Two years ago, I was invited to a $300 million start-up to advise them on choosing the right "big data technology". We managed to find the right technology for their "stated" needs. However, every cell in my body wanted to tell the VP of Analytics that their pointed questions reminded me of the story of the drunk man under the lamppost [1]:
A policeman sees a drunk man looking under a streetlamp for something and asks what the drunkard has lost. He says he has lost his keys and together they look under the streetlamp. After a few minutes, the policeman asks if he is sure he lost them here, and the drunk replies, no, that he lost them in the park. The policeman asks why he is looking here, and the drunk replies, "here is the light".
The streetlight effect is a common disease of business analytics. Managers are drunk on all the buzz about big data and analytics. They have the data (the light!), and they need someone to help them find the keys to value creation. So they hire data scientists to turn that data into insights. But that does not mean that data scientists will be able to find the "keys" to value creation. The keys may lie elsewhere...
I described the problem and recommended to the VP an "Action Research" to "develop a strategy for analytics". This would help them prioritise the business questions and focus on those that could create the highest value (ROI). The VP of Analytics politely dismissed my point and said they already know such questions. Today, however, I heard that the company's revenue has dropped significantly and many employees have been laid off. It seemed foreseeable two years ago, as analytics was the core of their business model.
Asking the right question is the foundation of successful business analytics. According to Jeremy Howard, asking the right question means a) identifying the right objective , and b) choosing the right lever . Howard, one of the most famous data scientists of our time said [2]:
The important thing is to look at the question you are trying to answer. And I split that into two parts. The first is the objective you are trying to achieve. For example, insurance is aimed at the objective of maximising the profit of each customer. The second thing we look at is what I call the levers . What are the things I can change to affect the business? And in insurance, one of the most important by far is the price you set. With this in mind, maximising profit by changing price, I can now go and say what kind of model I need to build that would hook those two pieces [price and profit].
Right objectives source in business strategies. Does the company want to increase sales and market share, or is it aiming for higher profits? Does it aim to develop a brand or a new product? Or does it want to reduce the cost per customer or the cost per product? Some may want to reduce risk rather than increase value. In short, the objective must be in line with the business strategy.
Finding the right lever can be more challenging. For example, a data science team asked to find a way to increase sales of an e-business might focus on one of the following questions:
- What is the competitive price for each product?
- Which products are often bought together?
- Why did the customer, who read the details of various products, not make the purchase?
- Why do our customers rarely come back for a second purchase?
- Why do current customers buy from us, but not from our competitors?
- How can a purchase become "fun"?
Clearly, every company is faced with a list of questions that need to be answered. These questions must be prioritised based on their ROI, i.e. the value they can create as opposed to the cost of the required technologies, skills, datasets, etc. This is where it gets complex, as there are very few people capable of carrying out this prioritisation. Nevertheless, the complexity should not lead companies to neglect prioritisation, otherwise they will end up in a situation similar to the following case. According to Gartner [3, 4, 5],
A global car manufacturer decided to carry out a sentiment analysis project. Six months and $10M later, the findings from big data were distributed to all dealers. All the thousands laughed out loud because they all knew about the insight provided.
That is why I advise my data scientist fellows to focus on the keys, not the light. I hope to find time to write about how this prioritisation can be carried out. I invite you to share your views and related experiences.
References:
[1] David H. Freedman (2010). Wrong: Why Experts Keep Failing Us. Little, Brown and Company. ISBN 0-316-02378-7.
[2] https://www.youtube.com/watch?v=yPGzOw_KcBk
[3] http://searchcio.techtarget.com/news/4500251611/Seven-big-data-failures-to-watch-out-for
[4] https://datafloq.com/read/top-reasons-of-hadoop-big-data-project-failures/2185
[5] http://blogs.gartner.com/svetlana-sicular/big-botched-data/
