As you roll out data cataloging and governance in your organization, you want to strategize your roll out to ensure that it is high-value for your end users from day one.
Given how diverse data teams are, you will hear a lot of ideas on how a third-generation data catalog can be used in the organization based on everyone's needs and preferences. As a DataOps champion, in a situation like this when you are looking to nail down your first use case, it is imperative that you keep an iron-clad focus on just one specific use case to begin.
🌟 Best practices for designing a use case
🎯 Define a use case.
The most important part of any use case roll-out is knowing who this use case is for.
- Start each use case with a single user persona in mind.
- Make the user persona as “niche” as possible.
- Involve these people before implementation starts and truly understand their problems.
🚀 Use an agile experimentation-led implementation
Embrace an experiment-first mindset when implementing a use case. Take learnings along the way and implement them iteratively as you refine the process. A few things to remember:
- Each use case will take a cycle to reach maturity.
- Aim for an MVP with a single team or business owner.
- Collect feedback from end users.
- Improve and refine the process before taking it to production/scale.
✅ Ownership and accountability is key.
Like any project, for the use case to be successful, you need to find folks who will nurture it.
- Ensure that each use case has an owner who is responsible and accountable for driving the project or use-case (and incorporate this into OKRs if needed).
- For projects that involve multiple stakeholders or teams, make sure there is alignment on project from key leaders on the time commitment and owners before you prioritize the use case.
🎖 Don’t bite off more than you can chew.
Prioritization is key, so pick a problem and solve it end to end.
- Data teams are ambitious by nature, so it is normal to think of multiple amazing use cases to target.
- Remember, to effectively use the team resources, it is prudent to prioritize one use case, totally smash it, and then head on to the next one! ⚡️
⚡️ Creating a use case document that serves as a North Star
The use case document is where you and your team strategize and plan the use case. This is where you apply the principles from the "best practices" section above. The purpose of this document is to ensure everyone has the same understanding of the business problems that you want to solve.
Let's use an example to go through the process step by step:
1️⃣ Define business problem(s) and create measurable outcomes.
You should follow the "What? So what?" framework in this step. This is how it works:
- You ask yourselves "What do we want to solve?" Here you should clearly define the problem that you are solving for. For example...
- Question: "What do we want to solve?"
- Answer: We want to be able to trust all our public reporting KPI data.
- Immediately after, ask "So what will the outcome look like?" For example...
- Question: "So what will the outcome look like if we can trust all our public reporting KPI data?"
- Answer: We could reduce the data engineering dependency by 90% in the public reporting process.
2️⃣ Nail down your target user persona.
- Primary end-user persona: For example, in the example above, finance analysts would be your target end users. You could then pick Emma and John, who are the senior and junior finance analysts in the team.
- Secondary end-user persona: You should also think about what other teams might benefit from this. For example, the compliance team could be the second end-user persona.
- Understanding user pain points: For example, you should ask, what are Emma’s top three pain points? When does “trust” break for her in reporting KPIs?
3️⃣ Design user flows and an MVP.
- Draft Atlan user flows to be implemented: For example, here's an ideal user flow for Emma.
- Log into Atlan.
- Search for key metrics in Sales and Marketing, tagged “Public Reporting”.
- Understand the definitions of the metrics, along with their nuances.
- Trace the column-level lineage for “verified” data tables.
- Notify owners in case of any issues with the data.
- Scope and execute an MVP: Pick a small part of the overall use case (you can slice it by business domain, data sets, or other smaller pieces) to scope an MVP. The goal of the MVP is to implement a quick POC, take it to your end users, and get their feedback before you scale.
In our example above, to begin with, you could limit the scope to public reporting KPIs for sales and marketing.
- User interviews and observations during onboarding: Once you have your MVP, onboard the end users and record their feedback. Ask them if they could go through the user flows that you envisaged earlier. Check these questions:
- Did it work?
- Does it answer all the questions the end user persona had?
4️⃣ Identify internal owners and stakeholders.
In the context of our above example...
- Emily from Data Engineering, who is responsible for implementation, could be an owner for this use case.
- Matt, a Senior Finance Analyst who has business context and can ensure the use case is implemented, could be the stakeholder.
5️⃣ Identify learnings and the path to scaling up.
Host a retrospective with the core group of stakeholders to document their learnings. You could, for example, leverage project plans, Data Doc templates, etc in the meetings.
Every use case is different, but to give you some inspiration, we have created a template to help you get started with your first use case on Atlan. ❤️ Check it out below.
You can always reach out to your Data Success Manager if you want to chat more about defining use cases.