When searching for interesting data use cases within an organization, businesses most often look for data that is known to exist, easily available and structured, i.e., where the light is...
"During a dark night, a policeman sees a man in the distance, searching for something under a streetlight. When he approaches, he notices that the man is fairly drunk. He asks what he is looking for, upon which the drunk man replies saying that he has lost his keys. The policeman proposes to help and together they start looking for the keys. After quite some time of searching, the policeman asks if the drunkard is certain this is where he has lost his keys. The drunk man replies “oh, no”, and informs the officer that he has lost them 500 meters further down the road in the park. The agitated policeman asks “Why on earth are you looking for them here then?” Upon which the drunk man simply replies, “because this is where the light is.”
What is streetlight effect?
This old joke illustrates a type of observational bias that is called the “streetlight effect”, also known as “the drunkard’s search”. It implies that people tend to search for something where it easiest to look. Unless the right methodology is followed, we have found out that this bias will also manifest in data projects, ultimately leading to organizations producing low-value and little-actionable insights for its end users.
When searching for interesting data use cases within an organization, businesses most often look for data that is known to exist, easily available and structured, i.e., where the light is. This will for example lead their search to data that is stored in a data warehouse, or contained within an ERP or CRM system, because those sources are easily searchable, and the data are easily extractable. Someone from the data team will identify an interesting data set and start exploring it. He will then iterate a few times with the business users to make the dashboard interesting and useful, but at that point, you will unfortunately have lost most of the business value already. Because from then on, you are developing a solution looking for a problem.
Chances are high that the dashboard will end up collecting digital dust on someone’s laptop, simply because it has been designed the wrong way round. At DataMotive, we believe that data projects should not be data-centric, they shouldn’t start from the data itself. The best, most value-adding data projects start from the business. They start from the organization’s strategic goals and key objectives, which makes it possible to identify those data use cases that best serve the organization’s needs.
Problem with data-centric
The main problem with the data-centric way is that the end user is presented by information that is not directly tied to a decision he must take or to an activity he must undertake, as these decisions or activities were not considered when initially designing the dashboard. It is paramount to take the business process aspect into consideration if you want a data solution to be used within the organization.
Turning the mindset around, and not starting from the data at hand, but rather from the business side of things, leads to more creative data solutions that ultimately lead to a greater adoption by the end users. At DataMotive, we believe that a good data solution should (1) have business value and a positive ROI, (2) trigger action by its users and (3) have been designed with the business users’ process in mind.
If you want to know how we do this at DataMotive, feel free to reach out to firstname.lastname@example.org!