The Unexpected Benefits of Democratising Data Access
June 13, 2019
Author: Manjit Johal
It’s pretty easy to understand the massive spike in analytics companies in the last few years: there’s been an explosion in generated data, and collection techniques have become much more sophisticated.
One example is the retail customer who participates in a company’s in-store loyalty program but also shops online and engages with the brand on social media. In isolation, those data sources are useful, but incomplete. Combined, they are much more powerful.
The more useful data you have, the higher the resolution in which you can understand your customers.
Understanding the customer
Greater clarity leads to all sorts of benefits: more targeted offers, product ranges which meet customer needs better, understanding how much customers are likely to spend, and plenty more. At the end of the day, these all contribute to higher sales.
Today, data access can also be tied together with machine learning to better anticipate customers’ actions.
For example, companies can learn more about how their customers approach purchases, or internally they can improve inventory models for more efficient operations.
The better you understand the customer, the better your business can compete with its rivals.
Leveraging data internally
But gathering data isn’t just about the customer – it also brings huge benefits for the employees working with that data.
Traditionally, the separate departments of a business rely on intuition built up over years of experience. The person making the decisions is the person who’s seen it all before. But now you can take accessible data and put it in the hands of employees. You can provide crucial insights and feedback into the decisions they’ve made and help them improve based on objective results, not just gut feelings.
Of course getting the ‘right’ data (i.e. the most relevant data) is crucial. However, the real challenge is reducing the time it takes from that data being generated to it being in your user’s hands.
Historically it could be weeks – even months – before data bundled its way into the right hands. By the time the user had that data, its value to the company was reduced. There’s a sort-of half-life with data: the longer it goes unused, the less powerfully it can influence business.
Let’s take an example.
Say an employee is running promotions for a variety of products. Some will inevitably perform better than others, but it will be weeks until you see those collated and analysed results.
Imagine instead that this information was fed to the employee in real-time: they could amend and tweak each promotion to maximise the impact of that promotion. They’d get higher-performing ads and a more lucrative return on investment.
Streamline the data process
But getting that data into employees’ hands in real-time can seem daunting. Some streaming mechanism would have to be deployed which might be alien to a number of organisations.
My initial advice would be to start slow. Try to refresh that data and make it available once a day, at first. Once that’s established, aim for four times per day. In time, head for every hour, or every minute, or as close to real-time as is feasible or necessary. The leap from monthly to on-demand data ingestion is enormous – unless you break it up into manageable chunks.
One way to make data real-time is to automate the ingestion process as much as possible. Stop relying on users manually creating and updating spreadsheets. Well-oiled automation won’t just be faster and more cost-effective, it will also be more trustworthy and will result in far fewer mistakes.
Trusting your data
It’s not uncommon for users to completely lose confidence in the accuracy of their data, something which substantially reduces the credibility of any system.
Increasing the velocity of data movement around your organisation might also surface some unexpected benefits. For one, it will highlight inefficiencies and flaws in your existing processes. A nightly extract that’s produced intermittently because of a flaky database, or a supplier that’s not producing extracts as expected, are two common examples of this.
One last benefit will be the democratisation of data; making it freely available to teams that aren’t responsible for producing it. This increases transparency between different departments which previously sought to protect and hide their data.