Accento 2019 was the first conference we (that's Uta and Nicolai) ever organized and so we gave ourselves some wriggle room. Of course we strived to make it the best event we could but our top priority was expediency - when you don't know what exactly lies ahead, it's better to achieve 90% now than 100% when it may be too late. Quick, but not dirty, so to speak.
2020 will be different. We'll have fewer fundamentals to decide and put in place (think conference name, domains, social media accounts, ...) and we'll have more time to work on the event. Most importantly, we'll have more experience, and this is where your feedback and this retrospective comes in.
We asked everyone at the conference - from attendees to speakers/trainers, from personnel to Disy's developers - what they thought of the event and where they saw room for improvement. Then we took that input and the Jira issues we created over the last eight months or so and locked ourselves into a room. For two afternoons, we discussed everything we did and everything you told us. (That means from here on out "feedback" also includes our own opinion, particularly on behind-the-scenes stuff.)
We used large sticky notes to persist the collective impression:
- clusters like location, program, marketing, etc. in pink
- attached items like check-in, talks, website, etc. in either green, yellow, or orange, depending on how happy you and we were with the item
Going through all the items would be excruciating, so we'll spare you some of the details and discuss our take-aways from the retrospective as well as the implications for 2020 on the level of clusters, focusing on the items that need the most work. Not to get too negative, we'll also present a thing or two that you were really happy with (which made us really happy with it, too).
We want to point out that we're still interested in feedback. If you didn't tell us anything earlier, have additional thoughts because of what follows, or want to outright comment on it, please do! You can send a mail or tweet at us.
Overall, the Südwerk was very well received. Layout, catering, our setup/tear-down schedule, check-in - all of that got mostly to exclusively positive feedback and so we already settled on going back there in 2020. That said, let's discuss the few details where opinions were mixed and where we could improve before looking at two of the many aspects that we'll gladly repeat next year.
We got conflicting feedback on the location's size: Some attendees said it felt a little empty, with the location allowing way more people. Others said it was just right and more attendees would make it too crowded.
Personally, we fall into the first camp and since we're hoping to grow the event, we assume that we'll see more crowd and less empty space next year.
One negative aspect was the comparatively high main stage, which is a little taxing on the neck if you're sitting in the first few rows.
Unfortunately, the stage is part of the location and isn't height-adjustable, so we can't do much there. Putting a greater distance between the stage and the first row of chairs may be an option, but it has two downsides:
- as a speaker, it's a little weird to have a pit between yourself and the audience
- assuming higher attendee numbers, we'll already struggle to get everybody into the curtained area
Not sure how we will handle this in 2020.
Then there's the length of the smaller room, which made it tough to see everything that's going on in the front.
One remedy may be to get a higher stand for the screen and a lectern for the speaker. Another one would be putting fewer rows of chairs into the room, so we don't entice attendees to sit too far back, but of course this means that fewer people can attend that track. Whatever we decide will have twice the impact because we hope to have three tracks next year, with the third in a room with the same layout (it's directly above the one we're talking about here).
The only outright negative item we attached to the location was the screen/sound recording setup. While getting sound (on the main stage) and slides (both stages) from speaker to projector/screen was simple and worked like a charm, recording that for later was complex and fragile. The pipeline was entirely self-built with HDMI splitters, capture cards, OBS, and run-of-the-mill laptops. For our own nerves' sake, we have to find a better solution for 2020.
Coming to the positive aspects, let's discuss the Südwerk itself. It's a large dancing hall with a beautiful wooden floor, high ceilings, and a long glass front that lets in a ton of light. That's a very elegant venue to start with but when it's empty, it's a little barren. So we filled it with about half a dozen sofas, organized in three seating corners, two high tables to get some work done, some plants and sponsor roll-ups to add a little color, and of course our Mario Kart corner with a table of free books.
We liked it a lot and so did you (which, arguably, is the more important part) and we're happy to further improve on that in 2020.
Good food is important for mind and body, but at conferences it's hit and miss. Some have amazing meals, others offer tasteless fluff that needs a label to let you know what it's supposed to be.
We were lucky and the Südwerk prepared great food! Not fancy, just really well-made and very fresh. Everbody dug in and quite a number of people mentioned to us how much they liked it. It may seem minor, but great food turns into great mood.
