Welcome to the inaugural post of the Vendelux data and developer blog!
We’re excited to open a window into the technology and innovation that fuels our work here.
I’m a strong believer that innovation and speed are what drives a startup and there’s no better way to get the juices flowing than some dedicated time spent hacking on random ideas to see where they take us.
And since Vendelux is a data company – we’re an AI-powered event intelligence platform that provides event organizers and marketers with data-driven insights and tools to optimize their event marketing budgets and ROI – what better place to start hacking then with our data?
So a couple weeks ago, we held our first (of many) DataFest.
What is DataFest?
DataFest was an in-person, week-long data hackathon where we, the data (ops/engineering/science/analytics) team spent time together, working with folks we might not normally work with, hanging out in the NYC office and beyond, and having fun building out prototypes of data-related projects.
Projects and teams were selected ahead of time, where the idea was to do something that is innovative, fun, maybe work with some different people, but also contribute something helpful to the company. DataFest also marked the first time our rapidly growing team was able to all get together for face-to-face collaboration.
As a hybrid team where I myself am remote 90% of the time, some amount of in-person time is invaluable.
Nothing beats the energy that feeds off each other or the ability to pop over and get a quick piece of feedback or advice.
Don’t get me wrong, if we were always in-person, those quick check-ins could also be distracting, but for a week – it’s amazing!
What did we accomplish?
Over the course of the event, the team tackled big questions, built some impactful tools, and explored new ways to leverage data for the company.
We played with LLMs for better event understanding, embedding models for similarity search, analyzed sponsor churn, and tried to automate some overly manual tasks.
We also spent time together outside of work, hitting some indoor golf, making fettucini and ravioli and spending happy hour together.
And then, the best part – we demoed everything we built during the week. The expectation was not perfect production code, just a prototype. Overall goals were mostly around fun, morale and camaraderie, but I also believe it’s possible to get something of value for the company out of it as well.
The demos showed that we were able to build out enough of a POC that informed us whether a project is worth pursuing, and how much effort it would take to get to a fully productionized solution.
And the energy the team felt and also created, was amazing. You could feel everyone else’s excitement about the things we were building. They could not believe all of the amazing things that we built in such a short time and were eager to know when these huge developments would be in the platform.
And that energy was contagious, building upon itself and reinvigorating the team to keep building and push the whole company further.
What did we learn?
First, we learned that we should definitely do this again. At least once a year, maybe twice. We also showed that we have a really talented team that when motivated and given a problem to solve, we can come up with quick and innovative solutions.
These are things we did that enabled our success:
- Choose projects and do a little bit of planning ahead of time (i.e. get your datasets in order so you can dive right in)
- Do enough planning and logistics work so the team feels like things are taken care of and they’re not worrying about where they’re staying, where their next meal is coming from, etc.
- Have some unstructured time. Some people are drained by having to be “always on” and spending time with others, so some downtime or time where we’re not expected to be together is worth it as well.
- Presentations!!! People may push back – they take time, it can be uncomfortable for some – but having a demo be a part of the week ensures people are thinking about the “why” they’re building something and that they consider how impactful their solution can be.
Finally, these are things we didn’t do or lessons we learned that we can improve on next time:
- More cross functional and collaborative – we could have involved other teams more and collaborated more on projects by bringing customer success, sales or app development teams into the fold and building cross functional teams to solve some of the problems.
- Voting, winners and prizes – as much as we were all winners that week, it would have been exciting to see a vote of what the entire company thought was the most impact, was the coolest tech, the best presentation, etc.
Bottom line, it was an amazing week – we laughed, we cried, we data’ed and we set ourselves up for some amazing data improvements to the Vendelux platform.
But… Tell me about the PROJECTS!
Ok, but what about the actual projects? Yes, that’s the real goods, so I’ll give a brief description of each of the cool things we did and why we did it.
- Speaker/Sponsor Extractor
- What: Automate the extraction of speakers and sponsors from public sources
- Why: Speeds up our ability to capture this data, meaning data on our platform is more up to date and accurate, more cost effective and more efficient
- Event Similarity/Recommendations (link to more specifics about this one)
- What: Come up with a method to recommend events to our customers – if you attended this one, check out that one
- Why: As an event platform, knowing which events to attend based on factors like ROI and the persona of attendees is invaluable. If a customer has had success at one, knowing where else they can duplicate that success is a huge win.
- Sponsorship analytics
- What: What percentage of sponsors renew their sponsorship for this event year-over-year?. How does that compare to other events in the same space?
- Why: This data is important for both organizers and sponsors. Organizers want this number to be high and if it’s not, it essentially shows their “churn” compared to comparable events. Sponsors want to know where to put their sponsorship money. If others are leaving this event, it likely isn’t a good event to sponsor anymore and your money will be better spent elsewhere.
- “Event Vibe” LLMs
- What: Using data from social media, what is this event like? How formal is it? How are the speakers? Essentially, give a qualitative summary of the event.
- Why: Again, the decision to attend or sponsor an event is a big and expensive one. The more information you have about it and whether it makes sense to spend your money there the better you will be able to get a good return on your investment.
- Ops team – Process Improvements for DataOps workflows
- What: Dashboards and automations around particular workflows
- Why: Any operational improvements have a huge impact on the ops team, which in turn improves the quality, quantity and cost-effectiveness of our data collection and validation efforts.
Have other opinions or thoughts to share? How have you done a hackathon/ tech-fest/datafest? Let me know in the comments below!
