To close out Q2, we hosted an in-house hack-a-thon as an opportunity to create without constraints and validate ideas. For three days, our engineers broke into teams and began work based on a specific hypothesis. The week culminated in a formal demo—and an internal vote to declare a winner. 

In the end, participants unanimously agreed they would do Innovation Week again. All they wanted to change was the time-frame (ad hoc ideas kept spinning out of original hypotheses).

Better yet, two company challenges were solved and one initiative that we expected to take two years was completed in three days. Just let that sink in...

If you’re interested in running a problem-solving innovation week, here’s what worked well for us:

1. Set guiding questions for innovation

Before anyone kicked off, they had to submit hypotheses that were constructed based on guiding questions. This wasn’t to constrain teams but to channel efforts into everyday challenges that, if solved, would ultimately support and align the whole team:

For us, those questions included: 

• One of our biggest differentiators is we can truly do server-side only mappings. Moving tags from client to server-side has many benefits but one of the major concerns for clients is cost of hosting. What are creative project ideas optimizing the cost on our current platform?
• Since we’re supporting numerous customers, we maintain many discrete environments, which means we must be proactive and meticulous when it comes to avoiding environmental defects. What are creative project ideas reducing the complexity of operating our current platform?

Of course, this included the caveat that these are should not be treated as hard lines against a project. Any project a team feels passionate about was worth adding as a submission.

2. Formalize tests and rubric

For a couple of weeks, each team submitted proposals that included: 

• Hypothesis: A statement aligned to innovation week goals which can be tested to prove your prediction
• Test to run: This can be one or several experimental ideas that could validate the hypothesis
• Criteria for success: well-defined measure of what success and failure look like
• Goals in running proposed test

Once accepted, all participants received the rubric:

Innovation Week Rubric

Aligned to Business/Industry needs (1-5)

How aligned is this project to solving a real use case that will be valuable in retaining or growing our customer or business?  This could be focused on internal (processes or technology that help us with scaling needs) or external (a feature that will help win business or reduce churn)

Defining Clarity (1-5)

How much has been learned from the process? This can be from a point of both success (We were able to significantly simplify a complex problem and have a great demo) or failure (We definitely would not pursue this, its a fucking technical nightmare).  One major focus for the week is to help product reduce the unknowns for larger ambitious projects, remember if significant clarity can be found it will help product prioritize the feature in the future  

Demoability (1 - 10)

How easy is it to show off what you have done.  What is the actual product feature completeness?  A one here is hey we didn’t build anything and a ten here is a feature-complete project that could be merged and released into our platform.  Bonus for multiple paths (I tested both Kafka and NATS to see which queue would function better as an internal queue) 

3. Define the process: 

Before any team could get started, they had to vet their hypothesis with the Growth team, who would challenge the potential impact and ask questions from the perspective of a team that isn’t already in the weeds of the product every day. Teams could use that feedback to refine their hypothesis if desired. 

Then, on a Wednesday morning, sprints began. Tests were run with the impending deadline of Friday afternoon at 3pm EDT, when a demo had to be ready. At that point, teams had to have Keynotes prepared, outlining their test and outcome. Tests ranged from straightforward questions about a the impact of a hardcoded javascript script added to an output event to use of machine learning to create custom mappings for customers. 

Will we do Innovation Week again? 

We’ve already mentioned the key results. 

• One team came up with a viable idea for optimizing hosting costs by ~70% (for our customers and ourselves). 
• One team successfully ran a test that, if applied largely, could ensure against environmental defects. 
• And one team accelerated years’ worth of engineering work into three days. 

Other teams, too, ran small experiments that will ultimately impact and improve our product. Nobody felt the time was wasted and every experiment taught us something. 

So yes, we’ll do Innovation Week again. And it’ll probably happen soon. 

What’s worked well to help your company innovate? If you have ideas or frameworks that have been successful, please let us know! Our Head of Engineering is always looking for new ideas. 

Photo by Joshua Sortino on Unsplash