The Rubik's Cube of Business Intelligence
Arguably the greatest modern management thinker Peter Drucker said in his 2004 Harvard Business Review article that “successful leaders do not ask 'what do I want to do', they ask 'what needs to be done?'” That sums up the most important point I hope to come through in this short paper on challenges facing today’s business intelligence decisions.
It all starts not with the data (velocity, variety or volume - the Big Data 3V) as most people will say; rather it starts with the constant explosion of data platforms or approaches or tools, simply put, data technology. For example, the latest Big Data landscape (version 3.0) saw 100+ tools or ways in each one of the following 3 areas: (1) infrastructure (2) analytics software (3) applications. Then on top of that, there seems to be another 100+ approaches in terms of open source, data source/APIs. There will be millions upon millions of different combination of technology composition. It is almost impossible to find THE best technology to solve a complex data problem. In a sense, the executive demand of “give me the best BI tool” is unanswerable because every data technology or tool is the best tool for a specific set of data requirements. This is why the 32 year-old Microsoft Excel hasn’t faded away a bit from the BI landscape.
In a sense, the executive demand of “give me the best BI tool” is unanswerable because every data technology or tool is the best tool for a specific set of data requirements
This is what I called a classic “BI Dilemma of Rubik’s Cubic Proportion”. The original 3x3 cubic was invented by a Hungarian Professor of Architecture Erno Rubik in 1974 and it has 43 quintillion (43,252,003,274,489,856,000) permutations! For modern BI, besides the exploding data technology tools we just discussed, if we add the other 5 dimensions of BI’s magic cube, namely data, talent, profit, vision and execution, the # of combinations or approaches or permutations has already well exceeded 43 quintillion! That is the ultimate challenge.
The second dimension of the cubic is data. Among all the latest trends of capturing all types of data: text, voice, image, streaming, signals and others, the biggest breakthrough potential in this area is the data integration. Roughly 80 percent of all data work is data cleaning and integration while only 20 percent spent on gaining meaningful insight. The critical thinking here is how to easily, instantly and accurately mesh all kinds of data into one data repository that can be easily, instantly and accurately analyzed into actionable tips for executives and all other users. In a word, the innovation here is the simplification. Commercial leader such as Informatica has conquered a large list of data formats while its open source counterpart Talend has 500+ data connectors. However, there is no one size fits all data integration solution.
The vision and profit are really tied together as the third and fourth dimension. Today it is almost a cliché for a business leader to say data is their most critical asset. But few executives can articulate a clear and winning business vision for data. The problem is not they don’t understand the importance of clean, accurate, timely data. Rather the real mental difficulty is to imagine, enumerate and visualize what a good data insight can do for the business in terms of customer experience, bottom line profit margin, growth opportunities and like. Data leaders like Amazon, Google boldly aim for code-less, instant data storage, transformation and visualization. They must balance the human and technology cost of early vision with foreseeable profit. Nothing in the data business really comes free. Therefore, finance leaders should consider capitalizing this expense in the corporate budgeting process.
Now onto the fifth dimension of talent. Today and in the future, the bilingual (business and data) leader is the direly needed BI people but they are also severely under-valued. That explains the nascent rise of CAO (Chief Analytics Officer) or CDO (Chief Data Officer). But the problem is most of them report to CFO or CIO instead of CEO which still bounds them into a merely line item on the overall company or IT budget. Likewise, twenty first century data scientist or data manager should learn to speak both languages as well to survive and succeed. However, the fact is we just don’t have a lot of such folks from our current colleges or work places. So the second best solution is to find such dual language translator.
To get all the above five dimensions (technology, data, vision, profit, talent) to work optimally, not perfectly, we must excel in the final and sixth dimension of the BI cubic: execution. Execution is such a catch word that most of us forget the true meaning of it–simply, get things done! From the smallest task of renaming the data field, merging two columns into one calculated field, to the complex predictive data modeling, to compiling a practical BI roadmap, to envisioning a data driven business for the next 6-12 months, whoever the person, whatever the technical tool is used, whatever the business problem we need to solve, it just needs to be done.
The case in point is that Sales EVP and all his sales managers at a leading global maritime communications company are constantly struggling with multiple versions of truth regarding customer performance, not to mention the difficulty to plan future sales activities from dozens of manual or semi-manual reports from disconnected data systems. After months of meetings on requirements and technology options, there is no light at the end of tunnel until the CIO and BI Director took a pause and worked out a scrum BI project under the hook. After iterative prototyping and constant execution of each minute data decision, they did a surprise and quick demo of the MVP (minimal viable product), a simple but neat dashboard to the business users. It was a huge success with a maximum of user adoption within few weeks. It simply delivers the promise of data on the go (mobile), anytime anywhere. Rather than spending more time on traditional requirements gathering, tool selections, business feedback, the bilingual data executives there simply majored on execution, taking control of every data point decision. They took the high risk of antagonizing the users by pulling out of endless meetings but that’s exactly where a strong execution is most needed.
The secret sauce of a successful execution of a data or BI project is to find the right bilingual leader who has a practical and whole sense of all the five other dimensions (technology, data, vision, profit and talent), quickly calculating the best permutation/combination of the Rubik’s cube, roll up the sleeves and just do it.
In summary, there is no one perfect and right way to initiate and complete a data or business intelligence project. Instead of constantly demanding the latest technology or brightest data scientist, business leaders should first identify or hire the bilingual (business and data) or hexagonal leader (technology, data, vision, profit, talent and execution) whose expertise is execution.