If the best part of a data conference happens in the hallway, we should probably put it in the programme.
Speakers and most attendees who have been to an event already understand the value of those conversations. Someone challenges an assumption, explains a failure, or helps connect a session to a problem waiting back at work. We know why being there matters.
Try explaining that to the boss paying for flights, hotels, a ticket and days away from work. Travelling to the other side of the world to hang out with interesting people is a difficult business case, however valuable those people might be.
“Why can’t you just watch it on YouTube?”
That is the comparison every part of the event has to survive. It is also a reasonable question. The manager sees a substantial bill alongside an apparently free alternative. The attendee is trying to explain something whose value depends on the quality of conversations that haven’t happened yet.
Counting sessions is easy. Measuring judgement, trust or the question that changes your approach is much harder. We need to make that value understandable without pretending we can guarantee or precisely measure it.
The large vendors keep pushing new things. For instance, Microsoft’s Fabric updates blog and the Databricks release notes make the pace visible: a continuing stream of features, previews and changes. Every few weeks there is something else to understand, evaluate or explain.
Quality Isn’t Free - Time, Money and Committment Come With a Cost. #
An event programme can become another delivery mechanism for that stream. Everyone scrambles to keep up, while the underlying question gets squeezed out: which changes actually matter to our work?
Meanwhile, YouTube offers more material than anyone could reasonably watch. There is excellent teaching, poor advice, and AI slop adding to the noise. Another video does little for someone who needs help deciding what applies to their situation.
Human interaction should be central to why we gather. Its quality depends on the people contributing.
In the data community, we have helped establish the expectation that speaking and creating content happen without payment. We have made this bed ourselves, over the course of many, many years. Generosity became normal, then normal became an assumption that expertise and preparation would continue to arrive for free. If you try to monetize your expertise, you will face criticism - that’s just the way we’ve built the community.
A substantial new session can take 50–100 hours to research, build, test and rehearse. That is serious work, fitted around jobs, families and everything else that demands attention.
Then an organiser uploads it to YouTube.
Sharing knowledge has tremendous value. But publishing a recording also changes the bargain for the person who created it. The talk loses some exclusivity. The material gets stolen or “inspire” what is essentially a carbon copy of the video. Reusing it elsewhere becomes less attractive to organisers seeking fresh material, even though the preparation cost remains the same.
We cannot keep asking for original sessions while making their useful life shorter and treating their creation as free.
Established talks deserve a place. New work deserves meaningful support. Commission sessions where the programme has gaps, pay where possible, and offer travel support, mentoring and editorial help. Agree on recording, publication timing and reuse with speakers in advance.
Ticket revenue should help fund the people and material that make the programme worth attending.
Keep sessions as the backbone, with an emphasis on judgement, experience, failures and lessons learned. It is always going to be easier to google the “how”, all while the “why” is almost always the more important question. Then curate them into learning tracks with explicit outcomes. An Analytics Architecture track might connect governance, deployment, security and operating practices around the decisions an architect needs to make.
Explain the core sessions, specialist choices and what participants should be better equipped to discuss afterwards.
Then build the hallway track into that path.
Build on the Track and Make the Whole Greater Than the Sum of Its Parts. #
After a group of related sessions, schedule an ask-me-anything with the contributing speakers. Give it a defined subject and a facilitator. Let attendees submit questions beforehand or ask them in the room. Connect the answers to the choices and tradeoffs raised in the sessions.
An architecture AMA could examine how to approach conflicting governance requirements or decide whether a deployment pattern fits a team’s capabilities. Speakers can explain where their advice stops applying. Attendees can hear how the same principle behaves under different constraints.
Put a meetup for people following the same track on the schedule. Give them a room, a time slot and a host who helps conversations begin. Offer prompts tied to the learning outcomes: which decision are you facing, what constraint keeps getting overlooked, and what have you already tried?
Make these gatherings visible when people book. Let attendees choose a track meetup and bring a question. Give them something specific to look forward to.
The manager can now assess a defined opportunity: learning about architecture, questioning practitioners about relevant constraints, and comparing approaches with peers. The organiser has made the interaction visible and given it a professional purpose.
That still leaves a difficult judgement about whether the opportunity justifies the travel. A timetable and learning hours cannot capture the quality of an answer or the value of a relationship. Calling a meetup “professional development” does not automatically make it worthwhile.
A brilliant recorded talk may teach more than a mediocre live session. Equally, a short conversation may reveal why an excellent tutorial is wrong for your circumstances. Quality matters in both formats, but comparing their value requires more than counting content hours.
Ask attendees which assumptions changed, what decisions became clearer and what they applied afterwards. These provide evidence worth discussing, even when a financial return cannot be isolated. YouTube remains the comparison; the programme must explain what being there makes possible.
The interaction needs boundaries, too.
An AMA should have an agreed duration and scope. Speakers need to know what they are committing to, with that contribution recognised in their support. Nobody should discover that accepting a speaking slot also signed them up for unlimited consulting.
Attendees need choices. Some will contribute immediately; others will learn by listening. Allow written questions, smaller conversations and quiet breaks. A structured meetup should make joining easier, without turning professional development into compulsory social performance.
Attendance Must Deliver Value. Participation Should Deepen It. #
I don’t pretend to hold The Answer(tm) to the perfect event. But I think we could benefit from stepping things up a notch. Just like AI forces new styles and mechanisms for teaching, the availability of content (be it good or bad) should prompt a review of the way we are doing events.
Start small: one curated track, one supported new session, one AMA and one hosted meetup. Ask afterwards whether the discussions helped people apply the material and whether speakers had enough support. Use those answers to shape the next programme.
Keep pairing new speakers with mentors and help them develop sessions for the track. The event should develop its future contributors.
Publish selected resources with speaker agreement and support for editing. The wider community should benefit too, but there needs to be a balance.
I want a conference where the sessions give us something worth discussing, the programme helps us find the people to discuss it with, and the budget recognises the work behind both.
The hallway track deserves a room, a purpose and a place in the business case.
Image by Anastasia Borisova from Pixabay