Velocity Conference 2017

Just last week I spoke at and attended the O'Reilly Velocity Conference in San Jose. My talk was titled Scaling a User Delivery Network for Real Time Audience Targeting, and you can find a link to the presentation page and the presentation itself. It was an interesting experience, to say the least, and while I'd recommend it to anyone (including my future self), there are a few things I learned from this one.


Have a solid idea in mind, and a clean, tight proposal. Make sure to read the track description and understand what the conference chairs are looking for. If you can't convey your idea well, you won't clear the first hurdle. Feel free to submit proposals to multiple tracks (I did), but know you'll likely only get accepted to one. My second proposal was for the Technical Decision-Making Track, and would have covered several lessons I learned in making the transitions from engineer to technical program manager to architect. But in retrospect and after seeing the presentations that were chosen, I realize my pitch was a bit chaotic and vague. Had I honed it down, and focused on a few key points like I did for my accepted talk, it might have been more competitive.


Everyone will have their own styles and methods, but I can talk about a few things that worked and didn't work for me.

I started out with a plan to have my slides/speech done by the end of April, use the month of May to get feedback and hone the presentation, and then use the month of June to practice practice practice. In reality (for a variety of reasons), it took me until June to finish my slides, get feedback, and practice, and I was even up the night before my talk further refining it.

Given that my presentation was supposed to cover a relatively linear storyline, I wanted to make sure to get the whole narrative out so I decided to write the speech before doing the slides. I started with an outline of the overall story, then slowly refined it into an essay of about 8,000 words. I thought this would be beneficial, because I kept flipping back and forth between my outline and my slides, and was having trouble keeping the narrative and transitions on track.

What I didn't realize until very late is that essays don't make good presentations. Unless you're doing a performance of some sort (e.g. a play or a magic act), the vast majority of a listener's attention will be focused on your slides. By the time I realized I needed to make changes to my content, I had grown quite attached to my essay, and had trouble making the requisite changes.

The other thought I had was that in the absence of having a lot of practice, instead of having vague bullet points and filling in with a lot of "ums" and "ahs," at the very least, having a well-written speech would give me something to fall back on in my presenter notes if I got nervous. It did turn into a crutch, however, during both practice and delivery.

I spent a good amount of time on my slides, and thought I hit a good mix of high-impact picture-based slides and bulleted slides, though I probably could have done a better job reducing the number of words on the bulleted ones.

Practice and Feedback

Give your presentation to as many people as possible before the day of. I gave the presentation to myself a couple times just to practice timings and navigation. I also gave the presentation to my wife and another close friend, and their feedback was invaluable. From timing (too long) and material density (too dense), to yawn factor (too high), to connection with audience (too low), I had a lot of work to do. Realistically, if you get some feedback from one person, you might be able to discount it, but if you get the same feedback from multiple people, you should probably re-evaluate.

I cut down quite a bit of unnecessary and overly dense material. While the talk was supposed to be technical, I was adding details that weren't critical to getting the main point(s) across. I also spent way too much time introducing some core concepts that weren't really necessary for the audience to understand later topics. After the cuts, I went from not hitting the core topic until 30 minutes in (out of a session length of 40 minutes), and running about 42 minutes, to hitting the main point at 15 minutes, and running about 35 minutes - a much better ratio.

I also added in an upfront "roadmap" of the talk, to let audiences know what to expect, and modified my "topic introduction" slides to connect directly to the roadmap bullets to bring listeners back.

Finally, whereas I thought some of my architectural diagrams were reasonably self-explanatory (a picture's worth a thousand words, right?) and mostly breezed past them to save time for other discussion, I switched that around, to make sure to walk through diagrams and explain all of the boxes and arrows, and breezed past some of the other, less-useful discussion.


Make sure to arrive early and scope out the room and set-up you'll have. Hopefully the conference organizers have given you plenty of information, but it's also helpful to bring your own adapters, power supply, etc. If you are scheduled at the right time (like right after lunch or a break), you may even have a chance to set up and do a full run through (even if quietly/silently to yourself) to make sure that everything works.

At the end of it all, just relax and give your talk the best way you know how. Audiences are there to see you, and want you to succeed, so they're far more forgiving than you fear.

Some Lessons Learned

As I went through the process, I started formulating a categorization of presentations:

  1. The TED talk - the uber-prepared, highly-polished, every-detail-memorized-presentation. This person has spent months preparing, and weeks (or longer) rehearsing until the prepared script no longer sounds rehearsed. He or she likely has had professional help and consultation, and every word has been pruned and honed ruthlessly until the performance approaches just that - a performance.
  2. The easy, fluent, conversational presentation. This person likely has a strong grasp on the subject matter, and has spent plenty of time preparing. They may or may not have notes or notecards, but have no trouble getting through the presentation. This is probably the largest category, and has a large continuum of styles and presentation skills.
  3. The halting, awkward, or otherwise bad presentation - we've all seen one of these.

My realization was that while I dislike fillers ("um", "ah", and the like) or incoherent narratives in presentations, and thus was attempting to hit "TED talk" status, I had taken nowhere near enough time to have perfected a 30-40 minute talk. But on the other hand, I had grown reliant on my "essay"/script, and was nervous to stray too far from it. While I'm comfortable talking extemporaneously about online advertising, the pressure of "producing" a talk for an audience pushed in a deleterious direction.

I harkened back to the various training presentations I had given at my former employer, some of which ran for over 90 minutes, and realized a similar pattern had emerged. The very first training session I had given I had scripted out in fairly extreme detail. As I gave the talk multiple times, I grew more comfortable with it, and could refer to the script less, make more connection with the audience, etc. When I revamped the presentation, I actually completely threw out the old material, started from scratch, and made the slides first, then followed it up with just a few notes on each slide. Nowhere was any script to be found. And while I had learned the proper delivery of the first version, the second version was tighter, more cohesive, and more relevant for audiences. With a third version I gave, I again started from scratch, and this time didn't even have notes at all.

I almost think for my next talk, I might actually use this strategy of creating and delivering a talk as a "first draft," then throwing it all out and starting from scratch to get to the real meat of the presentation.

I don't have a strong conclusion to this, as it was mainly just a place to gather a few thoughts after the conference, but perhaps that's another lesson to learn - finish strong so audiences aren't left wondering whether you're done!