Data for Change
Google “What is data democratization” and you will see all the top results talking about “access to data” as the key to democratization of data (unless you found this one which is awesome).
However, just giving data access — whether as raw data in a data warehouse or as beautiful visualizations inside a Product Analytics tool or a Business Intelligence tool — is certainly not data democratization.
So what is it?
Data democratization is the ongoing process of enabling everybody in an organization, irrespective of their technical know-how, to work with data comfortably, to feel confident talking about it, and as a result, make data-informed decisions and build customer experiences powered by data.
An organization that truly wants to democratize data needs to embrace the following principles (referred to as the trifecta of data democratization throughout this guide):
- Empower employees to feel comfortable asking data-related questions
- Provide the right tools to enable everybody to work with data
- See democratization of data as an ongoing process which might even require an organization-wide cultural shift
Before delving deeper into the above, allow me to digress.
Data Challenges
Mixpanel recently conducted a survey whose respondents consisted of folks working in product teams across a wide array of industries, at companies ranging from less than ten to more than a thousand employees.
The goal of the survey was to understand the good, the bad, and the ugly about the relationship between product teams, and well, data. Here’s the full report if you’d like to dig in.
When asked about challenges, the most common ones that came up were as follows:
- I don’t have access to the data I need
- I can’t trust the data
- I have access to data but lack the skills to find answers to questions
- The analytics tools my company provides aren’t designed for product teams
- Data experts at my company are too busy to help me
If one or more of the above-mentioned statements are deemed true by your employees, it is safe to assume that data democratization at your organization needs work.
What’s interesting is how these challenges map to the principles mentioned above.
It’s time to delve deeper into the trifecta of data democratization.
Start with making data literacy table stakes at your organization.
Data literacy should no longer be seen as a nice-to-have and everybody should be given access to resources they need in order to become as data-literate as they wish to.
For some, it might be enough to understand what data is gathered and what it looks like.
While others might find it worthwhile to even know why certain data is tracked, how it’s done, where it is stored, and in what format.
Therefore, data literacy solves one of the biggest bottlenecks in data democratization — access to data.
Access to data but what data and where?
Well, when someone says that they don’t have access to data, they can refer to raw data in a database, transformed data in a data warehouse, data in the form of visual dashboards, product usage data inside a product analytics tool, transactional data in a subscription analytics tool, demographic data in a customer engagement tool, data about marketing campaigns in a customer data platform, and so on. You get the picture.
And when that person can specify where they wish to access what data, providing access becomes a lot less complicated. Also, if that person is given access to the right data in the right place at the right time, it is far more likely that they will trust the data.
So the next time someone says they don’t have access to data, and they are unable to specify where they want access to what data, you have a data literacy problem to solve.
Different shades of data literacy
It is evident that data literacy is not limited to knowing how to write SQL queries or how to analyze complex reports.
Every team needs some form of data to execute their day-to-day tasks or to analyze the impact of their work. But different teams with different data needs require varying levels of data literacy.
Very different skills are required to implement data tracking, to derive insights from data, and to act upon those insights. Further, acting upon those insights by running data-led marketing campaigns requires a different skill set than that required to identify the right prospects to go after by looking at the same data inside a CRM.
Similarly, building predictive models and delivering personalized experiences in real-time rely on different types of data as well as very different skills. The former requires training in data science while the latter is a problem for data engineering to solve.
It is safe to say that data literacy, in some shape or size, has become a prerequisite for individuals to excel at their duties. And companies that invest in making data literacy accessible to their employees are sure to make their competitors play catch-up.
Now that we agree that data literacy is table stakes, the next principle in the trifecta of data democratization is to enable everybody to work with data by investing in the tools that enable them to do so.
That begs the question..
To answer this question, first let’s look at how different teams typically work with data.
- Marketing works with data to create engaging, better-converting content
- Growth works with data to run experiments and deliver personalized experiences
- Product and Engineering work with data to build features that are actually used and kill the ones that are not
- Support works with data to deliver faster resolution (by seeing what a user has done or not done inside a product)
- Customer Success works with data to deliver a better customer experience (by asking them the right questions based on usage patterns)
- Sales works with data to identify prospects that are likely to convert (by looking at the actions they have performed during the free trial).
Can a couple of tools really do all of the above? Hell no!
And this is just a very high-level overview of the most common ways of working with data. And this doesn’t even include the needs of data teams that need a whole bunch of other tools to ensure that the right data is made available in the right format in the right systems at the right time.
