The problem with big data and predictive analytics is that there is undoubtedly too much data and too much faith in data as the source of business insight. Everyone is talking about making businesses more data focused; but, in fact, we should be making data more business focused. There is no deficit in information, but there is a deficit in insight.
As we deploy analytics to help solve business problems, we should consider how much context we can teach computers so that they can prompt us to take the appropriate action when necessary. Data without context and then subsequent action has little value. Data with action is what drives business forward. A human with domain knowledge can look at data and figure out what to do with it, but a system does not do this on its own unless an engineer builds in that ability, since the system doesn’t know when the data is meaningful.
For example, with the power of analytics systems today, software might process a bunch of data and tell me that the average revenue I can get from all my customers is $100,000 but a particular customer that I am interacting with might have the potential to grow to $300,000. The analytics need to also tell me what I should do about it. Many analytics implementations today don’t automatically notify stakeholders about the insight that they have just uncovered or recognize the need for action.
The way to make predictive analytics powerful is to make data actionable through specialization, which requires engagement. Companies need to:
- Understand the data from the past – Descriptive Analytics. (What happened?)
- Use that data to deduce the key signals for success – Diagnostic Analytics. (Why did it happen?)
- Use the current context and forecast outcomes – Predictive Analytics. (What will happen?)
- Recommend actions using Prescriptive Analytics. (What should we do?)
The purpose of any analysis should be to get to the objective truth, the insights that will allow business people to determine the actual factors that drive the results, rather than spurious correlations that might lead them to specious predictions.
Analytics for specialized workers — those whose work requires deep expertise in their field — is about knowledge. Analytics should not be viewed as an IT project but as a business project, focused on a specialized domain area to help the knowledge workers in the business to better perform the tasks they have to execute every day. Without the right business context, it is hard to know which questions to ask, which is where the real value of analytics comes from for these people.
But if there is too much of a separation between the people with the business knowledge and the people with the analytics tools, then success is less likely. This is because a tool is only a tool, and if the analytics don’t have business expertise, domain knowledge, experience and a “nose” for what’s right, then the tool won’t offer the same qualitative input as a human worker. And that makes it hard for end users to connect the dots.
What are the important business outcomes that need to be brought about? That should drive your analytics focus. Companies might start with the Key Performance Indicators (KPIs) in their business and work back from there.
Many organizations that are investing heavily in analytics and hiring data scientists to slice and dice the data, without the requisite expertise, end up frustrated. They undeniably have more data than they had before; and in most cases (because there has been a focus on collecting it), the quality has improved. Yet results don’t come. The data scientists and software tools are rarely identified as a significant cause of the frustration, but those companies find themselves spending more time and money while failing to garner significantly better insight.
Now that technological advances have made it possible to accumulate colossal amounts of data at an increasing rate, it has become almost axiomatic that the answer to everything is in the data. But that isn’t actually true. Companies are making big bets on big data and analytics without the qualitative assessment that is required to apply deep domain expertise. This can often lead to big decisions being made with misplaced confidence.
Analytics may be the pathfinder, but the human still needs to hold the compass. The true battleground is not a certain type of analytics over another; the battleground is the knowledge worker, and the victory is in equipping the knowledge worker to make informed choices.
Donal Daly is CEO at Altify. He is the author of “TOMORROW TODAY: How AI Impacts How We Work, Live, and Think” and Amazon #1 bestseller “Account Planning in Salesforce.” Donal combines his expertise in enterprise software applications, artificial intelligence and sales methodology as he continues to transform how progressive organizations sell. Altify is Donal’s fifth global business enterprise.