The Hidden Benefits of Creating an Innovation Culture

By: George Bezerra, SVP, Head of Data

December 5, 2023


At TKWW, we love innovation, especially in Data. As the Data team, our job is to support the organization’s data needs across our core business areas, including Product, Marketing, and Revenue Operations. With a team of 55 data specialists, we are responsible for the full spectrum of data capabilities, including data analytics, data science, machine learning, business intelligence, data engineering, and dataOps.

We are confronted with new problems every day and have to use data to find solutions. We feel we’re best positioned to do so when we embrace an innovative mindset and curate a test-and-learn culture. We’re bold and unafraid to make mistakes—and we believe this approach makes us smarter and more efficient. We are always trying out new technologies, experimenting with new ideas, and finding ways to optimize our day-to-day work. If you’re a data expert or just getting started in your data career, we hope this article will be of interest to you—and provide some inspiration to cultivate an innovation culture. Here are some examples of our achievements in the last 18 months:

  • Experimentation platform: We prototyped and eventually deployed an experimentation platform, improving automation and self-service capabilities for A/B tests. Paired with other improvements, this allowed us to double the number of experiments we run per year!
  • Machine learning platform: We built a custom machine learning platform that helps us deploy models in production via an API in addition to providing development support, automation, and observability. Our new models built on this platform are having a huge impact on our product.
  • Modern data stack: We overhauled our data infrastructure by integrating best-in-class tools from the modern data stack into our platform. This includes the adoption of DBT for data transformation, Airflow for orchestration, Airbyte for ingestion, Terraform for infrastructure as code, and GitHub Actions for CICD.
  • Coding Copilot: An AI-based coding copilot solution was originally brought in to TKWW as a way to help increase software engineer efficiency. Our analytics team saw the opportunity and found ways to introduce Copilot into their Python workflow. Their approach is now getting broader adoption throughout the Data team.
  • LLMs and AI: Our data science and ML teams started a reading group focused on AI and quickly became the experts on the topic at the company. Since then, we built the infrastructure to support LLM-based work and already launched two LLM-based product features!
  • Data apps: At TKWW, we believe in treating data as a product and increase value delivery and provide self-service capabilities. In Q3 this year, our DataOps team built the infrastructure to support deploying Streamlit and R-Shiny applications developed by analysts and data scientists. In that same quarter we deployed 6 data apps, with applications ranging from forecasting to self-service opportunity sizing and semantic search.

Many of these advancements originated as ideas from individual team members. We quickly prototyped these concepts and, if they showed promise, expanded their implementation across the organization. While some attempts did not succeed, these failures resulted in learning and critical thinking opportunities. However, sometimes innovation projects do succeed, creating substantial business value.


Innovation’s hidden benefit: Higher employee productivity

There are obvious benefits of innovation, such as increased competitive advantage, operational efficiency, and revenue growth. However, innovation projects tend to be risky and their benefits are usually long term. As a result, those projects are often deprioritized in favor of shorter term initiatives that have a more direct impact. For many companies, the challenge of balancing the pressure to meet short-term financial goals often leaves limited resources for innovation and risk-taking.

However, innovation does not have to compete with short-term goals and, in fact, what we have observed is just the opposite: investing in innovation can boost team morale and productivity overall, including on short-term initiatives. When employees are encouraged to work on innovative ideas, they become excited, more engaged, and creative. They find more meaning in their work and, as a consequence, they work faster, more accurately, and they find creative ways to achieve more in less time. As a result, we get higher overall velocity on the short term initiatives in addition to producing innovative work that wouldn’t have occurred otherwise.

This concept is often misunderstood by executives. Leaders usually strive to prioritize only a few initiatives to eliminate distractions and ensure the team will deliver on the most important short-term goals, often reducing the bandwidth allocated to innovation to almost none. However, this results in the opposite effect; employees get demotivated and less creative, leading to slower teams and an overall drop in productivity. The hidden benefit of innovation is paradoxical: by allocating more bandwidth to innovation (and less to short-term initiatives), you actually increase the team’s output on short-term (and long-term) goals.


How we drive innovation at TKWW’s Data team

There are different ways companies and teams can foster innovation. For example, we split our bandwidth into 75% support and 25% innovation, where support is work being requested by the business and product teams while innovation are projects being driven by us. The 75%-25% split is just an overall guideline, and the reality differs by team. For example, DataOps tends to have more bandwidth for innovation than analysts.

With that said, the most effective approach we’ve seen to drive innovation is to create an innovation culture. Below are some strategies we employ to achieve that:

  • Bringing in ideas from outside: We have weekly reading groups where team members read and discuss technical papers and blog posts. We also encourage the sharing of relevant material in our slack channels and discuss ways for how to apply that knowledge to our work.
  • Sharing ideas internally: We have functional group meetings where we share our technical work with peers, leading to higher visibility on cool and new ways of doing things.
  • Sharing ideas externally: We’re finally reaching a maturity level where we can start sharing our work with the external world. This engages employees and motivates them to do “cool” things.
  • Rewarding innovative behavior: Data leaders encourage innovative and risk-taking behavior even when they fail (which is often the case). Our focus is on proactive behavior that triggers learning, not necessarily the success or failure of an initiative.
  • Support for learning and development: We are constantly challenging our team members to learn new skills and grow while providing them support on that journey.
  • Creating room for trying new things: When employees are under pressure to deliver quickly, they feel discouraged to try new ways of doing things. Managers help team members by pushing back on stakeholders and encouraging their team to expand on the original scope to explore alternative approaches that lead to learning.

Several of our success stories are from individuals who decided to try things on their own and built a POC or demo that got support from others, leading to a full-blown project later on.


Closing thoughts

Innovation is a great way to boost a team’s morale and overall productivity and it does not have to compete with short-term initiatives—in fact, the opposite is true. By reserving more time for innovative work, team members deliver more on both short- and long-term goals. This goes against the natural instinct of executives who deprioritize innovative projects to keep the team focused on short-term deliverables. The best way to drive innovation is by creating an innovation culture, where individual team members feel empowered to try new things and learn from their failures.