A data-driven culture seems to all the rage today but I don’t think this should be the culture. Data is meant to support your company, not drive it. The difference might seem minor to you but it can be the difference between NBA championships and underperforming expectations.
Let’s start by looking at the myths of what it means to have a data-driven culture before diving into how to create a data-supported culture.
Myth #1: Data Should be the Main Resource
The first myth of a data-driven culture is that data should be the main resource that everyone should be working with. I love data (maybe more than the average person) but I also understand its’ limits. Focusing on making data the most valuable resource can lead to optimizing the wrong things.
Instead, I believe companies and individuals within those companies should focus on managing their energy. They might spend their energy on data analysis, designing campaigns/experiments, or other activities that are important to their role.
Data should support all of these activities and this is why I believe that companies should aim for data-supported cultures instead of data-driven. Maybe this is semantics but words and intent make a difference.
Survey after survey on how organizations view data all say same thing:
- Data is an important priority and they will increase our investment in it
- Every team or executive believes being data-driven is important
- There’s too much data and it keeps increasing
I think that companies are trying to put too much ownership on their data and their people. They are expecting everyone to be data proficient (that’s my second point) overnight which may or may not be natural to them. These same organizations might also be dismissing the things that made them great and trying to evolve into something they are not.
If you have read any of these blog posts or my Growth Needle emails then you know I love basketball. There’s a growing trend of basketball teams using data to build and manage their teams. There’s no better example to this than the Houston Rockets. They have adjusted their style of play based on what the data says. Take a look at the following quote:
“Morey’s [Rockets GM] most fundamental insight involved taking an increased number of three-point shots. Three-point shots in basketball are more difficult because they are further away from the basket, but Morey recognized that the 50% uplift in points received for the three-point shot (compared to a two-point shot) made it more mathematically efficient than almost all two-point shots other than dunks and lay-ups. In particular, Morey (and others) realized that three-point attempts from the corner of the court (“corner threes”) had a higher percentage chance of going in (because the shape of the three-point line made corner threes closer to the basket) and were, therefore, more valuable. Many set plays are now designed specifically to get strong three-point shooters open for corner threes”Source
Despite their focus on data, they have yet to win an NBA championship since 1995 and have consistently come up short in recent years. They have been criticized for “boring basketball” and underperforming expectations. This doesn’t mean that they are bad team, they consistently make the playoffs but championships are what matters in this world.
I think the Houston Rockets are too obsessed with analytics and have tried to solve all of their problems with data. This ignores the large variables of human talent that are actually playing the game and can come up with solutions that aren’t statistically possible.
Myth #2: Every Single Person Should be Proficient in Data
The second myth is that every person should be data-proficient. Data should be democratized and everyone will go through the data to get answers to their questions. This is not what happens in the real world.
Some people do not want to go through data and just want the answers. Others are excited to go through the data but are unable to find the answers that they need. A good data strategy will take into account all of these different possibilities and provide a solution.
I constantly come across data scientists who are extremely excited about the potential of their work. They will share this with someone who simply wants the results and doesn’t particularly care for the math and beauty of their model. That is fine, not everyone needs to know how the sausage is made.
Some companies make data proficiency a key requirement for joining their teams. This is also fine if that’s how you want to build your culture. However, you should be aware of what blind spots this kind of data-driven culture creates.
For example, we always hear stories as to how tech products aren’t built for women because the engineers who make them are typically male. There are similar blind spots that are created when your team relies solely on data to make decisions just like the Houston Rockets keep missing the intangible variables of basketball.
Myth #3: In God We Trust, All Others Must Bring Data
I love this quote but not for the same reasons that you might expect. This quote states that if you want to make a valid argument, you better bring data. Your opinion and anecdotes have no place here.
Does this mean that companies should dismiss the opinions of everyone? Even the CEO who has 40 years of experience in the industry? What about the intern who has a hunch of something that could work?
Data by its very definition is designed to show patterns. Even the predictive kind of data that is becoming more prevalent, is based on existing data and patterns. If you can’t quantify something then you won’t have data to back it up.
This means that data may be slightly behind when it comes to innovation. By the time you can see the patterns of what products you should build, you will already be too late. This doesn’t mean that all data should be discarded but as I said before, we need to understand its’ limits.
Let’s now look at how to establish a data-supported culture and what this looks like in the real world.
5 Steps for Establishing a Data Supported Culture (Not Data-Driven Culture)
Let’s look at how your company can start using data to inform your decisions without letting it take over your entire organization.
Step 1: People
We start with People because it doesn’t matter how sophisticated your tools or your processes are if your team can’t use them. I also find that this is the biggest variable at my clients because they all have slightly different makeups and this affects everything in their data strategy.
Teams that are highly technical can take advantage of more advanced tools while teams that are constrained by their technical capacity need to find solutions that are user friendly. When companies get this incorrectly, they end up with tools and processes that no one can use.