Just like the location, the program got very good feedback. That particularly includes talks, trainings, speakers, and scheduling as well as the discussion of climate change and the afterparty. Once again, let's start by looking at where to do things better in 2020.
First and foremost, the lightning talks sucked. Well, not the talks themselves, they were great, but the way I scheduled them sucked. My idea was that with Q&A participants missing the start of the next talk (more on that in a sec), it would be nice to give them something else to do if they don't want to barge in. But that didn't work well and many people told us that they would have liked to attend some of the lightning talks but didn't want to miss out on an hour of content for just 10 minutes.
We get that and we'll do better next year - pinky promise. One option would be to fill an entire slot (or more?) with lightning talks, maybe even without other talks to attend in parallel. Another option is to have them during breaks - kinda like brown bag lunches. We'll see what exactly we end up doing, but it will be different from this year.
A pretty contentious subject were the separate Q&A rounds. They have several upsides and downsides:
- gives speakers a full time slot for their content
- avoids the "hostage situation" at the end of the talks where some people have questions but most want to leave but feel like they shouldn't
- allows ample discussion
- allows for short breaks between talks (i.e. more talks)
- breaks continuity from talk to Q&A
- makes attendees choose between Q&A and the next talk
- requires a dedicated place with some space, power, and screen per track
- requires conference personnel to ferry speakers and attendees with questions to said special place
Consequently, the feedback was mixed, although when tallied up it leans towards negative. We're really not sure what to do here - the upsides are very relevant and make the conference better, but the downsides suck. Going into 2020, this is a big question mark...
Another unsure thing for 2020 is the topic. Accento 2019 had Quick and Dirty (attendees liked it) and Ted's keynote on it was great (liked that, too), but not many talks picked it up (didn't like that).
For 2020, we consider coming up with a topic and then see during the call for paper (yes, we'll have a real CfP next year - see below) whether enough talks connect to it. If yes, great; if not, no topic.
The focus on Java and web dev was well-received and we'll keep it (that's what Accento is all about!), but we may add another approach to these topics. This year, it was mostly development, but there’s also operations and we'd like to offer content in that area as well. We think, in 2020, topics like "Optimizing SQL queries", "Newest Elastic Search features", or "Something, something, Docker" will get more coverage.
Final point on the the program: Nicolai wanted a panel. We had great people for a panel! But there was no panel!! Maybe, in 2020, there will be a panel.
We had two tracks and that was ok, but we think we need more. So, in all likelihood, we'll have three tracks in 2020. That should give us enough slots to dedicate some of them to lightning talks, panels, and more varied content, making the conference more well-rounded.
This year, all you could do if you weren't in a talk was play Mario Kart. Don't get us wrong, that was really cool! Many people played a round or two and had a good time. It was also a perfect fit for our topic. (Kudos to diva-e for coming up with the idea!) But it was the only thing we offered.
In 2020 we want to have more nuggets like that: more fun things to do in the breaks, to do with other attendees, to do while you slack off between talks to get a little rest.
The most important part of a conference is that content and speakers are top-notch and we really struck gold here. We see Accento as an international conference in Karlsruhe (hence all the English) and we are very proud that we had a few international rock star speakers there with Ted Neward and Simon Ritter leading the pack. Our German speakers were just as amazing with the likes of Marc Philipp and Peter Kröner who are internationally recognized experts in their fields.
The feedback on our speakers was overwhelmingly positive and we're happy to not only repeat that in 2020, but try to improve it even further. Not sure we can pull that off? Only one way to find out.
Selection, organization, pricing - according to you we did everything right. The feedback for the three courses can be summarized as good, great, and excellent, respectively, and we're happy with that.
Speakers and Trainers
We already covered that talks and trainings went over really well and so this cluster doesn't deal with that. Instead, it looks at the organizational side of getting clever people on stage and from this perspective, while everything went well, there's a lot of room for improvement.
The most critical aspect is representation: We had zero non-white, non-male speakers at Accento. For a first-time conference that has to put a lineup together in about two months for an event that's just two more months in the future that may be just barely acceptable. And that our only female speaker got stuck in traffic didn't help either, but we're still very unhappy with the result.
For 2020, we want to and we need to do better. One way to do that should be to reach out earlier.
