Press release: Six data quality traits that put the "intelligent" into "business intelligence"
You need to vet your enterprise performance data against these six data-quality traits.
Whatever your business intelligence platform is you’ll need to vet your data against the 6 data-quality traits.
Getting enterprise performance data to work for you is like… well, like getting people to work for you.
Let’s say you’re hiring for a C-level role. Whoever lands the job will be in charge of pulling important levers.
What do you do before welcoming them to your boardroom? You make sure their story and application data quality checks out.
Data-quality standards should be as unforgiving as your HR standards
When applying HR due diligence on potential new hires, you ask:
- Does what they tell you matches reality? (Accuracy)
- Are important details missing? (Completeness)
- Do other sources back their story? (Reliability)
- Do their answers match your needs? (Relevance)
- Did they show up on time? (Timeliness)
- Have They performed consistently? (Integrity)
Do these things right, and your new hire should bring powerful decision-making to make a big business impact.
It’s no different with your enterprise performance data. Whether your business intelligence platform is Finworx or Retailworx (built on top of Power BI), you’ll need to vet your data against the six data-quality traits above.
If you’re way ahead of us and business-intelligence ready, with infallible enterprise performance data quality, answer these seven questions to ask before adopting a business intelligence tool.
1: Data quality is about accuracy
Data quality number one seems too obvious.
The importance of accurate data escapes nobody. If 2+2 doesn’t = 4, or your financial data doesn’t add up, then you’re already in all sorts of strife.
It’s worth cautioning that business-data ‘accuracy’ is different to data ‘integrity’—data can have integrity and still be inaccurate.
Let’s say that in your IT assets, a vendor’s name must appear in the master list of vendors also. If the rule isn’t pragmatically enforced, then you have a data integrity issue, even if the data is accurate.
Before any business intelligence tool, you deploy can have maximum utility, your data accuracy must be established. This needn’t be a messy or complicated task at all if you take the right approach to data profiling.
But the ‘accuracy’ data quality is worth reinforcing—not least because without it, the other 5 data-quality traits we’ll cover don’t stand a chance of becoming valuable.
2: Data quality is about completeness
Data completeness for business intelligence is about ensuring that the BI you intend to gather for optimising business performance is fuelled by a full set of data.
In other words, before you deploy a new business-intelligence tool, you’ll need to know you don’t have missing data for fulfilling your business intelligence goals.
Establishing data completeness before going ahead with launching a business intelligence tool strategy can be achieved through a preliminary data warehousing project—something we do for international financial institutions regularly.
If you’re unsure, speak to us about whether or not your data circumstances warrants data management, to achieve data completeness in runup to launching business intelligence, or switching to Finworx or Retailworx.
3: Data quality is about reliability
Establishing ironclad data reliability is imperative for establishing a business intelligence platform able to give you an accurate circumstantial picture painted with data sourced from different canvases in your IT infrastructure.
If your ERP shows you a set of purchase orders that another IT system is telling you were made on a different date, which system do you believe?
Pouring unreliable data into a new business intelligence environment will create more cost and new challenges than it’ll solve.
4: Data quality is about relevance
Business intelligence capable of widening the scope for strategic decisions breaks down when you have data gaps.
When performing data management in preparation to roll out your business intelligence platform, don’t assume excess data—having too much data is equally problematic.
Fixing the problem of data scarcity into data overabundance creates a business-intelligence environment of redundancy.
Do you need the data you’re collecting for your BI launch? How much is it costing to collect? Did you pre-define its utility and role within your business-intelligence ‘data-to-decisions’ journey?
If you can’t define and prove the role and relevance different datasets will play in your new BI environment, then don’t collect them.
5: Data quality is about timelessness
Data timeliness follows the old adage that ‘time is money’.
The timeliness of enterprise data flowing to your BI view refers to how up to date the data is. Untimely business data isn’t just an inconvenience, it can be expensive.
Imagine that stock order updates in your ERP for countless stores across countless locations are delayed by 24 hours. This would cause no end of uncertainty and grief for teams situated at each location.
You might also end up with panic stocks being ordered by different departments, unaware that stock orders had already been taken care of centrally the day before.
Part of the role of business intelligence is being able to obtain information near real-time, so that old information can quickly be rendered useless with certainty.
In day-to-day terms, untimely information can be incalculably damaging to organisations in terms of time, money, and reputational damage.
6: Data quality is about integrity
Going back to the ‘hiring’ scenario in the introduction, when you hire new talent, you need to trace their work history integrity across other organisations they claim to have worked for.
Equally, data integrity is about ensuring your business data is traceable across other elements of your organisation. Establishing data integrity throughout its lifecycle, in each and every department and workflow, will determine its long-term utility.
Note that good data integrity doesn’t confirm good data accuracy. As we’ve mentioned, data can be accurate with low integrity and vice versa, though error checking is one of the steps to establishing data integrity.
Preserving data integrity means implementing processes for data backup, accessibility and permissions, input validation and other routine checks that should be in place uniformly across your entire business.
Strong data integrity means all of that has to go right all of the time. Get things wrong just once, and data integrity can break down.
Need an Informed Data-Quality Health Check?
Enterpriseworx partners with global financial and retail organisations to continually optimise their enterprise performance data utility, long-term.
Whether you’re just considering data management and business intelligence, or whether you’re rethinking your approach.
A little insight now can help you save a lot of time and expense later, when circumstances force you to run expensive emergency data quality projects.
Speak to EWX’s team about where you are on your data-to-decisions path, and they’ll help you better understand what fork in the road you’ll need to take next.