Part of the struggle I think is not only the 'curse of the generalist' but that data teams tend to set themselves up to do one of the activities particularly well at the expense of the others.
If they primarily work like engineers, (e.g. big backlog of jira tickets) they may not feel like they have the time to also craft narratives. They are set up well to answer business questions at scale and over time, but don't give themselves space to set dig deep, gather novel data, reframe a key question, and present insights in a clear way that directs answers a question.
If they are primarily journalists/analysts (e.g. your question is X, take 4 weeks and figure it out), then will have to answer the question in a larger context, and the analysis will be more comprehensive.
I don't know if the answer is to let a single team switch modes as needed (I've tried this before - it is hard to manage), but it certainly would be helpful to have clearly defined modes and tell people what role they are in when they start an analytics project.
This is an excellent point. It's not just that individual data professionals will have more and less experience in certain modalities, it is that a team's rhythms (which can't be easily changed on a dime) are optimized for certain modalities. This idea naturally drives towards the conclusion that different teams should pursue these different modalities, but that seems like a Bad Outcome to me. Similar to what we've talked about recently with over-specialization.
Looks like its time for analysts to measure engagement, open and click through rates of the shared analysis to predict and craft the next story :-)
!! love that insight and totally agree.
Part of the struggle I think is not only the 'curse of the generalist' but that data teams tend to set themselves up to do one of the activities particularly well at the expense of the others.
If they primarily work like engineers, (e.g. big backlog of jira tickets) they may not feel like they have the time to also craft narratives. They are set up well to answer business questions at scale and over time, but don't give themselves space to set dig deep, gather novel data, reframe a key question, and present insights in a clear way that directs answers a question.
If they are primarily journalists/analysts (e.g. your question is X, take 4 weeks and figure it out), then will have to answer the question in a larger context, and the analysis will be more comprehensive.
I don't know if the answer is to let a single team switch modes as needed (I've tried this before - it is hard to manage), but it certainly would be helpful to have clearly defined modes and tell people what role they are in when they start an analytics project.
This is an excellent point. It's not just that individual data professionals will have more and less experience in certain modalities, it is that a team's rhythms (which can't be easily changed on a dime) are optimized for certain modalities. This idea naturally drives towards the conclusion that different teams should pursue these different modalities, but that seems like a Bad Outcome to me. Similar to what we've talked about recently with over-specialization.