Call for Paper
Another aspect we need to professionalize is the call for paper. This year, we didn't have one - instead we just reached out to fellow speakers and asked them whether they would like to come and present. If the answer was yes, they were in (we already knew their talks would be interesting - that's why we contacted them in the first place) and all that remained was to discuss the content. While that worked out pretty well, it was a painfully manual process that is really hard to parallelize for multiple reviewers. The lack of a deadline also meant that we dragged this out way too long.
Next year, we'll have a proper call for paper with a real deadline and a couple of people vetting the submissions - the process is gonna be way more structured (read "way better").
A positive aspect of the manual invitation process and of the fact that only one person took care of all communication with speakers was that it was very personal. Speakers told us that they enjoyed the direct communication and that they felt very well taken care of.
As we look to slowly grow the Accento team in 2020 we have to keep an eye on this and make sure speakers enjoy 2020 just as much as they did 2019.
Marketing and Sales
Starting a conference from zero isn't easy and while we worked hard to reach as many developers as possible, we want to do better next year. Here, we'll discuss our efforts to reach you folks, so you know that you need to come.
An underlying problem that hamstrung a lot of our marketing and sales efforts was timing. It took us until end of April to have any online presence, two more months to have a decent web site with a sizeable chunk of our program, and another month to settle the last talks and trainings.
So, all in all, we were pretty late. Organizations tend to plan ahead and even developers do that occasionally and with many we felt (and were occasionally told) that they were very interested in Accento, but already had conflicting plans.
Hence one of our core missions for 2020 is to have better timing, which means setting up a timetable for the entire year with what needs to happen when, so consequent steps have enough time to take effect. In general, we want to be earlier with everything.
Making a Splash
To get people to come to the conference, we need to (a) let them know that we exist and (b) convince them that it's worth coming. We think we covered (b) pretty well with an enticing program and a good website, but we could've done better with (a).
One way to make more of a splash beforehand is to create content together with our speakers. Picking two, three interesting details out of each talk and turning that into a short YouTube video or blog post would be an interesting and informative watch/read while letting people know that they can learn more about the topic at this thing called "Accento". You can't have speakers doing that for you (marketing is our job, after all) and we hope to be able to do that with them in 2020. It will help the conference, but it will also be a lot of fun to work on something with the amazing experts we have on stage.
You can't talk about making a splash without mentioning social media. In 2019, our strategy was "Nicolai tweets about Accento about once per day". Hm... Maybe we need to do better in 2020. Having genuinely interesting content (not just the promise of it later) would really help a lot in this regard.
More channels to make a conference known are event calendars (usually digital) as well as newspapers (the printed form), online magazines, and so forth. We did some of this and it feels very low yield, but we think we're gonna do it again in 2020. But better! There are tons of developers in the region and we're pretty sure most of them yet need to hear about Accento.
And whatever we do, having 2019 in the back pocket will help a lot. After all, we already did it once and nobody died of boredom or food poisoning - in fact people were pretty happy and we'll also have a few videos to show. That should make it much easier for potential attendees to form an opinion, which will make it less of a leap for them to attend. And who knows, maybe some of you spread the word about Accento and how you get no food poisoning there.
Tickets and Prices
All feedback we got on the ticket prices (conference: 150 €; training: 350 €; combi: 400 €) was positive and general opinion was that price/performance was very good. We’re happy to hear that, but there are still a few things we can do better.
While full-price tickets worked well, some others didn't:
- The student discount of 50 % wasn't enough - even 75 € are a lot to many students. Next year, we're going to make them considerably cheaper.
- Early bird tickets tanked because, at that time, only Disy and our mums knew what Accento was. We hope we can sell more discounted tickets early in 2020.
- We didn't have a late bird with a higher ticket price for the last week or so but maybe we should have. We sold a bunch of tickets in the last few days, some literally the evening before the event, which is very stressful because a lot of details depend on the number of attendees (chairs, catering, name tags, etc.). We know for a fact, though, that most late birds planned to come for a while and we think that, if from the get-go we announce a price hike for the last days, we can incentivize them to buy earlier.
- The other end of the unlucky middle ground between having to plan ahead and giving you time to consider your purchase was that we didn't sell tickets on the day of the conference itself. So we're thinking about letting you buy tickets on the same day if you forgot to do it earlier.
By the way, using Eventbrite as a ticketing platform went pretty well. They take a 5% cut, but that's fair for the service they provide and we'll use it again next year.
