In this age of information, many leaders strive for a more data-driven culture, but they're often missing an important component to realize this goal. To improve the analytic capability within your organization, you'll need to change the way your workforce thinks about data.
If left unmanaged, about 75% of your workforce does not put a high enough value on data to support a data-driven culture. Although you may be passionate about data-driven analysis, and your data scientists certainly hold data in high regard, the rest of your workforce will need to be persuaded.
As Aristotle teaches us, there are three methods of persuasion: ethos (credibility), pathos (emotion), and logos (logic). All three are important, but today we'll focus on ethos. If you're trying to build a data-driven culture, data credibility is an aspect you can't ignore.
You must give your organization a good reason for trusting your data, and it starts with data quality. I've worked with a number of organizations on building a data-driven culture, and this is the number one reason why people are reluctant to rely on data. If people are unsure about the quality of the base data going into your systems, all trust will be lost with what's coming out of your system. This is where your data scientists come in very handy.
First, remember that data scientists are more than number crunchers — they're data professionals. Data quality is fundamental to good data management, so your data scientists should be intimately familiar with how to make it work. Your ETL (Extract, Transform, Load) should include data cleansing and integrity rule checking; good Master Data Management (MDM) should be in place for dimensions that source from multiple places; and database constraints (e.g., referential integrity) must be evaluated every time new data is loaded into the system.
Furthermore, leverage your data scientists' analytical skills to develop robust data quality metrics and use them to help educate the workforce about the realities of your data measurement methods.
I'm working with a large oil and gas client to help them improve their inspection practices. An interesting aspect of measuring pipe thickness is that sometimes a reading can be higher than an earlier reading, indicating that the pipe has actually grown in thickness. As you know, that's not physically possible, so this is often interpreted as bad data quality.
Chief Analytics Officer Spring 2017
15% off with code MP15
Big Data and Analytics for Healthcare Philadelphia
$200 off with code DATA200
10% off with code 7WDATASMX
Data Science Congress 2017
20% off with code 7wdata_DSC2017
20% off with code AIP17-7WDATA-20