2 Simple Steps to Boost the Value of Your Data

A friend of mine from business school recently joked that the aftermath of the U.S. election is more likely to give them an ulcer than the months of campaigning. Why? All the news of vote audits. Just the appearance of audit in headlines is causing her anxiety.

As a former management consultant, I know that no word causes executives as much annoyance and stress as audit. There’s a unique set of headaches that this kind of process brings to a business.

Yet audit shouldn’t be a dirty word. An audit is a chance to tackle critical business challenges head-on and reap the benefits of resulting improvements. This is especially true when it comes to data audits. The ROI is greater — and with the right tools, the process practically stress-free.

The journey to better data

Most companies think of data auditing as a chore: something you have to do every few months to comply with data regulations. The general perception is that data auditing is tedious, unnecessary, and time-consuming, with little upside beyond the necessity.

It can be. But a data audit has the potential to be so much more. At its core, a data audit is a process by which you review all your data, checking vital metrics to ensure that your data set is of high value.

High-value data = Better insights and higher ROI. This is something every company wants; they just don’t realize that regular data audits are the way to achieve it.

Step 1: Prioritize audits

The best way to capture that ROI? Prioritize data audits. Opportunity rather than responsibility. Think of them as one of your easiest opportunities to capture growth.

Step one: approach data audits as an exciting opportunity, not an unwanted responsibility.

You need to change your mindset and recognize the potential for growth held in the correction of your data. After all, according to Experian, the average company loses at least 12% of its potential revenue due to data inaccuracies. That’s not even considering lost potential caused by inefficient decisions made on less than optimal analyses.

Just imagine the lift that a data audit could bring!

Step 2: Automate your audit

So how do you achieve this lift efficiently?

Step two: set up continuous, automatic data auditing.

While a single audit done at scheduled intervals is a considerable undertaking, continuous data audits build the benefits into operations, making them much easier to deliver. Instead, you can automate the audit within the analyses you conduct and within the data collection processes. By making the audit a part of your model, you get the ROI you need without the stress of a big production every few months. An even bigger bonus? You find errors sooner, mitigating their impact on the bottom line.

The ABC of data auditing

What are you auditing in this continuous, automated process?

3 areas you want to check during a data audit: the ABC of data auditing.

1. Accuracy: Is data correct? This ensures that the basic information used to make analyses are a firm foundation. If data isn’t accurate, it’s just noise.

2. Breadth: Is the scope of data appropriate? Does it correctly represent the issue analyzed? This ensures you have no gaps hindering your analysis.

3. Consistency: Is your data formatted correctly, without gaps or deviations from the correct methodologies? This ensures you don’t lose vital information due to anomalies in data collection or processing.

Each of these is equally important if you want better, more actionable data. Build checks on each into the entire data process, and you guarantee higher ROI.

Bigger isn’t always better

If auditing data for these ABCs can have such a significant impact on your company, it’s easy to assume that the big audits are where you have the largest breakthroughs. That’s where you find the most prominent errors, right? The opposite is actually true. Constant, small but automated audits make a much more significant difference — because they prevent those problems in the data in the first place.

Take a look at this simple script to audit scraped data:

var fs = require('fs');
 var casper = require('casper').create();
 casper.start('https://www.avis.co.uk/');
 
 casper.then(function() {
     this.sendKeys('#hire-search', "Aberdeen International Airport").wait(500);
 
     this.waitUntilVisible('a.is-first-result').thenClick('a.is-first-result');
     this.thenClick('a.select-location');
 
     this.thenClick('#car-search-submission1');
 
 
     this.waitUntilVisible('.vehicle-matrix', 
         function() {
             this.capture('results.jpg');
 
             // This first callback is triggered if the selector is visible before the 10000ms defined as the 3rd argument
             var firstCar = this.evaluate(function() {
                 var vehicle = document.querySelector('article.vehicle');
 
                 var name = vehicle.querySelector('p.vehicle__note').innerText;
                 var features = vehicle.querySelector('ul.vehicle__features').innerText;
 
                 return {
                     name: name,
                     features: features
                 }
             });
 
             this.echo('Car name: ' + firstCar.name + "\n");
             this.echo('Car features: ' + firstCar.features + "\n");
 
             fs.write('results.csv', "Car name;features\n" + firstCar.name + ';' + firstCar.features + "\n", 'a');
             
 
         }, 
         function() {
             // If Casper can’t find the selector in those 10000ms, we assume it’s not available or an error occurred
             this.echo('No results found');
         }, 
         10000
     );
 });
 
 
 casper.run();

It merely creates images that allow the program to monitor whether the information pulled from websites is accurate and consistent. This loop builds auditing into the scraping process to ensure you collect high-quality data from the get-go. It’s a simple code that achieves a tiny audit in the earliest stages of data processing — with a major positive impact on data quality. Sometimes the smallest steps matter most in Big Data.

Optimal value

The costs of bad data are staggering. They undermine attempts to improve KPIs and leave companies struggling. After all, inefficient use of data caused a $200 billion excess of inventory in the United States. The only way to solve this problem is to use data more efficiently. Data auditing alone won’t achieve this goal, but it helps ensure quality data you can depend on for better decision-making.

No matter how you approach any data audit, you must make it a priority at every stage. Automation makes this possible — so you can further automate every other data-driven decision that matters.

2 steps get you there:

1. Proper prioritization of data auditing.
2. Continuous automation of data auditing.

It’s time to do data auditing better. Learn more about the technical side of data audits and how to become a true “Data Ninja” in the scraping course at Evo University. It’s free, so why not enrol today?