Website vs App
The website got very positive feedback and Nicolai is very proud of that because he put it together himself (not the design - everything user-facing was done by someone competent). We didn't plan it that way, but it turned out to also work reasonably well as a channel to communicate last-minute changes.
But don't we want to have an app? Everybody has one! As we see it, there are two advantages to a conference app over a website:
- easy notifications to let attendees know of changes
- easy way to collect feedback, particularly on the talks
On the other hand, who really wants to install an app for a one-day event? And let's not even talk about the effort that goes into developing one that could be spend elsewhere to much greater effect. No, we won't have an Accento app in the near future.
Karlsruhe has a great development community with lots of cool companies and we really wanted to get them on board for Accento, so we asked them to partner with us. This went pretty well: We were happy that partners helped us spread the word, particularly internally, which enticed many of their developers to attend the conference, and they were happy being associated with an event that went over really well. Still, we feel this aspect has a lot of room for improvement.
This year, we had to be quick with pretty much everything and so we made all organizational decisions ourselves. We knew that in advance and never offered any involvement, but in 2020 we'd like to change that. Some obvious places where we can easily involve partners would be talk selection, side program, and exhibition area (see below). We plan to think long and hard about where outside input or support would be useful and discuss that with this year's partners to hear their thoughts before finalizing 2020's partner program.
It happened more than once that partners asked whether they could do something at Accento and we thought "well, that's a good idea, why didn't we have it?". We usually approved the request, which means all partners could do the thing now. The problem with that is that it's not great for partners if the ground keeps shifting underneath them. Next year, we'd like to give them a definitive list of what we offer as part of the partnership.
We will also take that opportunity to more clearly delineate different levels of support.
Many larger conferences have an exhibition area, where each partner/sponsor gets a booth. Often times, they are just glorified flyer dispensers, though, and - quite frankly - nobody cares about that. Instead, booths should have something that makes them interesting to the attendees: a coding task, an intriguing piece of hardware, an eye-catching demo... But that's no easy feat and the investment only really makes sense if there are a few more attendees. Hence we decided against an exhibition area.
We're not sure whether that will change in 2020 and it's likely to be one of the things we'll discuss with this year's partners before making a decision. There will only be booths if we think they make the event better for attendees.
Out of expediency, we only approached companies from Karlsruhe to become partners. In 2020, we want to reach out further to other companies in the region. And not just companies either, we'd like to involve the community as well. Like with attendees, we're sure that having a successful Accento 2019 to show off will help us a lot in this regard.
Now we're going meta. This is about how we handled the organization of the event, once again focusing on what we plan to change for 2020.
Evaluating our roadmap is straightforward: We didn't have one and in some cases that tripped us up (see Timing above). We don't consider that a mistake, though - without knowing what lies ahead and how exactly tasks connect to one another, investing into a detailed roadmap is a silly waste of time. Instead, for each task at hand, we wondered by what time we had to complete it and tried to make sure to achieve that by starting to work on it a little earlier. As you can tell by how well the conference went, that worked, but at the same time that occasionally resulted in stress and missed opportunities.
The situation is different in 2020: Now we know very well what we want to do, how tasks relate, and whether the times at which we did them were ok or not. So one big thing for next year is to start with putting together a roadmap, basically by taking the tasks we did for 2019 and putting them onto a timeline that starts in November 2019 and ends in September 2020. We have to admit, we're really looking forward to seeing all of Accento in one big diagram.
Buzzwords for 2020
This year's Accento went really well and we're very happy with that. As you can see, there are tons of details to improve, though, and we're full of energy to tackle them. To still see the forest for the trees, we derived a few high-level focus points for 2020.
We want to move from just in time to timely, making sure we complete each step at the exact time we consider best. As pointed out above, the roadmap we'll create will play a central role in that.
Division of Tasks
With a third and maybe fourth person on board, we want to make sure everybody can focus on the tasks they do best while keeping communication lean and efficient.
Nicolai is pining (and occasionally whining) for a third track and he will likely get one. Hence, one determining aspect of Accento 2020 will be organizing more talks, more speakers, and more content beforehand.
We want to reach more developers in Karlsruhe and more from in the region, maybe reaching as far as the bordering regions in France and Switzerland. And we hope that everybody who was there in 2019 enjoyed themselves enough to come back in 2020. See you there!