Sign In

More Data, More Problems?

2 minutes

More data does not necessarily mean better decisions. In fact, for many business decision problems, more data can make solving the problem harder, not easier. The belief that more data inevitably produces better results only holds for a small class of well-behaved problems. So, what should we be trying to acquire instead?

More Data, Better Decisions?

Everyone has become convinced that more data is better. But is it? Have business decisions gotten better, or worse? More importantly, can we accurately measure how increasing data has affected decision quality?

The assumption is that more data will result in better decisions. The action is to collect more data which becomes the entire focus. The result is that decisions are no better if not worse.

Data Has Limits

Analysts (Analytics experts) love data. It is their bread and butter. But data can only go so far in explaining phenomena with complex underlying mechanisms. Simplifying those mechanisms is therefore essential to extracting useful information from available data.

An analytics expert's task is to forecast future information. Taking past data modeling it to extract a mechanism and using that to forecast future information only works if past data captures all factors defining the mechanism. Factors affecting stable phenomena such as language or maintenance can be captured well with data. However, it is nearly impossible to capture all factors defining social phenomena. Thus, more data will compound conflicting mechanisms and worsen forecasts.
What will actually help analytics experts is reducing mechanism factors. If problem complexity is reduced this will result in reduced mechanism factors thus enabling better forecasts. The data effectiveness methodology simplifies problem to only rate calculation. The business problem framework defines the problem mechanisms and parameters. The inductive approach allows for problem expert guidance and review.

Consultants (Subject experts) also love data, and the expectation is that it will enable better recommendations. But as more data is collected, the problem can appear more confusing, conflicting, and incomplete. The solution isn’t more data, but suitable problem expertise to complement their subject expertise in order to recommend options.

Subject experts are tasked with recommending better options. They study past data to acquire problem expertise. This only works if past data captures all elements of required expertise. Basic business mechanisms such as staffing and pricing can be well understood with only data. However, data can increase confusion when the goals is to obtain needed problem expertise. More data will compound problem ambiguity.
What will actually help subject experts is complementing their subject expertise with problem expertise. Subject expertise coupled with problem expertise will result in qualitative expertise. Data effectiveness methodology utilizes the scientific method for option evaluation. The business problem framework creates a problem structure that combines bot sources of expertise. The inductive approach allows for problem expert review and approval before testing.

Customers (Decision owners) share the love of data, assuming its mere presence will help them make better decisions. But without the analytics expertise and time to analyze it, data becomes a burden that can lead to worse decisions often shielded from scrutiny simply because they are based on “facts and figures”. What customers really need are the optimal actions and not the data itself.

Decision owners have the task of making the best decisions. Discovering future information from past data and thus arriving at optimal decisions only works if the decision owner has the necessary discovery skills. Only precise actions are immediately useful everything else requires further analysis. Decision owner doesn't have the required expertise to make use of past data. More data will be misused and lead to worse decisions.
What will help decision owners is leveraging their problem expertise to solve their problem. When problem expertise is combined with optimization expertise, analytics expertise, and subject expertise we can obtain optimal decisions. The data effectiveness methodology creates the necessary problem framework for solution optimization. The business problem framework coordinates and maximizes the contribution from each source of expertise. The inductive approach enables problem expert to guide problem and reduce scope of work.

Data vs. Information

There is a fundamental difference between data and information. We need information to make better decisions, while data is only one potential source. The challenge is to structure problems so available data produces useful information and gaps can be filled by other sources.

We need to focus on information. More data will not result in better decisions. More information will result in better decisions. Advanata structures the problem to obtain information.

Structure for Information

The frantic rush to collect data should be considered in the context of the problems being solved. Some problems can benefit greatly from more data, while others, including many business decision problems, can actually suffer from it. Therefore, the goal should not be to gather more data, but to structure problems to maximize the information we can obtain from data so that we can obtain precise solutions to our problems.

One quick poll question (3 seconds):
I’d really value your input.

Sign up for early access to articles, delivered every two weeks. Please check your spam folder for the confirmation email.

Ready to see Advanata in action? Explore more resources, read related articles, and access the full application at advanata.com.

Looking for a personalized walkthrough or decision intelligence consulting? Contact info@advanata.com to get application access, schedule a demo, or discuss your decision optimization challenges.

Leave a comment