I work at a place where data quality is not on anyone’s radar. We have a reporting team in our group so we do our best where we can, but combining any datasets with other groups (like marketing & sales) is next to impossible as each team is silo’d and do things their own way - think free-form text fields to tag content…
How can I politely and succinctly say the above? Also, anyone else in a similar boat?
Implementation of AI requires strong unified data governance and data hygiene to produce company wide strategic solutions. The current company posture instead is focused on tactical level data collection and analysis which does not lend itself to consumption in relevant possible cross-department opportunities.
And if you want to tank it without overtly tanking it.
“We will need to establish a review and governance board to establish standard data structures and reporting that can be used to drive the initiative.
It will need to be cross team and cross specialty so we should start by establishing a group to identify those people so we can proceed”
A year later and you’ll be lucky if they’ve even picked out who can be part of the review process let alone agree on some convention and adjusting their tooling and processes to make that work.
They’re not wanting to work better, they just want to be able to say they “use AI”.
Get them an estimate of how much it would be to give everyone Copilot for Business licenses and they will start talking about the business having other priorities.
We’re unable to leverage some of the latest advances in AI due to our Leadership’s abundance of caution and strategically allowable risk profile.
“Each department in this company has developed their own way of doing things. Unfortunately, not every department is giving us information in a way that we can use it to our advantage. We need to get everyone across the company on the same standards so we don’t have any mixups. We can’t help you do your jobs better if we can’t process the information you give us.”
We need shared definitions to tell meaningful stories with our data. And then use a company specific example like how a customer’s journey can not be understood with differing definition between marketing and sales. The marketing team can’t measure the quality of the leads they’re producing unless they can directly link a customer’s whole journey from acquisition to churn. Otherwise it’s just vanity metrics. But don’t be too harsh, vanity metrics are really common in business. A company needs strong data leadership to create a culture of using data to justify decisions to a culture of using data to inform decisions.
Definitely try to use examples to help them get a glimpse into the issue. I like to explain documentation errors by pointing out when what are supposed to be sequentially recorded timestamps are recorded out of order in my work’s database. Sometimes the data quality isn’t there.
