Creating a strong and deep-rooted data culture is a requirement to leveraging data for a long-term strategic advantage.
- Building a Strong Data Culture Overview
- What is a strong data culture?
- How to build a strong data culture
In today’s world data is king. Capturing vast amounts of data is no longer a challenge; mobile apps are ubiquitous which makes click stream data abundant, social media providers serve APIs to users' profile history enabling deep individual understanding and IOT devices are on everything from coffee pots to tractor trailers delivering massive streams of events, just to name a few examples.
Companies are now focusing on shifting their efforts for how they use the data they capture to drive more value and the most fundamental component to extracting value from data is to have a strong data culture. The benefits of having a strong data culture create value for your company by driving better utilization of your data to enhance or create products and allowing your team to spend more time being productive. I’ve worked for or consulted with about two dozen companies with enough hands-on experience to have a good understanding of their strengths, weaknesses and opportunities; in this article I will share my perspective for how to build a strong data culture based off of the culmination of my own experiences.
There are four main pillars to having a strong data culture:
- Alignment across your organization that data culture is a priority
- A strong community to reinforce the data culture
- Clear and well-managed processes
- Enabled associates at all levels and roles
Let’s take a look at the attributes of companies with a strong data culture as it relates to each of these pillars to understand what a strong data culture looks like.
A strong data culture must start with everyone on the team believing that there is value in putting in the effort to curate a data culture, capturing metadata, keeping documentation for pipeline changes up to date, investing to make data access easier and much more. From top to bottom, every individual on your team needs to be able to see the value that is being created by having a strong data culture and how that value is generated by the company as a whole as well as from their specific role.
Companies that treat data culture as a high priority have the following attributes:
- Their tech teams plan to spend time to update documentation and metadata before a feature release
- Executives reinforce the value of knowledge sharing, documentation, self-service access, training as well as data tools and technologies
- Managers encourage team members to spend time to participate in the data community
- Metrics that measure peripheral aspects of the data culture, such as what percentage of documentation is missing in a data catalog, are published and made available for the entire company to see
- Data stewards are seen as an integral component used for proper data management and not a roadblock for tech teams to move forward and their career tracks are just as distinguished and appreciated as data scientists or developers
The data community is much larger than just data analysts, data scientists and data engineers; it is often your whole company. It is the group of individuals who interact with your data, provide access to data sources, and create data pipelines as well as those who generally enable others to effectively use your data. The community also includes operations managers that review dashboards, front end developers creating click stream events and product managers who want to understand the efficacy of a new feature.
The data community is critical to developing the circle of trust with your data, as seen below. The value you get from the community will grow over time as these factors continue to reinforce themself.
An organization that has a strong data culture is more likely to have a community that:
- Has participants apply social pressure and hold each other accountable to maintain standards
- Creates Communities of Practice (CoP) organically which provide the larger community with new ideas, standards and procedures
- Provides opportunities for mentorships and training programs to support individuals looking to grow their skill sets
- Takes pride to ensure metadata and documentation are kept up to date
- Proactively collaborates among roles and teams, their users know about changes before they occur
- Encourages new tools, libraries, methodologies and techniques to be used when creating or using data
- Desires to share data with one another, not to control it
A company’s processes include both the formal and informal ways in which their teams interact. A more formal process may be defined for how new users request access to your analytical data whereas an informal process may be used by users to find out who the most recent developer was for a data pipeline. Clear and well-managed processes are highly correlated with the quality of a data culture, the hallmarks of which include:
- Simple and straightforward processes
- No single point of failure or bottleneck in any process
- Support and guidance for those who are unfamiliar with a given process
- Stewards that continuously look to improve the process without making it over burdensome
In order for your data culture to flourish the individuals in your team need to be able to do their job effectively. While I don’t think it is possible to create a strong data culture simply by focusing on enablement by itself, it most certainly is possible to prevent a data culture from flourishing if you do not invest in enablement.
From my experience, companies that have better data cultures also:
- Have individual contributors at all levels proposing and driving changes
- Invest in the tools that the team need to succeed
- Provide frequent and value-added training for everyone who wants to grow
- Engage with self-service tooling for common and repetitive questions and engage with colleagues in abstract and value-add conversations
Now that we’ve looked at what characteristics come through when a company has a strong data culture let’s break down the steps for how you can build this at your company. While you will face different challenges if you are starting a company from scratch versus trying to shift the culture of a Fortune 500 company we will look at some more common ways that are applicable for organizations of any size to curate the data culture that fits your company. There is no perfect script that will work for everyone, it’s important that you interpret these as it relates to your current environment and to think through how you can apply some or all of these items.
