Business Predictions Maturity Model (BPMM):
The definite guide to evolutionary corporate adoption of predictive analytics.
The landscape of scalable big data analytics solutions is complex and often opaque for many enterprises. Trying to give value to the continuously increasing amount of enterprise data often ends in a plethora of different technologies and solutions with little interoperability.
Many predictive analytics projects still fail and gaining business decision-ready insights and foresights remains out of reach.
The BPMM addresses this issue. It defines 5 Maturity Levels and offers a 4-Step Approach for enterprises at any scale of how to successfully adopt pluggable predictive analytics.
Make it your own
PredictiveWorks. aims to provide a 360° business view onto predictive analytics:
It offers any easy-to-use point-and-click analytics solution for a wide range of use cases, but it is not limited to the technical analysis level.
The BPMM seamlessly integrates with PredictiveWorks. Each maturity level directly refers to specific platform components. This architecture supports an evolutionary approach towards predictive analytics:
Companies can decide by their own which maturity level they want to reach when.
PredictiveWorks. always comes with the right components for the right level.
Individual business sections launch hardly reusable software and data science projects which often prove fast that whatever works. Statements such as “we need a minimal viable product” or “we have to build a pilot application first” are accompanied by the dream to gain quick wins.
What is the real business benefit of such wide-spread strategies? Do we need to prove whether predictive analytics will work? Do we have alternatives when we want to become data-driven?
When we decide to build a house, do we demand a pilot, say a certain part of the rear wall, to determine whether a house works?
How to reach the next level?
Why not invest into actions to find out what is the right sustainable fundament for the future data-driven enterprise?
At the heart of each process that transforms data into valuable business insights and foresights is a single conceptual kernel: a data workflow or pipeline. This concept is supported by almost all existing data solutions & engines, from complex event processing to machine learning up to deep learning.
The problem: Each solution comes with its own implementation of the underlying pipeline technology. Building corporate data pipelines for a production-ready 360° data view on top of these individual implementations is accompanied by a high risk to fail.
The solution: Focus on the corporate adoption of the right pipeline technology first, as this is nothing less than the digital nervous system of a data-driven enterprise.
You are invited to find the right pipeline technology that fits all your current and future needs. As an alternative you may want to follow the PredictiveWorks. way and give Google’s unified pipeline technology CDAP a try. It is open-source.
CDAP is a comprehensive pipeline solution of former Cask Inc. and has been acquired by Google. It is now at the heart of Google’s Cloud Data Fusion service.
2. Data Integration
Companies at this maturity level understand that the adoption of (business-ready) predictive analytics does not start with the implementation of a certain prototype for demand prediction, say in an Internet-of-Things context.
At this level the focus is on the corporate adoption of a unified data pipeline platform that standardizes all aspects of data processing as pluggable and reusable stages.
The BPMM recommends to start with the data integration task to make data, whether located in a public or private cloud, SaaS services, RDBMS, NoSQL, streaming service, file systems, or anywhere else, accessible for analytics.
You are free to integrate all the different data formats, API, batch and streaming technologies by your own.
As an alternative, you may want to leverage a graphical interface, which allows users to quickly and easily build data integration pipelines (without the need to write code) from an increasing set of pre-built and pluggable data connectors: CDAP.
What if a certain enterprise specific datasource is not covered by the various data connectors shipped with CDAP?
Developing a custom CDAP compliant data connector is no rocket science. The benefit is obvious: You do not have to reinvent the wheel again and implement proprietary connectors to wide-spread CRM, ERP, SCM, HRM, or other SaaS application, but end up with a custom connector that automatically works with all other, say analytics components of a CDAP data pipeline.
No longer different technologies provided by a plethora of independent narrow point solutions which respond to a single highly specialized use case.
3. Pluggable Analytics
Companies at this maturity level are prepared to gain value from their data. Data connectors define the source and destination of a certain analytics pipeline and data experts begin to specify intermediate steps to transform data into more valuable insights.
Having a unified data pipeline technology at your fingertips, you can implement and deploy your own plugin components to match the needs of a certain use case from deep learning and machine learning to time series analysis or natural language processing. Or even applying business rules or SQL statements for data aggregation, grouping and filtering.
The benefit of leveraging a unified pipeline technology is that all these plugins have to be implemented only once and are immediately reusable for similar use case in the near future.
Suppose Google CDAP is your preferred pipeline platform, then you can drastically reduce the building time of your data pipeline. PredictiveWorks. offers more than 150 analytics plugins to generate insights and foresights without having to write any line of code.
4. Business Knowledge
Data pipelines are at the heart of each data-driven company. Each business question that can be answered by analyzing data and each data-driven business decision generates new or updates existing data pipelines.
In most cases the amount of these pipelines is going to increase fast.
At this maturity level, companies know that their pipelines define what they do with their data and relating them to the originating business case and its associated task manifests a huge treasure of business knowledge.
Having persisted this knowledge and made available to respond to future predictive questions or to ease onboarding of future business users is a huge benefit for data-driven enterprises and key to sustainable competitive advantage.
PredictiveWorks. is the world's first data analytics platform with an integrated pipeline knowledge system.
It introduces the concept of business (knowledge) templates and covers the complete range from specifying a business case, its decomposition into business tasks and the implementation of tasks as technical data pipelines.
Data pipelines and its business context are persisted and made available through a Template Market approach, accompanied by a recommendation system.
Companies at this level benefit from a knowledge base that makes building predictive data applications as easy as shopping for a certain product.
5. Knowledge Sharing
Companies at the previous maturity levels (still) work on their own to answer predictive business questions.
With PredictiveWorks. they benefit from a tremendous simplification of pipeline building, and, a huge reduction of one of the most important KPIs: Time-to-Insight.
At level 5, companies understand that sharing their business templates with others and taking advantage of shared templates boosts their business a second time.
PredictiveWorks. is also designed as a business template sharing platform.
For companies, who want to experience the advantages of this top maturity level, PredictiveWorks. offers pre-built prediction templates for the following use cases:
- Cyber Defense,
- AI Marketing and
These templates are offered as seeds to initiate a cross-business sharing community of approved prediction templates.