diminishes the value of the company. While these engagements are often necessary they are a deleveraged activity for the AI enabled enterprise.

Data(D) = internal propriety data and open data ingested into the internal data stores.

Valuation Factor : x * data storage, x = internal valuation depending on various factors and types of data.

Valuation Formula : MLV = T + MC -OM + D

As you can see this is a very simplistic view of how to value you Machine Learning and Data Science practice. My hope is that is helps companies stop looking at their practice assets as simply an extension of operations and as a asset.

Source: Artificial Intelligence on Medium

Leave a Reply

Your email address will not be published. Required fields are marked *

Back To Top

Display your work in a bold & confident manner. Sometimes it’s easy for your creativity to stand out from the crowd.