
Back when I started my career as a developer in the early ’90s, BEFORE the internet (yes that long ago), there was no such thing as data visualisation like we have it today. Instead, we had ‘queries’ that were basically information listed on black and white DOS screens or printed on noisy dot-matrix printers that took forever to complete. It was only several years later when I was introduced to and discovered the power of "pivot tables", that my interest in data visualisation showed its first spark. The rest is history; as you know, we now have the comfort of seeing our data in any shape or form we care to imagine. As long as we have amongst others, our data models, definitions and integrations in check, nothing can stop us from spotting the trend we were looking for or the feature/element that will boost our sales.
If you did your math right, you now know that I’ve been IT Business for the past 30 years, and have had the pleasure of seeing the visualisation of information evolve to where it is today. I’m delighted that more and more companies are gearing towards setting up self-service data analytics. with business intelligence tools like PowerBI, Tableau and Metabase. This new age of self-service, however, leads me to highlight that it comes with responsibility; hence what moved me to write this article.
Often enough, the data stack and models, integrations etc. are not always completely up to scratch. This can be from several reasons;
- companies are practically and often drowning in data both self-generated as well as from 3rd party apps and platforms.
- the data team lacks the discipline and management to update the documentation
- (this is very common) we don’t give enough priority to data governance and documentation
- good data engineers and scientists are hard to find
- there is no system request pipeline set up to give the Data/Business Intelligence teams the time they so desperately need to do all the necessary documentation and checks.
- Data/Business Intelligence teams are often understaffed
- and last but not least – the amount of "I need this report ASAP drop everything else you’re doing" kind of requests become the norm instead of the exception.
Time, discipline and prioritisation are paramount to ensure that they have documented, tested and validated the tables which they have made available to the team. Regularly maintained Data definitions, tracking plans and validation checks are key to ensuring that reports give consistent results irrespective of who creates them.
If there are no set policies and governance around your data, you end up with people using old inconsistent models, and with them creating their own dashboards -a disaster waiting to happen. This manifests into people, subsequently quoting different numbers to the ‘official’ ones.
Then it happens, all of a sudden everyone becomes the data ‘expert’, and more often than not, your data team ends up spending more time defending their dashboards, and numbers than anything else.
When you know you have a problem.
There are several ‘red lights’ to look out for, I’ve listed a few to identify some telltale signs.
You know you have a problem when:
- you have 100’s of dashboards and reports old and new, from different data sources that are not necessarily verified and validated
- (my fave) the CEO prefers ‘his own’ version of the revenue report, that he proudly put together himself and ‘trusts’ only that report simply because…. he did it himself 🤓
- Joe Bloggs in sales is quoting his own sales numbers that are not the official ones
- there are different definitions of key data elements within your organisation, and their respective values on your dashboard do not tally with the self-made ones.
There is hope!
What can you do to ensure data quality and consistency?
- Give your data team the time and space to document their work.
- Set up a "data pipeline" where data requests can be prioritised and managed.
- Disable as many old reports and dashboards as possible, and when someone screams re-enable their report only when you’ve had it validated.
- Clean up your tables and create tested and validated master tables.
- Setup and maintain your data definitions and make sure everyone has access to them and that they are simple and straight forward enough for anyone to understand.
- Establish "the main dashboard", that everyone can relate to and understands that is the focus of everyone’s activities, test it, validate it and start from there. All other dashboards should lead to and correspond to the main one.
It takes one to know one.
Slight confession, as I’ve managed data and information teams for years across 3 countries, and so I’ve not come this far without my share of mistakes. From those which I’ve pointed out some, I’ve been through personally and also admittedly at times have been the culprit of. So yes, it takes one to know one, which could potentially qualify me as an expert if mistakes are what makes one. While I’m grateful for all the opportunities I’ve had to work with diverse cultures, talents and skillsets, I know first hand that it’s not always as straightforward as we would like it to be.
An expert is a man who has made all the mistakes which can be made, in a narrow field. – Niels Bohr
In conclusion – manage your self-appointed data experts.
What I’d like you to take away from this is, your data/BI team (I admit that I may be biased here but I do believe it) is one of the most important parts of your organisation. They can make or break a business because your data and its quality depend on them, and to survive in this era, you need to make more data-driven decisions and less what Joe Blogg said decisions. To do that you need your information to be accurate, consistent and organised in a way that relates to your customer and their behaviour to your product/service.
To sum up, be wary of self-appointed data experts and their impact on your business. Ensure that your Data/Business Intelligence teams are the only data experts, back them up, empower them, prioritise the full spectrum of there work and watch your business grow.
Comments and feedback are welcome.
Drop me a line on [email protected], DM or comment
Originally published at https://www.linkedin.com.