What is a data asset? While there are many different types, here is how we categorize the most common ones:
- Raw data - Customer behaviors, financial performance, etc.
- Business Logic - ETL pipelines, tables, and datasets
- Knowledge base - Business stakeholder-facing dashboards, reports, and SQL queries
We’ll be focusing on knowledge base data assets, which have the most direct impact on business outcomes. These types of data assets become more valuable (and expensive) the further downstream we go due to the accumulated value. So if they're so valuable, shouldn't we treat them accordingly?
Instead, we see data assets get treated as one-time-use deliverables because there aren't necessary infrastructure and processes in place. What happens is the data team receives unplanned, "urgent" ad-hoc tasks (aren't they all…) in the middle of a sprint. These ad-hoc tasks tend to get prioritized and executed with speed in mind (not reusability).
The common questions arise:
- Wasn’t something similar asked last month?
- Who worked on it last time?
- Where is that code?
After a frustrating search for previous work, the realization that it's "just faster to start from scratch" follows. The task gets completed and the data team gets back to their sprint, but the asset disappears into the depths of some folder, never to be seen again.
With past work becoming difficult to find, data team efforts result in redundant work, lost institutional knowledge, and wasted time. We need to break this vicious cycle and stop treating these assets as disposable items.
We must prioritize efficiency in maximizing the value of our analytics efforts. How? By building high-quality, reusable, and readily-available assets. Not only will this decrease the amount of time it takes to produce new data assets, but it will also increase engagement with your data assets.
The Pareto Principle & Your Data Assets
To make the most of your data assets, following the Pareto principle is most effective: resolve 80% of your tasks by utilizing 20% of your organization's previously-created data assets. But how do we make that 20% easy to reuse?
The most important step is making it extremely easy to find and repurpose previous work, both your teammates' and your own. This creates a future-oriented workflow across the entire data team that collectively minimizes the amount of time spent creating data assets. We call these principles being ADLT.
Teams can’t reuse what they can’t access, so facilitating access to key BI Tool assets and SQL queries is critical. Establishing secure permission controls will enable data assets to flow more freely. Accessible data assets will break down information silos that create inefficiencies within data teams and across the business.
People trust assets they understand. To help your teammates confidently reuse work, create clear documentation that includes metric/field definitions, outlines data model dependencies, highlights previous usage, and anything else that would be helpful to know for an audience that is unfamiliar with the asset. Not only does documentation enable more consistent asset re-use, it also allows you to carry on with your tasks without being distracted by repeated questions about past analysis.
Once you've made your work accessible, it's also important to make your assets easy to find. This includes using shared folders, spreadsheets, git repositories, or wiki pages. Proactively publishing your work will make it faster for you and your teammates to find what they need, when they need it, independently.
Working on repetitive tasks is a waste of your valuable time. We encourage the following: STOP HARD CODING YOUR WORK. Instead, parameterize values in your SQL queries to create templates and build BI assets with modular filters to allow for reusability. The goal is to avoid starting from scratch whenever possible.
Resolving 80% of tasks with 20% of assets
How does all of this translate to the Pareto principle? Creating ADLT assets (assets that are accessible, well documented, located easily, and properly templatized) helps with several challenges that data-driven teams encounter.
The two most important:
- When we receive a task our assets have already resolved, we are able to simply copy-paste the SQL query and fill in the parameters with current use-case values or filter for the right values in our existing BI assets
- When we receive a similar task, we can use our generalized assets as a base for creating the new one (by editing the code or report, changing filters, etc.)
ADLT assets become an analytical swiss army knife that can be repurposed to resolve the majority of analytics tasks that are thrown your way.
Once data assets meet the ADLT principles, you must maintain them to mature over time. You can make your assets as accessible, as documented, as located easily, and as templatized as you like, but if they are out of date, it becomes a wasted effort. Continuing to maintain the reusability of data assets will save you and your teammates from recreating the wheel with each new analysis task.
The ADLT Framework & Your Data Assets Value
The value of BI Tools and SQL queries can be measured by the amount of actionable insights they generate for the business. This is dependent on the ability for business stakeholders to engage with the right data assets, for the right problem, at the right time.
Let’s take the example of BI Tools. Most companies have dozens (maybe hundreds) of dashboards that are published across the organization, but using these data assets isn’t inherently valuable. Dashboards and reports become valuable when stakeholders find relevant answers for the problem at hand. This means teams should aim to create data assets that help the right audience engage with the right asset, at the right time.
The more insights generated, the more valuable the data asset.
So how do we build valuable data assets more efficiently?
This is where the ADLT framework makes a difference. Developing assets that are strategically organized, clearly understood, and easy to find will increase the insights generated from stakeholder engagement. Making reusable BI Tools and SQL queries will reduce the amount of time needed to resolve future tasks and increase the long-term ROI of data assets.
The Efficient Path to Becoming an ADLT
Following these practices not only reduces wasted time and effort, but also increases the output of business insights. However, being an ADLT isn't easy. Analysts who want to boost the ROI of their BI Tools and SQL queries need to use the proper tools and adopt streamlined work processes.
Developing and maintaining a well-oiled data operation machine can be much more chaotic than you'd think. Thankfully, some messy problems have elegant solutions. Rupert will help you easily reduce the cost of your assets, increase their utilization, and accelerate data-driven insights. Check us out here.