There is so much talk about when it comes to what analytics models and approaches have worked for different companies. LinkedIn is awash with happy stories of companies showcasing their analytics successes (similar, maybe, to the way Facebook can be awash with people talking about achievements in their personal lives).
We all know that Facebook is not always representative of what is really going on with someone – people tend not to write about negative or difficult things. Likewise, LinkedIn tends not to be a channel on which to share the pain and sweat companies go through first to define a data strategy, and then to execute on it. Few articles set a perspective and give the big picture.
Applying data analytics at the highest level means that a company makes decisions that are based, or at least influenced, on facts. Having a structure in place that enables this kind of decision making is win-win – even if individual projects may, and probably will fail. But getting there is hard, and often companies are not sure where to start. One of the first things that companies have to do is to understand how well they are doing now, to know how best to move forward.
There are a few models our data science consulting team likes to use to judge their existing capabilities:
One of my favorite diagrams that sets a perspective on the level of sophistication a company has is the following diagram, from the excellent Competing on Analytics: The new science of winning by Thomas H. Davenport and Jeanne G. Harris. So much can be revealed by the questions that are being asked on the data and level of analytics savviness. A company has to start somewhere, but should not remain at the bottom of the chart.
Potential competitive advantage increase with more sophisticated analytics. Companies in the bottom third of this figure, in particular, should be trying to move further up it. Your typical Business Intelligence tool (BI) will typically cover your descriptive analytics, but moving towards predictive analytics can enable companies to see problems before they arise. Prescriptive analysis then helps to specify the actions an organization should carry out, while autonomous analytics can power decisions without human interaction. Without human interaction may sound scary as jobs may be affected but the fact is that it increases productivity and more time can be spent innovating rather than investing in a manual process. Google search is a good example.
Not all organizations need to be at the autonomous analysis stage – but all need to be moving away from the bottom of the chart.
Figure 1: Analytical maturity
Analytics competition looks at the issue from a slightly different angle. To be at the bottom of this pyramid means a company is literally flying blind analytically – not a good position for any company to be in. Stage 2 reflects companies that do not have a companywide but a localized analytics effort. This may be a necessary sidestep to reach a companywide initiative, but may necessary to get general acceptance.”
Figure 2: The five stages of analytical competition
Companies at stage 3 have good analytics aspirations and are starting to show some distinctive capabilities. At stage 4 companies have some broad analytics capability and analytics is used for differentiation. Stage 5 companies are those that are masters in analytics and use analytics as a primary driver of performance and value creation.
Moving up the data analytics league
How should companies who identify themselves as close to the bottom of both the two scales above start to move up it? It can be daunting – “data” really covers technology, mathematics, business and people. We like the abbreviation of DELTTA as a way to bring together the various disciplines needed to succeed in analytics:
Data: Companies require integrated, high quality, easily accessible data and distinctive source.
Enterprise: Manage analytics resources in a coordinated fashion across the company.
Leadership: Strong, committed leadership who understand the importance of analytics and continuously advocate for analytics development and use in decisions and actions.
Targets: Organization cannot equally apply analytics on all aspects of a business. Important to target specific business capabilities
Technology: Complex technology architecture is required to process all the data - this means not just the original build, but on-going data operations aswell.
Analytics: Hire and train high-quality analysts and data scientists.
Organizations must excel in all of these categories to be successful. The target capability is an essential piece of self-discovery in the context of the business. Trying to emulate others will not give companies the edge they want. Companies must go through a discovery process to discover their Target objectives.
The discovery process requires looking at your business from an internal and external point of view. There are a number of internal areas that benefit from analytics and each needs to be taken in turn. The exercise to look at each area can be a team building exercise and also helps team set objectives with innovation where analytics provides sometimes immediate measurements. The external areas attempt to better connect and understand customers (both consumer and businesses), partner and competition.
No reporting and modeling tool can provide all the analytics capabilities to succeed. Most of the skills relate to people. We see lots of articles that believe the end game of analytics is people losing their jobs as processes are automated, and what may have required manual analysis is easily given through AI. But companies need people to succeed at data – to understand the strategy, build and run the data pipelines and infrastructure, understand how to build the most effective models and then to understand what to do with all the information that is produced.
•Note: The figures in this article come from the book Competing on Analytics: The New Science of Winning by Thomas H Davenport and Jeanne G Harris•