By Brynne Henn - January 14, 2021
Building data literacy in an organization is a critical step to building a truly data-driven culture. Achieving a data-literate organization can create internal alignment around critical changes and help drive strategic decisions.
Unfortunately, there’s often a fundamental language barrier that divides business users and data experts. This barrier makes it difficult for these teams to effectively collaborate and quickly act on insights built on their data. For years, companies have tried to create a shared language through education and training, but this strategy hasn’t succeeded largely because the language barrier isn’t one-directional. On one side, business decision-makers find the data they receive difficult to understand. Without knowing precisely what to ask of their data teams, they avoid early collaboration with data teams for fear of saying the wrong thing. But on the other side, data teams often lack the domain expertise to provide the full picture of their analysis and find their lack of alignment with the business strategy gets in the way of delivering clear recommendations.
As we enter a new decade of analytics, it’s time to take a new look at our approach to data literacy. We need to stop asking business users to learn a foreign language just to understand why their metrics are changing. We have to enable a more consistent way for analysts to capture business context and communicate insights to their stakeholders. The key to unlocking data literacy is creating a shared language between data experts and business stakeholders, one that turns analysis from mere reporting and fact-finding to true data-driven decision making by starting with the metrics and focusing on the impact.
In most businesses today, the dashboards and reports experts turn to make decisions are focused on the data first. At a glance, data is presented in a quick, easy to understand format with the most relevant information upfront.
Presenting information this way is fine, as long as every stakeholder understands what they’re looking at.
The reality is, unless these dashboard consumers understand the minutiae behind the data, many details require deeper discovery. To make this concrete, consider the original dashboard: in their car, every driver can understand changes in fuel and speed, but when a warning light turns on, most are left scratching their heads. The dashboard is alerting them at a glance that something needs attention, but it’s not until the driver takes their vehicle to a mechanic or consults a manual that the driver actually understands the issue. In business, a dashboard might be communicating at a glance the different factors leading to retention, but when something changes, users often turn to an expert to explain, creating a delay between data and action.
Instead, adopting a metrics-first approach to analysis accelerates data literacy by establishing a common language across teams. Like speed and fuel, metrics are a universal language that help both business users and data teams understand how the business is performing — creating a shared context for discussions. Returning to the automobile analogy, an indicator light is not a metric, and fails the “shared understanding” test. With a shared understanding, any factors or combination of factors impacting the metric can be reported and discussed confidently.
One company leading with the metrics to understand data is Airbnb. As the company scaled and the complexity and volume of their data scaled with it, their Data Science team realized that they were treating data as merely cold, numeric terms. They were stuck answering ad hoc questions like: How many listings are there in France? What are the top ten destinations in Ireland?
As detailed by former Head of Data Science, Riley Newman, in his reflection post, Airbnb realized they needed to redefine how the company discussed and communicated the data on what’s driving their business. They did so by going back to the core metric that the business has cared about since its founding, customer satisfaction. They reframed their analysis, and the changes they were observing within it, around a shared “voice of the customer” metric. As he recalls, this change created a shared understanding across teams, saying,
“Data has become an ally. We use statistics to understand individual experiences and aggregate those experiences to identify trends across the community; those trends inform decisions about where to drive the business.”
By starting with the metrics that are to the business — like retention, CAC, and churn — and testing any new data against those metrics, you’ll build the domain expertise analysts need to communicate the nuance of their analysis and provide business users with the vocabulary and confidence to ask analysts about the data.
When analysts are treated as reactive, fact dispensing machines, analysis quickly becomes a mere measurement tool. Interactions between the business and data teams are shallow, often requesting one insight at a time: How many leads did we drive from our LinkedIn campaign? What are the top ten locations of our active users?
This type of reactive request will not create a data-literate organization. It reduces analytics teams to stats sheets, not proactive partners. Worse, this type of insight gathering not only obscures critical insights within the data, it often ties up the resources of a data team. For even the best analysts, a single request can take days, even not weeks, to understand what is relevant to a given business request, generate the queries to answer the question, and present the answers. This bottleneck between data and answers means analysts often only report on what changed in the business and leave it to business users to determine why it matters — without the confidence or literacy to make a data-driven decision.
Instead of getting stuck in how the data is changing, a data-literate organization should focus on both how the business is changing and what is the impact of that change. Focusing on the impact prioritizes both analysts and business users’ attention, enabling faster, more collaborative decision-making across the business. And with Augmented Analytics tools like Sisu, insights are stack-ranked by relevance, allowing every team to quickly understand the “what” and the “so what” in any change impacting the metrics, creating a more data-driven organization.
To achieve data literacy, and in turn, a data-driven culture, organizations need to take a new metric-centric approach that creates a shared language between data teams and business stakeholders. But with more data available than ever before, getting to a metric-centric approach requires a new interface to data, one designed for the speed, volume, and complexity of today’s data.
Augmented Analytics is this new interface to data. Not only does it accelerate the analysis of complex data, it directly addresses these two challenges of data literacy, creating a common language with a metrics-first approach to analytics, and focusing the conversation around the impact of a given decision.
To start building a common language for your organization, get in touch with our team.