Last updated on

My Experience Speaking at the dbt Coalesce Conference, and the Easy-to-Steal Prep Process Behind It

A few weeks ago, I spoke at the dbt Coalesce data conference. It was a super rewarding experience, so I thought I’d write about it here.

Why speak at a conference

There’s plenty of reasons you should speak at conferences, meetups, anything in public. I’m not going to get into most of them here, there are plenty of articles that have done it. Why did I want to speak at a conference? Mainly, to establish credibility within my industry and get more public speaking reps.

What is dbt Coalesce and why did I want to speak there?

Credit: dbt Coalesce website

dbt Coalesce is an annual data conference from the folks at dbt Labs. Many talks cov analytics engineering (the niche of data I most recently worked in) and a suite of mainly cloud-hosted data tools dubbed the Modern Data Stack.

I attended a few of the previous conferences virtually and was very impressed with the caliber of speakers and presentations. Honestly, the talks that disappointed me were often by folks whose writing I enjoyed so much that my expectations of their speaking were too high. And both years I thought, hey I’d love to present one year if I have a good presentation idea.

I’m glad that I achieved that goal in 2022. Here’s a behind-the-scenes look into some of the steps I took to make it happen.

Brainstorm ideas

Five months before the conference

Most conferences put out what’s called a call for speakers, in which they invite people interested in speaking to apply. This can be a bit intimidating if you haven’t done it before, but I figured there wasn’t much harm in applying. I didn’t have an idea ready, though.

One of the conference organizers, Jillian, posted in the dbt Slack community that she was hosting office hours in May for anyone who wanted help in generating or refining pitch ideas (it’s awesome that Jillian and the dbt Labs team did this and more conferences should do it!). I took her up on that and booked some time. Before the call, I started a Google Doc and wrote out three ideas and why the audience would find each interesting.

At this point, I wasn’t even sure if I was going to apply to speak – I told Jillian as much and said that at the very least, I’d be well prepared to apply in 2023. From there, we dove into the ideas.

Jillian reacted positively for what I considered the strongest idea, which was how we evolved our workflows as the data team at Dutchie grew from just me to 14 people. But she shared way more enthusiasm about the second idea, which was our learnings from the work I did in the wake of not one, but two company acquisitions in 2021. My third idea was showcasing some of the continuous integration automations we set up – Jillian was receptive but warned there were competing pitches for this topic.

Backing up to the second topic, why did Jillian think that was the best bet? The first reason was that the conference hadn’t covered this topic in prior years. Mergers and acquisitions (M&A) are rare, and so only a fraction of the potential speaker pool can credibly speak to them. Second, with tech worrying of a market correction, we could expect to see more mergers and acquisitions than before as companies’ cash reserves draw thin.

Pitch the talk

Five months before the conference

Jillian said I was welcome to submit both my first two ideas, but knowing the M&A talk had a much greater chance at being selected, I put all my effort into submitting a strong pitch for that topic.

More or less, there were three questions I needed to answer with my pitch.

  1. Why is my topic something that would be valuable to the conference audience?
  2. What credentials do I have that enable me to speak credibly on the subject?
  3. Am I able to be an engaging public speaker to actually deliver the message effectively?

Each pitch needed a short Loom video of you describing your topic and why the audience would like it. Mine was 2-3 minutes long and screenshared a two-slide deck with themes I expected to cover. I didn’t overthink it, and did it in a single take after a practice run.

The Loom video addressed the first and third questions. I also wrote up a short bio listing my recent job experience and education to establish my credibility.

And then I waited. A few weeks later, I found out the good news that the conference accepted my talk!

Recruit copresenters

Four months before the conference

For most people, this step will usually come during the ideation stage. Most conferences will require copresenters to be attached to the pitch when you submit it. I was fortunate that the dbt Labs team allowed me to tack on someone over the summer. While I tried to convince my teammates to join the pitch in May, they basically said “well, maybe. Let me know if the talk gets accepted and I’ll let you know then.”

Once the talk was accepted, I asked again, and gave them a few weeks to think it over. Cameron decided to join, while Oscar was too busy to prep with us but offered to moderate our chat on the day of the conference.

There were three reasons I asked Cameron and Oscar to join the presentation.

  1. Their perspectives on the project, and high quality of their work in general would make the presentation better.
  2. I could have spoken for the entire 30 minutes, but it would be a lot easier to split up the talking.
  3. The credibility and public speaking reps I spoke of earlier as motivation would be good for them, too.

Outline the presentation

Six weeks before the conference, four weeks before first draft of slides due

The first thing Cameron and I had to do was outline. What messages did we really want to hit home, what stories we wanted and didn’t want to tell, and how we should structure the content. We didn’t overthink this one, we just started listing stuff in a Google Doc for two 30-minute sessions. Looking back, we could have gone a little deeper here.

