A good friend of mine recently consulted on performance measurement for a government ministry that combined the two seemingly unrelated portfolios of Tourism and Economic Development. Their businesses couldn’t have a more divergent customer base. The Tourism portfolio focused on direct to consumer marketing: Families flying in from all over, attracted by the local sights, the beaches, festivals and hotel deals. The Economic Development portfolio was B2B, and worked to attract companies large and small to the area as well as fostering and sponsoring local startups. The metrics used to measure performance across these disparate portfolios were necessarily different, the customers were very different, and the mandates were definitely different.
But the Ministry decided a couple of years ago to spend some time implementing a data strategy that prioritized integration. Their strategy produced high-quality performance reporting for their leadership. My friend was also pleasantly surprised by the collaborative environment she discovered across the various teams, who seemed genuinely interested in each other’s problems and how they measured their impact on the people of their jurisdiction. Their data strategy focused on integration, but they’d also created a culture of openness across their very diverse portfolios of programs, encouraging people to share, collaborate and experiment. The result was both good reports and a positive culture.
A lot of organizations try to solve their data management problems by buying expensive tools and an expensive strategy and hoping that’s enough. While it’s often obvious this isn’t enough, the missing ingredient to those technology-first strategies isn’t usually as clearly defined as what my friend found in her ministry. In this short piece we’ll explain how this missing ingredient makes your data strategy a success, and how to set up your organization for a successful data strategy.
Usually an organization comes to realize they need a data strategy when they can’t get a coherent picture of their operations. Different teams produce different numbers out of their own reporting systems. Those reports float up to their senior executives, and those executives end up in conflict in daily meetings trying to reconcile what those reports appear to be showing about where the company is going.
The emotional cost of this conflict is difficult to quantify but it makes work difficult for everyone. It’s really hard to overstate how much of a toll this takes on an organization’s culture. Executives would prefer to make data driven decisions and not argue about the quality of their reports or their team’s work, but instead they’re forced into cycles of skeptical anxiety about what they say is happening in their business. Teams question whether they’ve got the right data to do their jobs, and when they don’t, customer experience suffers and their work with other teams does too, especially when there’s a handoff. And individual employees spend significant portions of their day wondering if they’re working on the wrong thing or missing something the data isn’t showing. The uncertainty caused by bad data creates conflict throughout the organization, from top to bottom, from the people tasked with creating the reports to the people who have to explain or use them or reconcile them.
Whatever a data strategy is for, at the highest level it’s about integrating data from various sources into a coherent picture of the organization. This is true whether you’re a government ministry with seemingly incompatible portfolios, trying to measure performance consistently, a consumer-packaged goods manufacturer with multiple ERP systems, or a startup with vaguely-defined customers. Organizations must collect data about their operations in different systems, and that’s just a fact of modern business. But it’s that semantic debt, those unintegrated orders or all those duplicate and slightly-off-kilter customer records, that causes anxiety up and down the reporting chain.
So a data strategy at start should be focused on integrating data across various sources. In the case of my friend’s client, they started with a data modeling exercise to get their arms around the ministry’s data. They wanted a single view of their business so they could produce integrated and consistent metrics. These kinds of concerted data modeling efforts are often deferred or even ignored entirely in a data strategy. Instead organizations will spend ever-increasing dollars on complex data engineering and data science solutions, all intended to overcome their basic lack of a good data model. Either that or they end up with toy solutions, fragile half-baked systems that cover a limited number of use-cases and fall apart a couple of months after launch. My friend’s government ministry did the work, and built a data model to consolidate their customer data and their transactions so they could do apples-to-apples comparisons across their portfolio. Data modeling at this level requires executive sponsorship and enough air cover to do the hard work of analysis, mastering and modeling. But it’s created a firm foundation for their analytics and the high-quality performance reporting she was pleased to discover. Alternative paths often produce accidental and unreliable outcomes that end up making everyone unhappy or, worst case, they look initially promising and then fall apart. A data modeling project is the centerpiece of a well-executed data strategy.
But clean and open integration doesn’t just happen in the data model, it has to happen across the organization. The data modeler’s job is to understand all the various business definitions of customer, order or product, across the business, on the way to building a coherent rationalization of those definitions. But to gather and reconcile those definitions the modeler needs to get the truth out of everyone. They need to talk to the people who work with customer, order and product data, understand their use-cases, and work to reconcile the different needs and constraints.
This is the missing ingredient to a successful data strategy, and the one my friend found made the biggest difference at her client. To be sure the quality of the reporting lowered the temperature on all those performance conversations. But it was because the organization had reset expectations about openness that they were able to get to quality performance reporting. They came to realize that reliable data integration requires transparency about operations. If one team doesn’t listen to another, or one group’s operational constraints aren’t acknowledged by the rest of the organization, then the data model will be missing pieces critical to the coherence of the big picture.
While it may seem obvious that to integrate data across the organization you need to make sure the organization can integrate it’s diverse perspectives safely and clearly, this step in the data strategy is often treated by my technically-minded colleagues as a nice-to-have. But it’s not. As a data modeler, I’m looking for all the perspectives. If Customer Service isn’t included because we don’t think their view of the customer is important, then we’ll be missing an important set of use-cases. If we don’t talk to a group because we don’t believe their needs are integral to the big picture, then our big picture won’t be as big as we need it to be. Somebody will end up with bad reporting, and the anxiety will spread. Even if all these disparate groups do is end up confirming the model, we’ve accomplished something important for our data strategy. This process of feedback needs to continue, as well, so that once the data model has been implemented it can adapt to new business practices and continue to be perfected. A first round of feedback and integration isn’t enough to make the process work. To succeed a data strategy has to live, continually refreshed with new use cases and new challenges.
The net for people planning their data strategy is to ensure they do the organizational work to make sure the strategy succeeds. You can’t make everyone’s days easier and fix reporting with just data engineering tools and whip-smart data scientists alone. You need to make sure everyone understands where they sit in the model, and how their use-cases get incorporated into the big picture. An open organizational culture will accelerate a data strategy that focuses on integration, and create a virtuous circle that makes for better days at work and high-quality reporting for everyone.
No comments:
Post a Comment