You also want to consider the following questions.
Who is going to own the data on the business side and the technical side?
Everyone can access the data but there should be a clear owner that people can reach out. Common questions will include clarifications on what is being tracked, what tools are available and best practices for analysis.
The technical side is also important because you will need support to maintain any data tracking and implementation. It’s also helpful to have a champion within the technical team who can advocate for the data. Without this support, you’ll find that your tracking is constantly breaking as you release new product changes or features.
Who do you need to hire?
If you’re missing any key roles, this is where you want to make note of that. You may need to hire data analysts, data engineers, data scientists, or some other combination of skills. You can think of your hiring in two broad categories: who’s going to implement/maintain the data and who’s going to analyze it.
Choosing the right tools can help in this area especially if you have people in your team who could analyze data if it’s provided to them in an approachable format. However, you should keep in mind that this kind of analysis is limited to what a full-time data analyst or data scientist could be doing.
Whose support do you need internally?
You should also consider who will support you in this plan. This would include technical resources such as engineers but it might also include legal counsel to help you sort through data privacy issues, other executives and leaders, and even your own team.
You should be thinking about who could benefit from this data and who could block this strategy from proceeding forward. At the end of this step, you should have an overview of the people that will need to be involved in this strategy and some initial ideas on what they need.
Let’s now look at Process and the questions you need to answer there.
Step 2: Process
Process is something that I didn’t use to pay much attention to until I started to run into the problem of having too much data. Some of my clients started creating dashboards and reports left and right (which was great) but then they started duplicating their efforts. Someone would analyze the marketing performance and 3 weeks later, someone else would do the same.
I then started to think about how companies should approach the process of analyzing and sharing data. In this section, you can answer the following questions:
How are we going to share and consume data?
This starts to touch into the world of tools but you can start to think about how people want to consume and share data. Some people will want to jump right into the raw data and play with it while others simply want summaries of what is going on.
Your data plan needs to address these different needs and take into account the technical capacity of your team (refer back to the MTS assessment). You should also think about how people could talk about data in a collaborative way using tools like Slack or Microsoft Teams.
Where are we going to store best practices and documentation?
As your team gets familiar with the data, you’ll start to develop best practices including definitions for your most important KPIs and fundamental reports. You should be storing all of this information in a centralized location where everyone can access it and suggest improvements.
In particular, you should be capturing the following best practices:
- What are the most important KPIs to your business and how do you define them?
- What are the most important reports and are there any specific considerations to understand before using them?
- What tools and data are available right now?
What issues or challenges are we going to run into?
Finally, we can spend some time thinking about what issues or challenges you might encounter. This can include issues such as data accuracy, pushback from other teams, lack of centralized best practices, and issues in training. We will cover many of these issues in upcoming chapters and you’ll have a chance to revisit them very soon.
Now that you have an idea of how your company will manage and analyze data, we move on to Providers, everyone’s favorite topic.
Step 3: Providers
Whenever clients ask me about tools, I tell them that we shouldn’t start the conversation there. Instead, we need to focus on what questions they would like to answer. This is more important for your data-driven culture than any specific tool will ever be.
This helps them relax because most companies tend to have a good grasp of what questions they would like to tackle. How they answer those questions is where tools come in.
I like to group data questions into general categories and work with companies to figure out which categories are currently most important to them.
Over time, a company will be able to answer all of these categories, but we need insights today and can’t wait until we have all the data possible. Prioritizing at this stage will also help us cut down the implementation time needed to get up and running.
Here is a list of the most popular categories that you should think about:
- What is the performance of our marketing dollars?
- What are the common behaviors of our best users?
- How do we best communicate with our users?
- What do our users think about our products or services?
- How do we easily visualize data from multiple sources?
- How can we manage data ingestion and privacy?
I find that companies can typically focus on 1-2 categories at a time without feeling overwhelmed or spending months in implementation.
Step 4: Training
When I think about training my clients, I typically focus on 4 areas that will tackle different outcomes. These are 4 training “tools” that you can deploy:
Group Training: we start with group sessions to go over basic concepts and ideas. These sessions are practical in nature and customized for each team (marketing, product, etc).
Individual Training: we then move on to individual sessions where we can really dive into specific problems and challenges.
Reactive Training: in this format, I’m reacting to problems that come up as people start using the data and mining for insights.
Documentation: the last step is to create documentation on best practices. This helps capture answers to common questions.
Step 5: AI and ML
Finally, we can get to the “fun” stuff. Once you have your basics in place, we could look at how your team could use AI and Machine Learning to improve marketing, product, sales, design, and much more.
A data-driven culture starts with the right intent but typically end up with unexpected results. Don’t give your data magical powers that it could solve everything. It can help you dramatically grow your business but you still need amazing human talent and a little bit of luck.