Data governance for startups and SMBs: what it is and why you should care

Historically, data governance was something only large corporations cared about.

But as most startups and SMBs today are hoping to grow their businesses with data-informed decisions, the practice of data governance is becoming more important.

And that’s why in this article, we’ll walk you through everything you need to know about data governance for small and growing businesses. 

But before we get to the good stuff, let’s quickly define data governance to make sure we’re all on the same page.

What is data governance, really? 

Simply put, data governance is the process of managing the availability, usability, integrity, and security of all company data. 

Ultimately, the goal of data governance is to make sure that the right people within the organization have access to clean and consistent data that they can use to make operational decisions. On the other hand, proper data governance also safeguards the data from unauthorized access and controls the level of access for different people within the organization.

With this definition in mind, it’s safe to suggest that many startups and SMBs should start paying more attention to data governance.

Should you worry about data governance yet?

Speaking of paying more attention to data governance, how can you tell if you should already start doing something about it?

If you ask us, there are two tell-tale signs that suggest you should establish some data governance in your organization:

1. Think of your company as a bumper car course. As you start adding more and more cars (= people) to the course, you’ll quickly notice that it would be nice to have someone to direct the traffic. In other words, the larger your organization gets and the more people use data in their daily decision-making, the more important it is that these people have a unified way of using the data.
2. Look at the type of data you’re looking to consolidate. Since we’re specialized in marketing data here at Supermetrics, we’ve noticed that whenever a company starts joining their marketing data with internal business data, it’s time to get serious about data governance. For example, companies with large online ad spend would typically want to understand the revenue implications of their activities, and the only way to get a full understanding of that may be to combine revenue, product usage, and marketing data.

A historical perspective into data governance

To understand why the issue of data governance is now more important than ever, let’s take a quick look at how we got to where we are today.

The 70’s on-prem data debacle 

The concept of data analytics dates back to the pre-internet era. Back in the 70’s, all company data was stored in on-premise databases. 

At the time, software was expensive to build — but the hardware needed to run that software was even more expensive. In other words, the in-house cost of developing software was cheaper than the cost of data storage. 

Unsurprisingly, the relatively high data storage costs led to a situation where most companies had to be extremely stringent about the shape and size of data they would store in their on-prem systems. 

And as the monolithic nature of these on-prem systems didn’t exactly give data engineers the opportunity to flexibly control the data inputs whenever they wanted, they developed a strict waterfall-style process for modeling, cleaning, and transforming the data that would end up in the database.

This process ensured that the data that finally ended up in front of business decision-makers was clean and consistent, satisfying the definition of data governance.

The cloud era of the 2010’s

The 1970’s model of data governance was all well and good until the end of the 20th century. But fast forward to 2010, and suddenly every company in the world is moving their data to the cloud — and for good reason.

With increasing data volumes, storing data in the cloud was (and still is) a lot cheaper than storing it in an on-prem system. What’s more, cloud-based data storage democratized access to data, making sure that anyone within the company could access the data they needed, when they needed it.

All of a sudden, anyone could pull data into their spreadsheets, BI tools, or data visualization tools with something like Supermetrics. Gone were the days of having to wait for someone in the data team to deliver formal reports and turn them into actionable insights.

Dismantling data anarchy in the 2020’s

While we can probably all agree that the democratization of data was ultimately a great thing, it also spawned a host of issues. 

Most notably, self-service BI was great to an extent… Until it wasn’t. When everyone and their grandma had the tools and access to build their own reports, you almost invariably found yourself in a situation where one department’s numbers no longer matched those of a different department. 

In fact, just last week, we noticed that the MQLs the Supermetrics marketing team is reporting don’t match the MQLs our sales team is reporting. But I digress.

You would think that with access to the same data, people would arrive at the same conclusions. But the problem is, you’d be wrong.

This so-called data anarchy is one of the reasons why we’ve recently seen many of our more advanced customers move to a centralized model of data governance and delivery. Instead of having 15 marketers or analysts pull their own data into Google Sheets with Supermetrics and drawing their own conclusions, they’re moving to a system with centralized governance. 

By imposing a unified data governance model, companies ensure that people are analyzing the data correctly, comparing apples to apples, and arriving at sound conclusions.

And if you’re interested in gradually moving away from data anarchy, here’s what we recommend you do.

Data governance best practices for startups and small businesses

Data governance is one of those pesky things that’ll never truly be done. That’s why you’ll want to think of it as an iterative process that consists of three key pieces: people, tools and processes, and continuous education.

1. People

The first rule of data governance is: if everyone owns it, no one owns it.

This means that you’ll want to make data governance someone’s responsibility as early on as possible. A good rule of thumb is that as soon as you have a full-time analyst in your team, they should explicitly own data governance.

Once this person has been assigned, it’s up to the leadership team to communicate what this change means to the rest of the team. After all, owning data governance is one thing — and getting people on board is another.

2. Tools and processes

Once you have an owner for data governance in your organization, it’s time for them to roll up their sleeves and get to work.

A good place to start is building a “data treasure map”, which is a document that answers the following questions:

• Where does the data come from (e.g. a list of all your marketing, finance, and sales platforms, CRM, CDP…)
• Where does the data go to (e.g. your data warehouse and/or data lake, and the BI/visualization tools where you have your dashboards)
• Who uses what data

At first, this treasure map can be a simplified drawing of your data sources and destinations. 

Later on, you’ll want to turn this into a full-blown Master Data Model (MDM), which is a catalog of:

• All the different datasets you have
• The tables you’re using and their detailed descriptions
• Access levels and the people assigned to them
• Partitioning strategy
• Naming conventions

If you’re hiring new people to the team or shopping for a data pipeline, having a data treasure map and/or an MDM can be hugely valuable. Think of it as sharing your medical history with a doctor instead of only telling them that your head hurts. In other words, a documented data model gives new employees and possible vendors context into what you’re doing and helps them come up with relevant questions and better recommendations.

Pro tip: For data governance to be successful as you scale, you’ll need to embed it within your existing processes. If people can avoid doing it, many will, so implementing mandatory minimum requirements for populating the MDM before adding data to your platform can add the necessary motivation. Getting this right from the start will almost always be more cost-effective than trying to fix it later. Look for any opportunities to make life easier for your people by automating as much of this as possible.

3. Continued education

Finally, for employees who are used to complete data anarchy, going back to a more centralized model of data governance can feel painful. That’s why it’s important to continually educate the team on the benefits of the new model as well as its dos and don’ts.

And remember, data governance doesn’t have to mean a fully centralized model where your non-data team members will have to beg access for days. It’s up to you to define what different people can or can’t do on their own.

Wrapping up

Data governance may sound like a big and scary issue for most startups and scaleups. But in the end, the goal is to make data more secure, accessible, consistent, and clean, which in turn will improve operational decision-making.

And that’s what we all want, right?

As with anything, the best way to get started with data governance is little by little. Start by assigning an owner, having them create a “data treasure map”, and educating your entire staff of the changes. After that, data governance becomes an iterative process, where the treasure map should be reviewed every three to six months.