QueBIT Blog: “Build vs. Buy” decisions in Financial Performance Management and Analytics

One of the advantages of being a Business Analytics consultant, is that you rarely know what a feasible solution looks like until you have taken the time to truly understand your client’s business, strategy and culture. This applies to financial performance management (including financial reporting, planning and customer profitability modeling), just as it does to predictive and prescriptive analytics problems which may not always directly impinge on the Office of Finance.

It is an advantage because it is always fascinating, and allows you to combine business knowledge and technical expertise with social skills, emotional intelligence and creativity.

As computer systems move from tackling well defined automation problems (like managing accounting transactions, customer order processing etc.) to the fuzzier world of Business Analytics (“give the right people the right information, at the right time!”), traditional non-agile approaches to the System Development Life Cycle (SDLC) used by software engineers simply do not work because they pre-suppose that you can begin by defining clear requirements.

That is not to say that analysis, planning and design are not good, and necessary, activities. On the contrary! But in order to be truly successful, any analytics system must be dynamic, flexible and more of a tool that can be wielded effectively by a business user. The key is to use a tool that is flexible enough to deliver value, while still providing enough structure and security to be trustworthy.

In today’s business world, Excel has emerged as the premier tool used for this purpose, even though it is well known that the over use of spreadsheets brings substantial risks. Famous recent examples range from scandals in high finance (http://www.forbes.com/sites/timworstall/2013/02/13/microsofts-excel-might-be-the-most-dangerous-software-on-the-planet/) to discredited economic analyses (http://www.nytimes.com/2013/04/19/opinion/krugman-the-excel-depression.html?_r=0).

And yet the fact that it persists proves that it serves a very real need. The article (http://www.theiia.org/intAuditor/itaudit/archives/2006/january/the-role-of-spreadsheets-in-todays-corporate-climate/) says it well: .. it is rare for one software application to support an organization's entire gamut of finance-related activities and needs. In addition, because organizations change over time, gaps exist between a company's business needs and a system's capacity to fulfill them. Consequently, many companies rely on spreadsheets to help fill these gaps.

So let’s talk about “Build vs. Buy” decisions in Business Analytics. If you cannot begin by defining clear detailed requirements for (say) your financial planning model, including (say) how you plan to implement allocations – in detail - two years from now, how can you even begin to evaluate pre-built products on the market to establish which will work best?

And let us not be derailed by the argument that such systems are “configurable”. “Configurable” means one of three things:

1. There is a limited number of ways in which it may be configurable, so you need to be very sure of what you want – at a detailed level – to make sure it is configurable in all the degrees of freedom that you need now, and in the future. Which brings you back to that pesky requirements question.
2.  “Configurable” may mean that with excessive coding, a specialist can bypass or dismantle pre-built functionality to accommodate your needs. This is the worst of all worlds: you end up with something overly complicated (because it includes both the original code and your unique custom overrides), and likely difficult and expensive to maintain. Furthermore, the cost of every successive customization as your business needs change increases over time.
3. “Configurable” may mean that it is in fact custom built from scratch, in which case you are really buying a tool and building your own solution, which we are arguing is in fact the most practical and cost effective approach – provided what you end up with continues to be configurable as things change over time.

Fortunately there are Business Analytics tools on the market today that can flexibly be deployed to serve a company’s analytics needs by empowering the business user, while still providing security, auditability, structure and controls.

As Business Analytics specialists and an award winning IBM Business Partner, QueBIT’s toolkit includes IBM Cognos TM1, IBM Cognos Business Intelligence (“BI”), IBM SPSS and R. With the right tools (and we concede that these are not the only ones!) and a common sense agile implementation strategy we help customers get closer and closer to the ideal of giving “the right people the right information, at the right time!”.

This is not a pure “Build vs. Buy” decision of course, as you have to buy the tools. But once you have the tools, you gain these advantages, in addition to having a great system:

• When the business is engaged with the tool, the results will be so much better. The problem with most IT requirements gathering is that a lot is lost in translation, and sometimes the business cannot articulate exactly what they need until they have started using it.
  • A flexible tool like IBM Cognos TM1 is like a “better” Excel: all the same benefits (and more) plus controls, auditability, power and secure data management.
• Good tools will always have open interfaces for getting data in and out from other systems. The issue of whether a pre-built “out of the box” system will “integrate” with some other system goes away. For example we have deployed the IBM offerings that we work with (TM1, BI and SPSS) for customers looking for Sales and Operations Planning systems that integrate financial modeling (sales forecasting in TM1) with predictive modeling (demand predictions using statistical techniques in SPSS) and a rich visual dashboard (BI), for a fraction of the cost of pre-built systems, but with the additional benefit that it can be extended, or integrated with other systems on the platform. In particular the real time integration of the predictive demand planning model with financials is very powerful.
• The same tools can be used for new applications, and each new implementation becomes easier and more cost effective as you gain in-house experience and expertise. QueBIT’s CARE methodology actively promotes our customers’ development of in-house expertise because we believe that it leads to better and more business in the long run.

What do you think? Are you willing to change your analytics requirements to match the design of a pre-built model if it will cost a lot less, or do you see rich advanced analytics as a source of significant competitive advantage?