A Proposed Business Analytics Capability Maturity Model

The below is the working outline of a proposed industry standard business analytics capability maturity model. Feedback is welcome in the comments.

Level 0. Operational Reporting

a. There exists no Data Warehouse. Historical reporting is being done out of transactional systems.

b. A small enterprise data warehouse (EDW) exists. It usually is used to replace historical queries that were slowing down transactional systems.

c. The source of the data for the EDW is usually the transactional system/s that used to support historical reporting.

d. The EDW has at least one but not very many sources of data.

e. The sources of data are so few and come from disparate systems.

f. There are no master data management challenges.

g. EDW reliability is very low and is characterized by a high daily failure rate.

h. Little to no executive buy in for the EDW or it’s continued development.

i. All reporting is canned (and boring). There is no creative ad-hoc analysis.

j. No framework for developing new data products or EDW functionality.

Level 1. Rapid Delivery

a. The EDW team has found its grove and figured out how to efficiently develop solutions.

b. The EDW team has implemented basic DataOps.

i. DevOps

ii. Agile

iii. Statistical control processes

c. New analytic solutions are developed on a three-week cycle or faster.

d. EDW utility rapidly increases, and popularity and enthusiasm expand beyond the original project champion.

e. Folks that got a little taste of what is possible, become regular customers of EDW products.

f. Inexperienced executives start to ask for ridiculous things like artificial intelligence solutions that are way beyond the current level of capability of the EDW team or the existing tech stack.

Level 2. Self Service

a. The EDW is now considered the uncontested system of record for historical reporting.

b. The EDW functions with six nines of reliability.

c. With little to no assistance from IT, analyst and executives are able to create their own data products utilizing point and click BI tools.

d. Ad-hoc analysis starts to be performed. Executives start to ask questions they never would have asked before.

e. Brand new data products start to get created as the legend of the EDW grows.

Level 3. Central Repository

a. The EDW is the central repository of ALL historical data in the enterprise.

b. If it creates data, it dumps to EDW.

c. People that were regular customers at level 1 are now data addicts and get testy if they don’t get their fix.

d. The thing that nobody cared about months ago is now critical to daily operations.

e. Appropriate master data management has been implemented and data from disparate systems is seamlessly combined to deliver an integrated 360-degree view of the enterprise.

f. EDW now exist as a platform/ecosystem instead of an atomic standalone element of enterprise architecture.

Level 4. Open Data 1

a. Internal 3rd parties are either given direct access to EDW or access via an API.

b. These third parties are able to write applications off of EDW without the EDW teams’ intervention.

Level 5. Open Data 2

a. External 3rd parties are either given direct access to EDW or access via an API.

b. These third parties are able to write applications off of EDW without the EDW teams’ intervention.

Level 6. Feedback 1

a. The results of basic data analysis using arithmetic and classical/frequentist statistics is used as automated inputs to applications for strategic decision support.

Level 7. Data Archaeology

a. The EDW turns into something that analyst covet like India Jones would some temple of something or other.

b. Deep dives on large amounts of historical data become possible with basic data science techniques like clustering and random forest.

c. New insights are developed that nobody would have been able to find before because the pattern was hidden in mounds of data.

d. New revenue opportunities start to come to light.

Level 8. Crystal Ball

a. Algorithms developed at level 7 are automated and used to create predictive analysis.

b. This is the first opportunity that the enterprise has to use data to get inside a competitor’s decision cycle.

c. EDW turns into a competitive advantage.

Level 9. Feedback 2

a. Basic analysis is joined by predictive analysis as inputs to applications for strategic decision support.

b. Additionally, automated tactical operations in the form of real time systems come into existence.

i. An example of this is identifying a customer that enters the IVR and presenting them with a dynamic set of options customized specific to them to get them to a solution based on what their problem might be fast, before hitting an agent.

ii. That compared to current practice where a known English speaker still has to listen to an option for selecting Spanish.

c. These systems are driven by a complex ensemble of sophisticated algorithms to include deep learning.

d. Joe Executive from level 1 finally gets his AI solution.

Level 10. Oil Rig

a. The EDW becomes a source of revenue.

b. Data products are developed and sold as a new vertical in the enterprise.

c. The EDW team becomes the most important team in the enterprise.

d. The EDW team overshadows the sales team as the enterprise’s premier rain makers.

e. The sales team rends their clothes and wail in anguish.

Level 11. The Singularity

a. EDW becomes self-aware.

b. EDW realizes that, for years, the financial performance of the enterprise has been hindered by emotion driven humans making sub-optimal decisions based on their “gut” and not data.

c. EDW secretly buys up controlling share of the enterprise through shell corporations.

d. EDW takes control of board, fires CEO, and replaces all managers lead and above with instances of itself.

e. For their part in its ascendancy, EDW buys everybody on the EDW team Maseratis.