So far in this series , we’ve discussed at length the ways in which you can get a Business Intelligence solution introduced and implemented in your organization. We’ve highlighted the need for building a visually pleasing set of dashboards that are properly scoped for the intended audience , and we’ve laid out a strategy for building a data-focused culture.

While all the above are incredibly beneficial and sure to provide dividends, there is one important (perhaps THE MOST important) facet to emphasize: the data itself! Without proper care and planning, your whole initiative will be DOA if the underlying data is not properly harnessed and presented.

Garbage in/Garbage out

It goes without saying that if your dataset contains gaps or incorrect records, then it is impossible to build a meaningful report out of it. Oftentimes, these gaps or inaccuracies are the byproduct of process failures (either flawed processes or processes not followed properly). Before starting down the path of rolling out a BI solution to the powers that be, it is a good idea to assess the data first and look for such issues.

Fortunately, you can use your chosen BI solution to identify those gaps. Try creating a few trial reports and compare their results to what you know is true. Dig into the deviations and assess why they exist. If possible, go back and correct errors or fill in gaps in your production systems. Update your business processes and/or work with the teams to minimize such errors in the future.

Jumping to conclusions

As the saying goes: “Correlation does not imply causation.” A study in the 1980s of New York City crime noted a correlation between crime and the amount of ice cream being sold. Does that mean that buying ice cream causes crime or vice versa? Of course not. A more plausible explanation is that both of these things are more likely to occur during the warmer months. When building out your dashboards, watch out for this sort of data fallacy so as not to mislead your audience.

Context matters

Most data points are mere indicators or pointers to the “truth” or “reality.” Taking a series of KPIs and presenting them in a vacuum without additional context is an easy way to confuse your viewers or induce them to create their own takeaways that may not match reality.

Consider each dashboard component as one color of the palette or one stroke of the brush of the larger painting. Each of the elements ought to supplement the others to help create a fuller depiction of reality. For example, the KPIs should be related to each other, and there should be supporting pieces (such as visuals or performance over time) that will augment the main focus of the report.

PEMDAS

It is easy to get tripped up by elementary arithmetic when dealing with complicated dashboard components, so taking the time to ensure that the data remains accurate as it is sliced and diced is crucial. Don’t forget your order of operations for any formulas or calculations that your dashboard will be computing. A common mistake made by beginners is to fail to recognize that the sum of averages is not the same as sum of the data divided by the number of data points. Fortunately, this error is typically obvious in the results and easily rectified.

Set it and forget it

A properly setup BI system should be able to automatically refresh itself with the latest data so you should not need to spend much time or effort maintaining the reports on a daily or weekly basis. That said, you should still allocate some time on periodically reviewing the scope, the elements, and the data itself. Incrementally adding value to your reports is key to keeping your organization’s data culture thriving!

The post Common Data Mistakes appeared first on Fairdinkum.