Product and Growth teams alone need at least half a dozen tools to do their job well (I’m talking about best-in-class tools and not do-it-shabbily-but-do-it-all ones).
Seriously! I’m not even exaggerating.
A craftsperson is only as good as their tools
Every tool is becoming a data tool forcing every team to become data-literate.
Product and Growth teams need a tool like Mixpanel or Heap for Product Analytics, a tool like Hotjar or FullStory to gather qualitative data, a tool like VWO or AB Tasty to run A/B tests, a tool like Userflow or Userpilot for guided onboarding, a tool like Intercom or an alternative for 1:1 conversations at scale, a tool like Customer.io or Userlist for lifecycle email campaigns, and a Customer Data Platform like mParticle or Segment for data unification.
There you go — seven tools already and we’re only talking about basic data tools for Product and Growth teams. Besides the above, every company needs a suite of other data-specific tools which, depending on the company size, can be managed by a dedicated data team or just reside with the Product team.
At the very least, businesses, especially ones that capture a lot of data, must invest in a Data Warehouse (DWH) like Snowflake or BigQuery to make all the data available for analysis and downstream action, a Business Intelligence (BI) tool like Looker or Mode that sits on top of the DWH and enables self-serve analytics, a data integration (ETL/ELT) tool like Fivetran or Stitch to move data from external systems (like the seven tools mentioned above) into the DWH, as well as a reverse ETL tool like Hightouch or Census to move data from the DWH to external systems for downstream action.
As you can tell from the above wishlists, in the context of data analysis alone, companies today need a Product Analytics tool as well as a BI tool — they both serve different purposes for different teams. And this is besides Google Analytics or an equivalent for web analytics that Marketing needs.
Phew! That does seem overwhelming but it is the future — embrace it or be left behind.
It is crucial to invest in a set of tools that enable individuals to work with data efficiently and make data-led decisions without relying on others. It makes everybody productive and keeps the team morale high!
Moreover, implementing best-in-class tools that do the job well is better than spending countless hours looking for the ideal tool, or even worse, deciding to build something that can easily be bought.
Build vs Buy is a topic for another day but I must say that whichever route you take, make sure to evaluate how your decision impacts your employees — particularly how it affects their day-to-day work and long-term goals.
Keeping that in mind, let’s address..
I’d like to start by saying that the size of a company and its growth trajectory heavily impact the pace at which data democratization takes place. Needless to say, building a data democracy is much easier in the early days of a company as it’s easier to mould the culture that supports it.
Larger organizations are likely to face a slew of challenges and data democracy can also be perceived as data democrazy!
Keeping that in mind, the larger an organization, the sooner should it invest in the process of data democratization.
For the sake of keeping this guide from turning into a book, I will only focus on the needs of medium to large companies that have dedicated Marketing, Growth, Product, Engineering, Support, Customer Success, and Sales teams.
So, why is data democratization an ongoing process?
Because it relies on data literacy which is also an ongoing process. The world of data is experiencing unprecedented growth and the rate at which tools and technologies are evolving is fascinating. But it’s also hard to keep up with and frankly, a little annoying for most people outside of the data space because of how it impacts their work.
At the very least, everybody in an organization, irrespective of their role, should be able to get answers to their data-related questions effortlessly.
Additionally, how various teams work with data and to what extent should become general knowledge within an organization. It should be easy for employees to know who has access to what types of data, where the data resides, and what is the process of getting access to that data or asking questions of that data.
Dataportal by Airbnb is a fine example of how larger companies can democratize data by allocating dedicated resources to solving this mammoth problem. Projects like Dataportal certainly require ongoing resources but the payoff seems to be worth the effort.
Finally, what kind of a cultural shift are we talking about?
One of the challenges mentioned in the Mixpanel survey was Data experts at my company are too busy to help me.
Data democratization needs a cultural shift that makes this challenge obsolete — a thing of the past in your organization.
Everybody who relies on data to excel at their job and meet their goals should become a data expert.
Everybody in the organization should feel confident talking about data and be equipped with the tools and the knowledge to work with data and to get answers to their questions without any dependencies.
Lastly, everybody in the organization should be given the opportunity to make meaningful contribution to data-related projects.
There is no proven one-size-fits-all approach to building a data democracy but empowering people seems to be a logical step in that direction.