line graph, bar graph, two women looking at data on a computer, large group of people walking and wearing masks in a city, and two males a manager and frontline employee checking on products in a warehouse
May 21, 2020
Data Intelligence

Four Useful Tips for Better Data Standardization in 2020

Data standardization is the critical process of bringing data into a common format that allows for collaborative research, large-scale analytics, and sharing of sophisticated tools and methodologies.

There are a multitude of benefits that come with standardizing your data, including cost reduction, productivity increases, and better decision-making.

One of the most critical steps to becoming a more data-driven organization in 2020 is to implement better standardization practices. Standardizing data practices makes analyzing and interpreting data faster and more accurate, which ultimately produces more informed decision making. However, depending on your organization, it can be a large undertaking. 

Here are four steps for better data standardization in 2020:

1. Know What You're Working With by Auditing Your Data Sources

Start with the sources.

Standardizing your data means examining the flow of data—from supply points to your database, then how it is disseminated throughout the organization.

Work with internal stakeholders and decision-makers to map out data sources, and acknowledge any silos you encounter. Sometimes, owners of specific business units can be protective of their data, not wanting to share or surface it in ways that they don’t feel they can control as directly. These individuals will make the audit much easier, as they can speak to the collection, storage, and process in regards to that data. Mapping out data will give you an idea of information architecture and processes—where data lives, if/how it changes, and how it is, then, passed through the rest of the organization. If it helps, use visual whiteboard tools like MURAL or Miro to map your data structure. These tools allow for remote collaboration, so you can easily work with stakeholders and decision makers.

Once that data is mapped, take some time to evaluate quality. Keep things like accuracy, breadth, and consistency in mind. Most importantly, refer to your business objectives. Is the data you are collecting relevant to your organization’s current initiatives? What’s missing? Can anything be automated? Keep these in mind as you map and examine your data architecture.

What your data audit should contain:

  • Data sources (first or third party)
  • How often each source generates data
  • Teams that use each source
  • Changes in the data, if/when applicable

2. Set Standardization Criteria Based on What Matters Most

Figure out what matters.

Every enterprise has its own data needs, goals, and sources, so it’s tough to give blanket advice. Therefore, it's up to you to determine what matters most to the business so that you can standardize your data so that it makes sense according to business objectives and current initiatives.

Remember that the goal of standardization is getting the data to a place where it can drive and inform business decisions. A good rule of thumb is keep things consistent and consolidated; after all, the volume of data you are generating and collecting is just going to increase as you scale. It’s best to keep things as simple as possible so you can easily standardize incoming data. 

Refer back to your data audit and consider the sources: Can all data within existing sources be standardized according to the criteria? Consider incoming data: Can the criteria accommodate that, too? Work with internal stakeholders and decision makers to create a system of standardization that considers these points.

We’ve often seen business stakeholders take a “give me everything” approach to data, as opposed to focusing on the right data, or just what they need. Doing so can tax the system, and it also makes it much more difficult to extract the right insights from the data they’re looking at.

This feeling typically comes from a place of not wanting to risk “missing anything,” or not wanting to explain what they’re really trying to learn to their data partners (“it’s easier if I just get everything and then analyze it myself”). Whether it’s a lack of trust, lack of experience, or something in between, the net result is typically the same.

3. Standardize New Data, Then Work Backward

Start with new data first.

Oftentimes, the systems for younger data sets are immature, and thus, easier to manipulate. Refer to the standards and criteria you’ve created with others and apply them to this data set. This should set the standard for standardization. 

There are typically two, primary categories when it comes to data standardization:

(1) Source-to-Target Mapping, where both internal external data is onboarded, and its keys and values are mapped to an output schema, and (2) Complex Reconciliation, where you create calculated metrics that differ depending on logic.

Standardizing new data is only half the battle; you need to enforce your newfound standards on historical data and your database(s). This could be relational databases, unstructured data, XML, documents, voice, and media. These are often disparate - spread across different tools/sources including spreadsheets, software, and more, which can lead to inefficiencies, wasted time and general frustration.

Luckily, applying standards can be less painstaking with automation. Data standardization is a manual process that’s ripe for human error and wasted time. Small errors and inconsistencies in your data can cause larger problems as they flow through systems, creating a snowball effect that can lead to mismanagement or uninformed decisions. Automation streamlines data standardization, and, once configured to your liking, allows you to rely on the same process with similar, trustworthy results.

4. Make Standardization Part of the Narrative, Then Tell the Story

Get better at telling stories.

Data is only valuable to the extent it can drive informed decisions; the best data hygiene practices in the world can be rendered relatively ineffective if the business isn’t using said data to tell the right stories. All too often, two (or more) stakeholders look at the same set of data and draw completely different conclusions, which means each is likely to craft a completely different narrative. Thus, it’s critical that the business is aligned on how data is leveraged, which insights matter most, and what data may not  matter as much—internally and externally.

Once you’ve drawn the lines between your data, reasons for standardization, and your business strategy, package it into an easily consumable format, and share it with the right people internally. Display tangible examples of how standardized data has, and will, be able to drive and inform business decisions. Point to specific benefits that might include: cost savings, process efficiencies, risk mitigation, and others. 

Sharing your data standardization story and promoting its benefits is a great way to get a firm foundation on establishing a data culture/community within your organization.

Wrapping Up

Better data standardization is critical in 2020 and beyond; if done correctly, it will set you, your teams, and your organization up for success in the long run. The more data that exists, the more important it is to control that data. Proper standardization makes analyzing and interpreting data faster, which produces more actionable insights and, ideally, better and more informed decision making.

Data Studio
Google Cloud Build
App Engine
Cloud SQL
Cloud Storage
Cloud Functions
BigQuery ML

Improving Data Quality for Better Decision-Making

guide improving data quality for better decision making

View our actionable, four-part guide for enterprise leaders to learn how to improve data quality, reduce risks, and make better decisions. View the guide now.

View Guide ->
RevUnit icon
RevUnit is a technology studio that helps supply chain clients identify and implement data solutions that actually prove ROI. We help organizations across industries like transportation, freight, logistics, retail, and manufacturing achieve business results through innovative data solutions — powered by AI/ML and the cloud. We’ve done it for clients like ArcBest, J.B. Hunt, and Walmart.

Related Posts