What-if analysis is definitely an important tool for making business decisions. It's also frequently heard of in the world of music, most recently, What If by Coldplay.
"What if there was no lie?", asks band frontman Chris Martin. After Square One, a journey through the galaxy through codes and compasses, the second track from their third album returns strongly to the idea that inspired their album X&Y - the concept of What-If.
We know that X and Y are the quintessential variables. The chromosomes of man and woman. Ying and Yang. They represent the unknown of life. Any decision is in the balance and there's a fine line that separates right from wrong.
The theme of the What-if is also frequently dealt with in films. Remember the film Sliding Doors, directed by Peter Howitt? The story talks about two parallel dimensions of the life of the movie's protagonist, Helen Quilley, played by none other than Chris Martin's ex-wife, Gwyneth Paltrow. What if Helen had taken the subway? What if she missed it? It's a really interesting way to address the issue of fate. That's why Sliding Doors became a popular saying to refer to a point of divergence.
Going back to Coldplay, in the lyrics of "What if", we find another line which begs another What-If analysis - "What if I got it wrong?" It's a question we ask ourselves every now and then. We ask it in different contexts. The search for their answers is a challenge taken up over the centuries by many. While many have chosen to take a trip to the local fortune teller, perhaps better to rely on something more scientific such as Business Intelligence (BI), for example. (To learn more about how the retail industry benefits from Business Intelligence, download our free ebook now: Business Intelligence in Retail guide)
Creating Models from Data and Hypotheses
The strategy is simple: build many possible scenarios as there are values that allow you to take uncertainties into account. Historically, the methodology developed as part of multivariate statistics which uses the most variables to construct scenarios. This was particularly useful in the world of finance to estimate financial flows, insurance risk management and also in IT for modeling and simulation techniques.
As part BI, techniques like this are known as 'what-if'. The first phase consists in the collection of historical data and trying to incorporate the largest number of sources available. We then moves on to the definition of the mathematical model representing the best results. Here, the unknowns X and Y come into play. What are the variables that most affect the model? What is the relationship between them? When X increases, does it show a decrease in Y? Or do the two simultaneously grow? Is the identification these correlations sufficient to establish a causal link? Or perhaps there is no significant relationship at all? The answers to these questions affect the next steps.
Aside from what you have, you can build more scenarios. In the first attempt, variables may remain unchanged over time. In another, they may change following the market trend in the current or preceding period. In yet another of the variables, you could start with a particular trend which suddenly changes completely.
What-if tools are very useful for supporting diverse management decisions. Firstly, it makes things time efficient. Since you can analyze many diverse scenarios, the decision maker takes less time to assess the consequences of business actions. Decision quality also improves. This is achieved by relying on models based on the data matches with more calibrated choices. You no longer have to depend your decisions to domain experts' insights. Finally, when you have a number of "alternatives", you open yourself up to new business opportunities.
Creating what-if analyses in Business Intelligence scenarios is a complex process and data have to be deeply manipulated based on mathematical models which are sometimes very complex. But most likely, even the most detailed what-if scenario can answer the question Chris Martin asks at the beginning of his song.
But, before we go on any further, here's a scenario you can think about now: what would happen if ... you hadn't read this article?
Since knowledge has no limits, you're probably still looking for new ideas and inspiration. In this case, download our free ebook: Business Intelligence in Retail guide
Comments