People don’t know what they mean once they discuss data quality.
A number of years ago, our data platform team aimed to pinpoint the first concerns of our data users. We conducted a survey amongst individuals interacting with our data platform, and unsurprisingly, the primary concern highlighted was data quality.
The initial response, characteristic of our engineering mindset, was to develop data quality tooling. We introduced an internal tool named Contessa. Despite being somewhat cumbersome and necessitating significant manual configuration, Contessa facilitated checks for traditional dimensions of information quality, encompassing consistency, timeliness, validity, uniqueness, accuracy and completeness. After running the tool for a few months with a whole lot of information quality checks we concluded that:
- Data quality checks occasionally assisted data users in discovering, in a shorter timeframe, that the info was compromised and couldn’t be relied upon.
- Despite the frequent execution of information quality checks, there was no noticeable improvement within the subjective perception of information quality.
- For a good portion of issues, particularly those identified through automated data quality checks resembling consistency or validity, no corrective actions were ever taken.
Survey and objective measurement are useful tools, but nothing can replace a discussion over coffee and cake, as Jane Carruthers writes in her book, “The Chief Data Officer’s Playbook”. Indeed, I like to recommend this to anybody, as one-on-one conversations helped us discover one other essential angle of the situation. A few of these conversations unfolded as follows:
“Hey, you say, that data quality is poor, what do you mean by that?”
#1 Pricing business analyst: “We’re working on establishing price for the ancillary product X. Within the dataset we use, we’re missing data on what was the actual revenue from the product X per each order. Now we have this dataset , but it surely accommodates only expected value of the revenue from X at time of the acquisition. We will see also the actual revenue per product, but not on the order granularity.”