How to make data culture a priority
How you prioritize data culture within your company will directly impact how much time team members spend to curate the culture; therefore, it is generally easier to prioritize the efforts to make a strong data culture from the top down. When you’re in a leadership position, whether that is an executive or a team lead, it is imperative that you don’t just tell those that work under you to “build a data culture” or to “spend time to build a culture”; that is only one minor piece of the puzzle. There are a broad range of actions that, together, start to describe the behaviour that a leader should exude when driving organizational culture changes:
- Define a clear vision that describes where you would like to see the culture evolving and share that with the team
- Promote giving others the time that they require to execute the tasks that build up to culture
- Use positive reinforcement as your number one tool used to show that you are prioritizing the quality of your data culture. Spend time with individuals at different levels in the organization, convey to them in specific terms how you see their work providing value and listen to them to understand how you can help them and how they can help you. Too many leaders fail to adequately hold skip-level conversations with associates that report 2+ levels below themselves, as a leader these associates will gravitate to your vision with a passion if you take the time to talk to them.
- Be specific with feedback and appreciation, don’t ever be vague. If building a proper data culture is important enough to prioritize for the company then you should be prioritizing the time to understand and appreciate the value that everyone is creating. As a leader you can prioritize giving the space for culture to grow but you cannot create it yourself; by giving explicit feedback to your team that they can, and will, relate to it which will help them buy into your vision.
- Identify areas to sacrifice - you will inevitably hear others tell you that there isn’t time or that they already have too much to do. If you are not willing to hire more resources be prepared for these conversations with specific areas where you want to reduce effort in order to ensure that more time can be spent on building a data culture.
- Be patient with your team and don’t expect that just because they’ve been told to be better at documenting their data that the quality and depth of the documentation will change over night. The end goal is a moving target, reward incremental progress.
- Define metrics for which you can measure progress. This doesn’t need to be overly complicated and can be as simple as having a quarterly questionnaire. If data culture is important, it only makes sense to use data to understand how you’re progressing!
How to build a strong community to support your data culture
Your data community is the heart and soul of your data culture. Just as with our photosynthetic friends, helping the community thrive comes down to providing the right environment. You cannot force how fast it happens but you can adjust the environmental factors that impact how it grows.
Being in a leadership position naturally lends itself to having more influence in how a community can be cultivated and in determining what mix of factors should be used to grow the community. Here are some ways that you can encourage community growth within your organization:
- Hire for the right qualities - I recognize that you may not even have a budget to bring on new resources, but all companies need to hire to survive, at the very least to augment those who leave via attrition. When you do hire, make sure you are looking for qualities that relate to collaboration, detail orientedness and empathy. Find the people who are going to spend time to not just improve themselves but who will want to improve others as well. Hiring someone for their experience in another company that has a strong data culture can be a great approach to kick start your own culture, but don’t make experience an overweight criteria; overall culture fit with the culture your company is building is much more important than previous experience.
- Give visibility and responsibility to the communities of practice (CoP) by being a guest speaker, helping the CoPs organize guest speakers from other leadership positions or industry leaders outside of the organization. When the CoPs make recommendations - for example, they may recommend that data stewards should be hired - take them seriously and follow up the CoP organizers for how to implement their proposals.
- Be open and decisive when making organizational decisions from issues that arise when behaviors conflict with the community spirit that the company is trying to achieve. One common example is when individuals who control access are over protective in their willingness to share it, when you hear that they restrict productivity to a detrimental degree clarify the expectations of the role with them and the culture aligned with your vision is more beneficial for the organization.
- Structure your teams and reporting hierarchy in a way that encourages collaboration across roles and teams. There is no single solution here for every company but in general don’t create silos, keep your data creators and data users close together, if not on the same team. Try to create team-level performance metrics that different roles should cooperate on to solve in addition to tactical objectives for individuals that will encourage them to act in a way that is aligned with your desired data culture.
- Encourage spontaneous and frequent feedback from everyone in the organization and reward minor success stories with publicity and tokens of appreciation. One of the cheapest, quickest and most effective ways to do this that I’ve seen time and time again is the use of spot awards; when associates can nominate each other and give the reward to each other they want to try harder for each other and they reinforce the cycle with positive behaviours.
