Throughout my substack, we’ve looked at a variety of topics, problems and solutions in the data space. From a technical point of view, we now have most, if not all, of the solutions we need for problems. After all, what technical problem in data do you think hasn’t been solved? What use case doesn’t have tooling, technology or software to solve? I need only refer to Matt Turck’s diagram once again. The data Summits last week show even more progress towards complete and easy-to-use tooling.
This wasn’t true 10 or even 5 years ago, but now, from a technical point of view, we need to focus on teaching folks who want to enter the space, upskilling the people we have or possibly making existing tooling easier to use with the help of AI. I know of many who are moving from engineering into sharing knowledge in some form, whether that is teaching directly, teaching on another platform, running an independent data community, running a vendor-associated data community, writing books, or even running courses at established universities. We can also reimagine existing tooling to make it better to use (think moving to Dagster from Airflow), which will continue to happen for a long time - and rightfully so.
So, in my mind, there is one big frontier that is often difficult for data folks. It’s not always difficult and it varies from company to company, department to department and season to season. I’ve experienced it working well and badly in my career. A general way to describe this could be: “The Human Interfaces of Data”.
What exactly do I mean by this? Pretty much what it says on the tin. Where do data folks and tooling need to interact with people elsewhere in the business? What is difficult about this today? What doesn’t work? What does work? How do we move from friction and difficulty to a better position? It’s not just limited to commercial stakeholders - it can be upstream SWE teams and 3rd parties, too.
I realise that it’s such a big problem, with such variety depending on the organisation (specific humans), that there is no way that I can know the answers to these questions on my own, there is no way that I can generalise, either. I need to collect more data…
My plan is to interview a number of people who are right on the edge of this interface and for them to purposefully be approaching it in different ways. Data practitioners - IC and EM, consultants, vendors and VCs (I’ll try to find some who were practitioners before) - our whole ecosystem. It needs to be our whole ecosystem, as each one of us, in each part of the ecosystem, in each different org, gets a fractional view and experience of the problems. My aim is to collect some of these views and experiences and piece them back together in the hope of being able to generalise afterwards.
I’ll ask all of my guests a set of standard questions, then I’ll drill into their answers a bit afterwards with some follow-up questions, finally publishing each interview as a post with some commentary. It won’t be every week as I coordinate with guests, but it most likely will be the longest series I’ve done to date.
At the end of the series, I’ll try to collate themes from all of the interviews and see if I can create some kind of handbook as to how data folks can win in different situations. I expect there will be some un-winnable situations found too, but these are just as valuable to uncover. Knowing when to save your time and energy and to move on is very helpful.
Humans, the final frontier.
The Questions
Give us a bit of an introduction about yourself:
Your education, briefly
How you came to data in your career
A bit of an overview of the places you have worked in data and what roles you have held
If you’ve moved on from a data practitioner role, when/where/why
Can you describe the human interfaces you have had as a data person:
Ones that existed
That worked well
What
When
Why
That didn’t work well
What
When
Why
If you compare the times that did work well and didn’t, can you pinpoint the circumstances that were the cause of the difference?
Ones that didn’t exist, but should have
Why should they have done?
What were the consequences of them not existing?
What are you doing now in your current role as X, that helps make human interfaces in data better?
Tooling
Process
People
Thesis/Investment
Other
Who I’ve got signed up for the series, so far:
Ilan Man @ Brooklyn Data
Taylor Brownlow @ Count
Michael Rogers @ Lyst
Caitlin Moorman @ Indigo
Abhi Sivasailam @ Levers
Let me know if you’d be interested in being a guest! Ping me on LinkedIn or dbt/LO slacks.