As businesses increasingly leverage data-driven decision-making, the ability to use and understand data at the company-wide level becomes mission-critical. While tech behemoths like Netflix, Airbnb and Spotify have strong data cultures built over the last decade, most companies often face challenges getting up and running with data. Data Culture co-founders and former WeWork data leads Gabi Steele and Leah Weiss, and Monte Carlo CEO and co-founder Barr Moses, share what it takes to make an impact with data at your company.
It’s 2 a.m. You’ve been staring at an Excel spreadsheet for five hours, trying in vain to understand how to take your raw CSV file and transform it into something that will actually tell you how many sales qualified leads your outbound marketing campaign drove last quarter.
At your last company, you had a team responsible for handling all of this, but at your new company, you’re running the show when it comes to data. Second to an energy drink, or better yet, a good night’s sleep, you wish above all else that you had all this data in one place you could actually work with, like a Looker dashboard.
Even though your data pipeline is murky as a swamp, what’s abundantly clear is that your company’s siloed approach to data just isn’t working.
Sound familiar? You’re not alone.
Here are three telltale signs your company should rethink its data strategy — and our suggestions for how to fix it:
1. Lack of clear data ownership
As the demand for real-time, actionable data grows, so does the cast of characters working with the data, including data analysts, data scientists, data engineers, and even data governance representatives. Although there is frequent overlap between these roles, each profession requires different expertise. If you’re just getting started with your data initiative, these distinctions can be hard to keep track of.
We often find companies first realize they need data to answer business questions such as:
- How much revenue is that advertising initiative driving?
- How many customer support agents do we need to hire to meet demand?
- Which markets should we prioritize in Q3?
Frequently, companies’ first move is to hire a data analyst to define and model the data. Hiring a data analyst is an awesome necessary first step, but also insufficient if your goal is truly to become a fully data-driven organization. It’s important to keep in mind that the role you’re hiring for doesn’t necessarily include the skillset required to do all things data, e.g. model the data, maintain the extraction and loading workflow, and execute the transformations required to arrive at those insights.
2. Lack of data democratization
Very often, the ability to work with and understand data is reserved for a select few in the organization, and despite the rise of data analytics training and data science boot camps, it doesn’t make sense for every data user in the company to become expert data modelers.
Data democratization, a term coined by data legend Bernard Marr, refers to the ability of all data users in an organization to access and understand data. Without data democratization, there can be tension when employees don’t understand the work required to build and maintain data models or run queries, and escalate these complaints to their managers when the data team tells them how long a request will take.
One customer success analyst at a leading food delivery service told us that this “ad-hoc querying” was the bane of her existence. “It prevents me from doing strategic work, and my stakeholders don’t understand why I couldn’t pull these queries for them … yesterday ... but I’m responsible for the entire customer success organization!”
We find that these bottlenecks are often the result of both a communication breakdown between different teams involved with the data, as well as a lack of democratization of analytic skills that would enable all data users to understand and work with the data themselves.
3. Your data infrastructure doesn’t meet your company’s needs
One of the most common challenges companies face when investing in data is bridging the gap between data infrastructure and analytics. The modern data stack has enabled a new path forward when it comes to quickly setting up a strong technical infrastructure, but a lot of data infrastructure is inaccessible or inappropriate to a company's use case, which means that investments of money and time never translate into anything that positively affects the bottom line.
A few other places where data tooling can fall short:
- Overcomplication: Data solutions are not always intuitive to use, particularly when it comes to onboarding new users. This can introduce bottlenecks that prevent teams from leveraging the full potential of their stack.
- Manual processes: Some solutions, particularly in the data governance space (think: data quality, data catalogs, and metadata management) still require manual input, a timely and cost-intensive process. If data governance and compliance is a priority for your company, data teams need a new approach that will keep pace with the growth of their business and reduce manual toil.
- Involved onboarding and set up: As previously mentioned, most organizations need their data as soon as possible. Data teams can’t afford to spend weeks or even months onboarding or getting up to speed with a new solution.
Together, these three signs — a lack of clear data ownership, a lack of data democratization, and inappropriate tools and technologies — paint an alarming picture of a company that isn’t set up for success with data.
How to build a data-driven culture
The good news? It’s never too late to start treating data with the diligence it deserves and building a data-driven culture at your company. Here’s how:
Communicate value of data early and often — to everyone
The first question every company serious about data should ask is: How will investing in data help me solve our business problems?
It’s important to align with key decision-makers, stakeholders, and data users to understand what their needs are and how the data team can help. We’ve found that one of the easiest ways to communicate the value of data is to start with one or two strong use cases, whether that’s partnering with marketing to build more sophisticated revenue models or working with the operations team to design a set of dashboards that quickly surface relevant data about the growth of the business.
Once you have data (literally) to validate your claims, your data initiatives can grow into more sustainable programs with enthusiastic data champions from other parts of the business serving as early cheerleaders.
Self-serve, automated tooling
Earlier, we mentioned that manual solutions and processes can make it harder to set started with data. While we’re far from the days of months-long onboarding and physical databases (you know, server farms and data centers in your office building), there is still room for improvement. Fortunately, more and more solutions providers are realizing the benefit of self-serve, cloud-friendly, and AI-enabled solutions to make tasks like data modeling, data exploration, and data discovery easier than ever before.
Many organizations start with a mix of stop-gap data access solutions and data models. A comprehensive, automated data solution — a modern data stack — allows you to systematize the whole process, allowing you to build reusable, replicable data models, dashboards, and reports based on known needs of your business.
Great solutions will free up data engineers to work on projects that actually move the needle for their companies, and let data analysts explore the data for themselves. To achieve this, your data team should carefully design a data stack and conduct infrastructure analysis to assess what makes the most sense for your organization.
Data observability and reliability
It doesn’t matter how much time and energy you invest in your data infrastructure if you’re working with bad data, you don’t know where this data is coming from, or you can’t trust it’s up-to-date and accurate. An easy way to frame this problem is through the lens of software application reliability. For the past decade or so, software engineers have leveraged targeted solutions like New Relic and DataDog to ensure high application uptime (in other words, working, performant software) while keeping downtime (outages and laggy software) to a minimum.
By applying the same principles of software application observability and reliability to data, these issues can be identified, resolved and even prevented, giving data teams confidence in their data to deliver valuable insights. A great data observability approach will provide end-to-end visibility into your pipelines, monitoring and alerting for issues before it affects downstream consumers, giving data stakeholders confidence in their insights to deliver business value.
Scale your impact by building a community around data
Community building takes time, patience, and a lot of listening, making this part perhaps the most challenging — but rewarding! — element of building a data-driven culture.
While data “owners” might be members of the data team, nearly everyone at the company uses data in their day-to-day jobs. To get started, we suggest partnering with data consumers in other functional areas, soliciting feedback and input as you build out your data stack and define your data strategy.
This community building extends within the data team, too. If you create a culture where business analysts and data engineers represent opposing forces who don’t understand each other, it’s very hard to succeed. Data engineers should do their best to bridge gaps between data and the business, bring data consumers into their process, and build community around data.
After all, hiring one data analyst at your company won’t solve all your business problems (though it's a nice start!), but building a data-driven culture will.