Creating processes that enhance your data culture
Using process to build a strong data culture may sound counterintuitive. Data cultures generally want to enable openness, sharing, simplicity, and maximizing data use but processes can often be thought of as the antithesis to all of these. Processes don’t have to be a bureaucracy laden tool that slows down your organization and they can be used to further enable your team to be more productive if done correctly. As a company’s size grows their processes will grow as well, both in terms of the number of processes as well as their complexity. These are ways to implement and manage processes in a way that will support your data culture:
- Create processes with the expectation that they will change. No process will ever be perfect when it is created and, even if it is, the external environment around the process will inevitably change which will make a process antiquated. Make sure your process owners are reviewing the efficacy of the process often and also that the owners of a process rotate as well to bring new ideas and perspectives into the mix.
- Clearly define processes as the goal that they are trying to achieve and the value that they create, do not create processes as “this person’s” process or “that team’s” process. One of the most common attributes that I find teams with poor data culture have is that they over-anchor on the processes that are in place and the individuals who manage the processes do not see the value and reasoning behind what the process is trying to accomplish as much as they see it as an opportunity to flex power and control.
- Take the burden of complexity off of the individuals or teams that are going through a process. For example, whether a user is requesting access to a database or a team is trying to integrate with a new external vendor and will need to send the vendor sensitive data, do not expect that the user or the team will know all of the steps to complete the process. If the process cannot be easily described in 1-2 sentences then ensure there is a sherpa to help users navigate the process.
- Leverage data stewards to solicit feedback from your data creators and data users. Data stewards should be highly engaged with your entire data community which puts them in an ideal spot to proactively collect feedback on how to improve processes.
- Find the right balance between single approvals and approvals by committee. Putting a single person in charge of a specific action, whether approving data access requests or reviewing data security protocols, can make processes fast and efficient but be wary about fiefdoms where individuals express too much control or troughs in productivity when the single person is not available. Conversely, committees can be bureaucratic and lead to everyone wanting to voice their opinion and they can drown productivity with paper work. I find that for most use-cases a group of 2-3 highly enabled individuals who are collaborative by nature and have the ability to make decisions on their own tends to be the most effective way to manage processes.
How to enable users who will create your data culture
Enablement really just comes down to giving your team the resources that they need in order to to succeed. Since the four pillars to a strong data culture are so intertwined with each other hopefully you already have an idea of what steps can be taken in your journey to enable users. Here are additional ways in which you can build on some of the aforementioned concepts to help to create enablement within your company.
- Ensure that every single person in your organization has the opportunity to impact the data culture. Provide the mechanisms and pathways for individuals to surface new ideas and have the opportunity to have an impact from any position within the organization. When someone has an idea and they are unable to share it because someone will not listen, or worse if someone listens but does not care, then you will likely lose that ideator’s passion from your community. You never want this to happen because the people who proactively bring ideas for improvement are the most valuable in curating a data culture.
- Treat data as an asset, not as an expense. When investing in your data consider the full lifecycle of your data from how it will be used operationally, who will need to access it analytically to what governance and monitoring will be required. Invest appropriately in the full lifecycle of the data, not just the initial capture and storage.
- Provide the tools that your organization needs in order to execute efficiently. At their most basic level, all of the tools that you use (e.g. databases, data catalogs, chat apps, etc.) are intended to get more value out of your data and users. Listen to your team and understand what problems they are facing. Picking the right tool for your team may take several iterations but try to test and learn quickly so that you can find the capabilities in the tools that your team values the most.
- Provide training to your team to keep them up to date on current industry best practices and trends. This can be done through internal programs, hiring external consultants and sending your team to conferences to name a few. In addition to the knowledge that your team will gain it will reinforce to your team that data culture is a priority and that you are invested in them.
We’ve laid out some ways in which you can start to build a strong data culture but it’s important to remember that you need to drive the culture that is right for your team. A strong data culture at one of the largest credit card card companies may have strict processes and controls for data access which reinforces how the culture treats data security. On the flip side, if you’re a small retail business that same amount of process probably won’t work for you if you’re looking to quickly integrate your data ecosystem with a dozen suppliers.
Continue to be consistent and patient with your team as you put your data culture into practice. It could take anywhere from a few months to a few years to shift your organization to the data culture that you want but make sure to check in on the progress that is being made often to ensure your company is moving in the right direction.