Draft the slides

One month before the conference, two weeks before first draft of slides due

From there, we created a skeleton slide deck and added a slide for every main bullet point in our outline. I’m a visual person, so being able to actually see how much could fit on a slide was super helpful towards thinking about structure. By doing so, we made a few pivots in how we expected the average slide to look. However, doing this also meant that we started to stray from our outline – something we got called out on later.

We started to meet more frequently (about 3x a week) and now for an hour at a time, refining the slides until we felt that we had a solid foundation.

Get feedback on slides

Three weeks before the conference, two weeks before final slides due

When you’re been working on a project alone or in a small group, it’s usually a good idea to get some additional eyes on it. Why? Sometimes you get too close to the problem, and another person might instantly notice something you missed.

So we jumped at the opportunity for members of the conference team to advise us as we were working on our presentation. We ended up meeting three times with Jillian and Mark, who worked in Ops and IT and had his share of experiences with M&A.

In the first meeting, the feedback was mostly constructive but delivered in an effective manner. Shortly put, we weren’t focusing enough on our thesis. Jillian at one point asked to see our outline, and Cameron and I nervously pointed out that our outline had become stale in the process of drafting our slides. Mark also made a few great points about specific slides where we should really emphasize tying the content to the thesis.

As Cameron and I debriefed afterwards, we felt the neurons firing – most of our individual slides had fine content, but we never actually declared a thesis, so the presentation felt like it was missing a piece. We brainstormed a quick list of options, and eventually went with the simplest option – how would we summarize our takeaways from the slides in a sentence?

In our case, the thesis became “The problems you tend to encounter in M&A-related data projects are going to be similar to the problems you encounter in everyday projects, just much more severe”.

Once we had the thesis, everything else came together. We met again with Jillian and Mark less than a week after the first meeting and their feedback was night and day much more positive. From there, we keyed in on a few remaining areas for improvement before submitting final slides a week before the conference. During that time, we met everyday after work for an hour, sometimes two.

Practice the presentation

Two weeks until the presentation, one week until final slides due

Honestly, I was surprised that the first time we did a full practice run ad-libbing the presentation, we sounded great. That was a benefit of knowing our topic inside and out. However, we expected to have 27 minutes to deliver the talk at the conference, and our practices were pushing 38 minutes.

Fortunately, just practicing more shaved several minutes off as we became more familiar with what we were trying to say, and pausing less often. I also eventually had to give in and create a script I (mostly) memorized so that I could more reliably hit my time target. Cameron, to his credit, was able to keep ad-libbing but at an increasingly consistent cadence.

Like with the previous section, we also wanted to make sure that our verbal presentation resonated with folks. So we leveraged our networks and rehearsed in front of a few teammates from the Future and Dutchie organizations. To our surprise (and delight), the feedback was almost universally positive with a few small critiques and suggestions. Thanks again to everyone for your help with the finishing touches!

One thing Cameron and I did in the final days before the conference that really helped was to actually practice on Streamyard, a livestreaming tool that we’d be using on the day of. Most of our presentations the last few years have been on tools like Zoom or Google Meet, so having cameras, screenshares, and chats all in different places on the screen was a small adjustment that we were prepared for on conference day.

All in all, we probably combined for about 10 hours of rehearsal time in the final weeks and we almost knew our presentation by heart. Our timing had become more and more efficient and predictable, to the point we were actually a minute below the allotted amount on the last few practices.

Deliver the presentation

By this time, the last thing to do was actually deliver the presentation. It was largely uneventful, which is a good thing. We paced ourselves well, and I graded my performance as maybe 90% of my best practice runs, which I’ll take. I like to slightly overprepare for presentations in case nerves end up getting to me (sometimes they do, sometimes they don’t), and that way I’ll still be satisfied with the outcome. Here’s the link to the presentation, if you’d like to see it for yourself.

dbt Coalesce switched to a Slack-based Q&A this year in part due to the hybrid structure, so afterwards Cameron and I hopped on a Zoom call and started answering questions on Slack.


Overall, I’m pleased with how the experience of preparing a talk and presenting at dbt Coalesce turned out. Our slides looked good, the audience seemed to enjoy what we had to say, and Cameron and I had complementary skillsets during the prep process and worked well together.

It was a positive experience speaking and I definitely recommend others considering speaking at a conference, whether dbt Coalesce or another one, go for it. The preparation process is repeatable, you’ll have folks in your network willing to help you out, and by budgeting your time you can make the prep fit to your schedule.