There’s a commonsense rule in business which says that quality of input will always determine quality of output. It’s worth remembering that the exact same principle applies to your data initiatives, too.
Accurate reports and reliable insights do not appear by magic. And of course, even the most powerful data analytics platform can never be a crystal ball. Feed it with incomplete, inconsistent, duplicate or corrupt data, and the results are guaranteed to disappoint.
Forrester found that one-third of business analysts spend more than 40% of their time vetting analytics data before it can be used for decision making. It highlights another big problem with quality deficits: you can spend more time dealing with them than with real business challenges.
So how do you know if there’s a data quality problem within your organisation? Here are the signs to look out for…
You are stuck in the data processing mud
You embarked on your data project expecting certain things to happen. You’d have answers to key business questions at your fingertips. Reporting would no longer be a monthly ordeal, and you’d finally be able to focus on value-added strategy.
Poor data quality can cause the reality to look very different. It can easily lead you to be mired in what’s been dubbed ‘data janitor’ work, where data scientists can spend between 50 and 80% of their time collating and preparing unruly data.
The people grappling with these tasks are meant to be data heroes rather than data processors. You’ve invested heavily in them, and yet their potential is going untapped while they attempt to render your data fit for purpose.
Same question – different answers
What is our projected overhead spend for this month? How many new leads are in the pipeline? When you’re in a meeting of department heads, you can discover how even seemingly black-and-white questions can produce a range of weird and wonderful answers.
Data items haven’t been coordinated. Key fields have been left blank. There are multiple versions of reports scattered across the business, resulting in inconsistent answers. Instead of coming up with solutions to pressing business problems, time in meetings is spent debating who has the right figures.
All of this quickly breeds lack of trust in the numbers: a sure-fire way of sapping confidence in your data initiatives, causing stakeholders to revert to decision making on a hunch.
You’re not getting timely insights
According to one survey, when employees need data to make a decision, just 3% have instant access to it. For the majority (60%), it takes hours or days.
From supply chain management through to retail, the landscape can shift suddenly. You need to decide and act fast. If data is incomplete, or if there’s a failure to convert it into a usable format, you lose the ability to step in to prevent minor issues escalating into major business problems.
When and how to fix your quality issues
Data does not land automatically at your feet in a complete, usable format. This is simply a fact of life. So how do you respond to it?
A lot depends on when quality issues are addressed. Allow flawed data to travel downstream, and it’s a recipe for the type of laborious validation exercises, gap-filling and line-by-line adjustments that can leave you on a constant treadmill.
Your mission should be to address it at source; to fix the process – not the faulty output. The right architecture, capture, storage, extraction and integration approach is crucial. For expert, targeted help in addressing precisely the data quality issues your organisation is facing, speak to Triangle today or view our data quality services.
