During a recent workshop with my team, I wanted to randomize the seating arrangements. I organized a small icebreaker activity using their birthdates as the random token. To wrap up the exercise, I shared a heatmap that I found online which displayed the most and least frequent birthdates.
The team quickly noticed a massive spike in births during the month of September. They logically concluded that since September is nine months after December, the Christmas holiday season was the primary driver for this trend. It was a plausible explanation, so the team accepted it without further question.
They did not challenge the data because they had found a good hypothesis. As I discussed in my previous post on data validation, checking the validity of data with a thorough explanation is a vital leadership skill. However, there was one major problem that we overlooked: geography.
Why Data Sources Matter
The source of that specific heat map was the Australian Bureau of Statistics (ABS) 2017. European people are generally not familiar with the acronym ABS. While the data was true in a literal sense, it was not applicable to our specific context in the Northern Hemisphere.
In Australia, December is not just the Christmas season. It is also the peak of the summer holiday season. The summer holiday season correlates closely with birth rates in most developed countries. When we use the same calendar, but we apply Southern Hemisphere data to a European context without making adjustments, we miss the true explanation behind the numbers.
This is a common mistake in business, but also in our daily lives. We find one plausible explanation and we stick to it.
Actionable Takeaways for Data Integrity
To ensure that your decision-making remains sound, I recommend that you follow these steps when you analyze any set of data:
- Contextualize your source: You must understand exactly where the data originated. Even a reputable source can still be the incorrect source for your specific problem.
- Challenge inconsistencies: If a data point does not feel correct to you, you should investigate it immediately. Do not ignore loose items simply because everything else makes sense. I am not referring to the outliers, but to inconsistent data points.
- Always ask why: Data without context is nothing more than noise. You must understand the true drivers behind the numbers before you act on them.
- Check data temperature: Some data must be hot to be meaningful, meaning it must be very recent. For example, analyzing raw material prices that are one year old will probably not lead to a winning strategy when the market is constantly fluctuating. Birthdate trends over the year should remain stable, unless you look at a unique period such as the COVID-19 pandemic years.
By remaining curious and skeptical of true data that lacks context, we can make better, more accurate decisions for our organizations.