First Published on March 3 2011, 2:47 PM
In the previous post I described what the #newBI is not. In this post we will explain what the #newBI actually is, according to my view.
One step back, let’s look to a company from a very high perspective. All the people working in the company, do activities that fall into two, very large, categories: operations and control. Operations are rather self explanatory, let’s just say that they’re organized in processes. Control means studying operations, getting to know them well, find patterns, making hypothesis to improve them, changing processes, measuring results etc. Control is not necessarily an executive function. Depending on the company organization, almost everyone, at least sometimes, acts like a controller, at least of himself.
The top brass will need, most of the time, few aggregated metrics. Every business unit head will need more detailed and diverse information. Every line manager and information worker will need to go in a deeper detail. Even the single worker who works at the assembly line, might be interested in quality statistics to see his ratio of discarded pieces.
In my vision, Business Intelligence is everything IT necessary to support the control function. It is opposed to ERP, that supports operations.
ERP implicitly implements data cycles. Every atomic operation (or transaction) sparks other operations down the process. People, according to specific conditions, execute these operations within the scope of their responsibility. So operations go ahead.
What do people do, while exerting the control function, with data? The conventional BI answer is “sort, filter, drill, slice and dice” etc. Well, actually they do that, but they need to do much more. If conventional BI answers, with good approximation, the question “what happened?”, the real questions are “What’s going to happen?”,”If this goes wrong, how long am I going to take to recover?”,”Everything seems OK but, is there any trouble ahead?”, “If I do this, then what happens with that?” etc.
Why people export all those data to Excel? Most of the times, to get the answers to such questions.
This is not particularly new. The answer to this needs go under the name of “Analytics”, the new BI buzzword for the beginning of the decade. Analytics are the results of data mining or predictive tools, the output of simulation or statistical analysis applications. We already have all the technology we need for those.
What I’m arguing is that we got it all wrong. They’re usually batch applications, that produce datamarts and KPIs libraries, consumed by the conventional instruments, reports, dashboards etc.
That’s not what most users need.
Most users need something more interactive and more easy to manage. They need to change prices and see what happens, using budget data first, then projected data based on the actual and historic curves. They need to maintain a best and worst case. They need to know how an increment of 10% on sales will impact on the workforce, the associated costs and, in turn, how this will impact on the cash flow. They need to maintain KPI systems that automatically tell which are the nonperforming products and let you select them as a category. They need to calculate break evens and tipping points. They need to tag the data they find interesting and retrieve them automatically. I do not think is natural that, to find an answer to such questions, people should download data in Excel and do their calculations separately. In this way every iteration is a pain, because data are not live, they’re disconnected from the flow. Every simulation done, brings on a new simulation to be done in an endless cycle.
There are very few applications that allow the user to operate seamlessly in this way. None is comprehensive.
So this is the missing link.
This is my definition of #newBI, a Business Intelligence where users can actually use data, not just visualize them.
I too often get requests along these lines. Too often I see companies with departments filled with Excel daredevils who do nothing than producing piles and piles of data upon requests. We are so used to this idea that we find perfectly natural ask the IT to launch a batch to make a simple simulation and get the results tomorrow. While the brute machine time required to make the calculations is a matter of technology, there’s no reason why all the levers and triggers should not be in the hands of the people making the analysis.
Almost all the business decisions have an impact on the figures but they are hardly modeled in business systems. Let’s suppose your sales go well and you get a new sales manager. It will impact sales, obviously, because she will need to travel to get up to speed, maybe together with other sales people who will be doing something different than planned. It will impact production and procurement, because she’ll generate new sales. It will impact finance, because she will generate travel expenses. It will impact IT, because she’ll need a new laptop, new licenses, new accounts etc. These impacts are known, by each department, but the impact on every department and on the company as a whole is hardly or ever calculated. Get some of these bumps along the road and you risk to derail, at least in some areas.
I hope I gave a clear image of the new generation of apps I’m envisioning.
There are already some apps that implemented some of these ideas, but they are still few and marginal.
Yet, I’m not the only one who thinks like this, but this is a subject for another post.
What do you think about? Am I completely fool? Let me know your idea!
4442 views and 3 responses
Mar 11 2011, 2:41 AM
Kyle Chastain CPA, MCITP responded:
What a breath of fresh air. I started my Business Intelligence career on the user side versus the IT side. Many times I’ve felt like the polar opposite of my IT counter parts. However, here is an IT person that understands and clearly communicates the purpose and reason for Business Intelligence. The reason business intelligence exists is the end result; a sound business decision made in a timely manner that can be communicated effectively.
Mar 11 2011, 6:25 AM
Augusto Albeghi responded:
Thanks Kyle, far too kind!
When IT and business start to fight head-on, that is a real show stopper. For making what I envision real, good alignment is obviously a prerequisite.
Do you have some examples to share?
Mar 18 2011, 7:52 AM
Kyle Chastain CPA, MCITP responded:
An example that demonstrates both the good and bad is a purchase reporting system I developed. The IT department had been studying the problem for 2 years and had budget of $1.5 MM to implement a solution. However, the IT dept. didn’t understand the user’s urgent needs and was getting distracted by architecture, support and control over the data. So the purchasing department was using brute force man hours to gather some data.
At the time I was working on capital project reporting system for accounting. So in my spare time I wrote a reporting system in MS Access for the purchasing department. It cost them about $30,000 vs $1,500,000 and it took less than 2 months to write. It’s now on the desktops of about 30 users and produces about a dozen reports on their 46,000 monthly transactions that span 3 countries and 3 companies ($1.5 Billion of annual transaction volume). It was a quick development and deployment. It’s very easy for the end user to use and gives them reports and graphs they can download to Excel or Word.