As data leaders, you must have often come across situations where you need to strategize an organization-wide roll out for a platform.
When it comes to rolling out a third-generation data catalog platform like Atlan, a really effective strategy is to think of use cases for end users that will add value in their day-to-day life. But, before you launch any of these use cases, you should first validate your assumptions by interviewing your end users. This will ensure that whatever efforts you invest are in the right direction.
The main objective of a user interview is to understand your end users and their pain points. It should enable you to visualize a user's current workflow and, consequently, envision a desired workflow — one that truly adds value to their existing process.
Set a clear goal
Before you conduct a user interview, it's critical to first get clarity on what you want to learn through this research. There could be different reasons for conducting a user interview, so find your "why". For example, it could be because you want to:
- Validate your hypothesis
As a leader you have certain assumptions around different end user problems. These assumptions are your hypothesis, which you should validate through an interview before acting on it. For example, let's say that your hypothesis is "Analysts lose productive time in asking/explaining to others definitions of business terms." To figure if that is actually the case, you can conduct an interview with a few analysts.
- Understand the daily workflows of end users
One important factor that goes into defining a good use case is knowing the current workflow and the desired workflow. A user interview can help you understand your end user's workflow and identify areas of optimization.
For example, a product manager might tell you, "To create X report, I need Y data, so I often reach out to data analysts and schedule a call with them to understand it." Now that you have this information, you can think of making this workflow more efficient. Perhaps a better flow would be:
- Product manager needs Y data;
- They log onto Atlan and find context around it;
- To clarify questions, they tag the data analyst in Atlan Chat;
- They can now use this data for reporting.
- Address recurring challenges
You want to increase your team's productivity by solving their pain points. A user interview will give you great insights into challenges your teams face and hence allow you to design a use case that helps the team.
- Get insights on how to roll out changes
If you want your teams to adapt to anything new, you need to find the easiest route to get them onboard. Conducting a user interview will give you firsthand information on what worked for your team in the past and what didn't.
Design your interview panel
In order to successfully conduct a user interview, it is important to choose the right panel of interviewers and participants.
- You need to make sure that you go in with a positive mindset to truly listen to your users.
- Build rapport with your users so that the discussion doesn't feel like an interrogation. Smile, keep a warm tone, and make them feel comfortable.
- Keep the format of the interview semi-structured — ask questions that lead you closer to your overall goal but don't read off questions abruptly from a script.
- Conduct a user survey beforehand to get responses from a couple of different folks. Then, based on the answers, invite a few users for a chat.
- You may skip this step if you already have a very clear idea of potential folks to have discussions with.
- In any case, ensure that you have selected a representative sample of your target audience. For example, if you are interested in understanding the pain points of the data science team, don't just invite data scientists. Speak to data scientists, data engineers, analytics managers, etc.
Design your discussion guide
The most important component of your preparation is designing a discussion guide.
A discussion guide is a document in which you formulate the questions you want to ask your participants. The point of the discussion guide is not that you ask every single question in it (and in sequence!). Treat it more like a reference document, a "skeleton" for your discussion, rather than a script.
Below we have shared an example for how you could structure your discussion guide and follow it up with user flow structures.
For example, imagine your core assumption is: Data analysts lose time in finding the right data for analysis.
User flows are essentially a visual representation of the steps that your users are following to go from problem to solution.
When you are preparing your discussion guide, make sure that you think of questions that will allow you to understand the participant's user flow. Once you have this current user flow, you can identify steps in the process that you want to optimize in order to reach a better outcome. Mapping out both your current and future user flows helps you truly understand these steps.
Below is a visual framework you can use to create your own user flows.
Example user flow
Let's say that you interview a BI persona in your organization. They create new dashboards for analysts from different teams. After interviewing them, you visualize their workflow:
Your current flow enables you to understand which pain points can be solved by introducing Atlan in the users' lives. This, in turn, allows you to visualize what an ideal journey for this user will look like with Atlan.
You can then create a new user flow to communicate changes with the team:
It's totally understandable and expected that participants might get excited about something at times and start veering off track. In these situations, remember that you have this person only for a limited amount of time (30 minutes to an hour at max), so you really need to take control of the discussion. Be polite and try to bring the conversation back on track.
Some of the phrases you could use are:
- I am so sorry to interrupt. This is great and I would love to know more, but being conscious of time, can we please also discuss...?
- This is super helpful, but I'm also keen to hear your thoughts on ...
- This is great. Maybe we can set up a separate session to chat about this in depth, but given time constraints today, can we please talk about...?
Tips and hacks
- Record the interview. It can be hard to always capture everything in notes, so we recommend that you record the sessions to fill gaps in your notes later.
- Block time to review notes. Try to block time right after the sessions so that you can always summarize your thoughts while they are fresh. Notes are a gold mine of insights, so invest time in putting together great notes.
- Team up. If possible, two people should take these interviews. One asks questions, and the other takes notes and prompts if the first person misses an important question. (This is also great when you want to check whether your understanding of a point from the interview makes sense.)
We have transcribed a user interview to help you understand how a typical user interview flows:
Goal of the interview: test the hypothesis that data analysts lose time in finding the right data for analysis. (Go into the discussion with a clear understanding of your objective.)
Interviewer: Thanks X, super excited to chat with you today. Just to give you some context — our goal through these discussions is to understand your daily workflows better and how data is weaved in our lives here at <org name> in different roles. If you have any questions for us, happy to take them towards the end of the call. Does that sound good? (Introduce yourselves and start with the intent of the call.)
(At this point the user would not or say "Yes".)
Cool, so what I have learnt is that typically in the data world, roles can be the same but day-to-day looks different. So if you could begin by telling us more about your day-to-day that would be perfect. (Start with a good prompt — typically asking them to tell you more about their day-to-day.)