I recently wrote a post praising transparency and citing our open source Akash project as an example of our commitment to it. Here’s the thing though — the software that powers the Akash Network isn’t done yet. We’re moving fast, but aren’t even at a point we’d consider a good MVP. So what would possess us to open this in-progress codebase to the world? Today I’ll give you a window into that thought process and offer some thoughts on how it’s going so far.
Not if, but when
We always intended Akash to be an open source project, but shared an implicit assumption that we’d make our GitHub repos public when we were “ready”. When would we be ready though? When it’s finished? When we hit some (arbitrary) milestone? If so, which milestone and what are the criteria for hitting it? This was essentially the conversation when that implicit assumption was surfaced about a month ago.
Ultimately we chose an arbitrary milestone — March 19, the date we planned to publicly announce the Akash Network. Choosing a date made the milestone criteria easy enough, but raised valid concerns about scope. Given a fixed set of resources, defining both scope and date is folly, so we effectively said “On March 19, we’ll ship whatever we have”. Fine, but what if “whatever we have” isn’t good enough?
Yes, it’s good enough: Our leap of faith
I believed that Adam, our CTO, would have a hurdle to overcome. He deeply understands the tech we’re working with, has a vision for where we’re going with the platform and as a result also had a visceral understanding that we’d be nowhere close by March 19. He’s also a bit of a perfectionist. So I was surprised when he responded to the “Ship on 3/19” idea with “Hell yeah, let’s do it.” His rationale was inspiring: we’re doing good work, we’re evolving constantly, and we should trust the community to understand what iteration looks like.
Which brought me to my own hurdle. As a product manager, I’ve had some unpleasant experiences relying on executives’ people’s understanding of iteration. Without proper expectation-setting, demoing an early-stage project can devolve into a conversation about what’s missing, what’s wrong, or how the project has missed the mark. And here we were talking about releasing code into the wild with no ability to “properly” manage expectations. It made me uncomfortable, but what better way to reinforce Akash’s words about transparency than to take the action of developing our product publicly? So we decided to be brave and I braced myself for some retroactive expectation-setting.
So, how are we doing?
The Akash repo has been public for 10 days now and the reaction has been refreshingly positive. Those already familiar with us (investors, telegram members, industry peers) tell us they appreciate the openness, and we’ve received no negative feedback on the early state of the codebase.
That said, we’re still operating with a small sample size. Since 3/19, the Akash site received 1395 unique pageviews, while the repo only received 88 unique visitors (NB: these metrics are measured slightly differently, so take 88/1395 as a rough ratio).
While the sample size is small, our repo is clearly getting traffic from a wider set of visitors than the few who’ve have given us direct feedback. We can draw a couple conclusions from the GitHub traffic analysis.
Few people use the UI to explore the code and very few people read docs past the readme. However 39% of unique visitors have cloned the repo — more than we expected (NB: some unknown amount of clones are from Travis). So the interest level seems high, but we don’t know who’s cloned it other than us and Travis (no forks yet), nor what they’re doing with it. We’d especially like to know who has set up and run our prototype environment. Have you?
Our content (site, blog, Hackernoon, Hacker News) is predictably the biggest driver of traffic to the repo. But social is lagging badly. So we should continue paying attention to content quality and frequency, but up our social game (we have some plans afoot).
Now, keep it up
There’s a certain freedom in these early days. It’s liberating to be candid and is easier when your audience is relatively small and relatively friendly. We intend to continue being transparent as we grow and we view our actions today as setting a pattern of behavior. But stay tuned because our commitment will be tested — we will inevitably screw up sometimes. We’ll miss a deadline, ship an ugly bug, have an outage; something is guaranteed to happen that we would rather sweep under the rug. We look forward to being held accountable!