In this post, Tom Arundel discusses five key areas of focus for a Digital Data Center of Excellence, and why one is key to scaling digital transformation. Tom is currently Global Learning & Insights leader at Quantum Metric, and formerly Director of Digital Product Performance at Marriott International.
As large, matrixed organizations strive to be more data-driven, they often end up with data silos instead of customer insights. As a result, digital product managers are tasked with setting roadmap priorities without knowing what features might be most valuable to users or understanding the dollar impact of issues causing friction in the purchase funnel.
Instead of spending more time listening to users and eliminating customer friction, prodcut managers are often busy chasing arbitrary metrics. Increasingly, data is either unavailable, inaccurate, or inconsistent. And to top it off, new third-party technologies aimed at improving user experiences often impose risks to page performance, security, and privacy of consumers.
In this post, I’ll explain why a Data Center of Excellence (CoE) is key to scaling your digital transformation, as well as strategies for incorporating it into your digital operation. We’ll provide five areas you can focus on.
So what is a Data CoE and why does it matter?
This opened the door for new and specialized digital customer experience (CX) tools to hit the market. In addition to standard web traffic, page flows, and click metrics, there were session playback, heatmap, and page performance measurement tools, not to mention third party marketing, advertising, and personalization technologies.
This panacea of new real-time reporting tools opened up new insights about customer behavior and removed bottlenecks in ways that democratized data. However, it also created a new set of problems. Suddenly, digital product managers were flooded with data but no insights. To top it off, as new third-party code proliferated across their pages, there was increased risk of slow page performance, privacy, and security issues.
To make matters worse, different data about the same customer was scattered across multiple different vendors in various reporting tools, owned by different departments. As accessibility to full-time analysts and data scientists was scarce, product managers were increasingly pressured to become part-time analysts, spending valuable time trying to make sense of data to make informed decisions.
Enter the Digital Data Center of Excellence (CoE)
This is an internal group of cross-discipline experts (often spread across Analytics, Data Strategy, Digital Product, IT and Marketing teams) that act as data stewards, with the goal of accelerating adoption of data across the enterprise. They provide leadership, best practices, support and training, in order to remove barriers to communication that can impede product launches and cause friction in digital user experiences.
A CoE understands deeply the people, process, and technology necessary to enable a successful data-driven organization. It serves as the information hub between digital product lines, internal stakeholders (IT, data scientists, etc.), and external technology vendors and partners. Instead of separate teams working in organizational silos that slow momentum and effectiveness, a CoE works on one mission, one vision, one strategy and one goal.
A Digital Data CoE can take on several areas of responsibility. I suggest five key areas that can enable more data-driven organizations. Let’s break them down:
- Tag Management Oversight – As companies have moved away from hard-coding tags on pages (and removing dependency on IT deployment cycles), tag management systems (TMS) have taken a central role in activating and deploying third party technologies that capture data about user interactions. A successful Digital Data CoE should be intimately involved in defining the people, process, and technology that runs the TMS. When properly architected and governed, a well-organized TMS team (comprised of business systems analysts, developers and Quality Assurance) can quickly and seamlessly deploy and/or fix tags, activating new product and marketing insights, and improving accuracy of reporting downstream. As we move to a cookieless world, TMS teams also need to plan for the next phase of future server-side and API-driven data capture. Tag management is a critical foundational element to the next four areas below, including data quality, page performance, privacy and security.
- Analytics Oversight, Standards, and Training – Analytics tools require documentation on how data is captured and configured allowing experts to train stakeholders on self-service access. In this role, the key to success of the CoE is removing bottlenecks that prevent data from being easily shared and interpreted. To start, tracking requirements should be clearly and consistently defined. This includes developing standards for data capture, data dictionaries and documentation that can be easily shared. Once data is captured, instead of spoon-feeding reports to product managers, a successful CoE advises and trains on self-service access, and how to interpret data across the various tools. A CoE can communicate potential pitfalls in data quality, and highlight misleading trends, but ultimately doesn’t need to be (nor should it be) a single team of “data hoarders” that stakeholders depend on entirely for insights. Stakeholder training is never a one-time exercise, but rather can be accomplished through routine open office hours, along with up-to-date documentation on knowledge platforms.
- Data Quality Oversight – Developing trust in data integrity can be challenging when data is captured across various tools, sites and platforms, then fed into multiple downstream reporting systems. However, it’s possible to govern data quality by creating a CoE “Swat Team” of data experts who collaborate on routine data quality reviews. The team can be comprised of a cross-disciplinary team of tagging experts, analysts, dashboard managers and data strategists, with the common goal of identifying and eliminating data defects across multiple reporting tools and systems.
- Page Performance Monitoring – As the technology stack grows, so do the number of tags, images, CSS, content platforms and backend APIs that power—and also slow down—front end user experiences. Good performance and session playback tools can quantify impacts to revenue and conversion, as well as help isolate technical root cause. It is important for the Data CoE to have a hand in monitoring and analyzing page performance, but they cannot be solely accountable. Product owners, IT managers, and other stakeholders must share accountability and be trained to analyze and understand when their platforms or code are slowing page load times.
So why should I care about this? And what’s next?
If you work in a large digital operation and don’t have a Data CoE, there’s a good chance your business is still operating in data silos. This is likely resulting in sub-optimal efficiencies, lost customers, deteriorating loyalty, and millions in wasted revenue opportunity. Not to mention the slow page performance, risky malware, and sub-optimal experiences for your customers.
By building a CoE practice within your organization, you’re freeing up the decision makers to focus on what they do best – building great products. One industry expert compared the CoE to the coaches and the product managers (stakeholders) to the players in the field.
And by educating stakeholders on data self-service, full-time analysts are available to develop insights, instead of running reports. This also frees up full-time analysts to document best practices and create and host training/office hours, all with the mission of building trust in the data.
Developing a Data CoE is a journey, and it won’t happen overnight. Small, thoughtful steps can lead to great success.