Agile BI: Welcome to the Business Model Generation
How a Business Model Canvas and BI Model Canvas can help accelerate your modeling work.
By Lawrence Corr
[Editor's note: Lawrence is presenting the keynote address (From Business Modeling to BI Models) -- in which he will share ideas about data modeling techniques that actively engage the business -- at the TDWI World Conference in San Diego (August 18-23, 2013). He is also conducting a session on agile dimensional modeling at the conference.]
Traditionally, BI and data warehousing has been reactive: lagged behind operational development (in technology as well as chronology). Historically, data warehouses or data marts were often built after operational databases had been found to be inadequate for reporting purposes, and significant BI backlogs had built up.
Today, agile BI is going proactive. The two different worlds of operational and analytical development are becoming parallel worlds in which new BI systems need to go live concurrently with innovative new business processes and their matching operational systems.
With this proactive aim in mind, Business Model Generation by Alex Osterwalder and Yves Pigneur (Wiley, 2010) should make interesting reading for agile BI practitioners eager to actively involve BI stakeholders and business-model innovators early on in the design of future BI systems. With more than 650K copies sold (in 26 languages) and millions of downloads of the Business Model Canvas (see businessmodelgeneration.com) described in the book, it belies the opinion that business people have no appetite for data modeling -- it just has to be the right kind of data modeling.
The book describes a "business model" as "the rationale of how an organization creates, delivers, and captures value". One of the book's main contributors (it has more than 450 from 45 countries) describes it as "the logic by which an enterprise sustains itself financially."
The Business Model Canvas is a deceptively simple visual template for consistently mapping out nine building blocks (customer segments, customer relationships, distribution channels, partners, revenue streams, costs, key activities, resources, and core value propositions) of a business model.
Business Model Canvas practitioners seeking to understand and improve existing business models and develop innovative new ones are encouraged to use the canvas collaboratively: draw it big on the wall. Post colorful sticky notes for each of the building blocks. Use words and simple pictures to describe them to aid 'big picture' understanding. Sketch alternative models using its consistent layout for easy comparison.
Using the canvas in this way is similar to many other "stand-up" modeling and planning activities that are part and parcel of agile practices such as modelstorming, Scrum, and XP. From an agile BI perspective, the results can usefully be viewed as high-level conceptual data models for designing data warehouses and BI systems intent on measuring business-model efficiency.
One BI specific adaptation of this business modeling approach is the BI Model Canvas, part of business event analysis and modeling (BEAM), which is a set of collaborative techniques for modelstorming BI data requirements and translating them into dimensional models on an agile timescale. This canvas provides a consistent layout for modeling the 7 dimensional types -- who, what, when, where, how many, why, and how -- used to describe business events and how they are measured.
You can download a copy of the BI Model Canvas from modelstorming.com. You may find there's something about the visual and kinesthetic pleasures of placing and replacing stickies on a star schema canvas that really animates the dimensional modeling process and increases collaboration compared to modeling directly onto a blank whiteboard: stickies are easy to move (when you realize you have more dimensions or facts to squeeze in) or need to fix mistakes. They accelerate modelstorming because people don't think they have to define everything perfectly the first time and several people can post data definitions at once.
If you find either the Business Model Canvas or the BI Model Canvas useful additional tools for agile BI, welcome to the business model generation.
Lawrence Corr is an experienced data warehouse designer, coach, and educator. He specializes in helping organizations to use simpler, more inclusive BI design techniques. He is the lead author of Agile Data Warehouse Design: Collaborative Dimensional Modeling, from Whiteboard to Star Schema.