Week 10 - The Feedback Challenge

Jon (Product Lead)

As we’re working on building an MVP, we keep coming up against an issue:

• The technical complexity of a data solution that serves large volumes of data is inherently difficult to descope down to something simple that can be built in a few weeks.?
• At the same time, it’s been difficult to get feedback on our idea from potential users without being able to put a working prototype in front of them.?
• Without feedback and validation on our idea, it’s hard to justify committing 4-6 weeks to building a prototype that we’re not sure users want.

The plan for this week is to get ourselves out of this predicament. The way I see it, we have two options:

1. Launch ASAP with a prototype we know won’t be seen as ground breaking, but is enough for potential users to critique. Providing us with enough information to make a call on whether it’s worth continuing down this path, and if so, what we should build next.
2. Take a leap of faith and bet on our interpretation of the insights we’ve gathered to date. Scoping a functional product that could be built and released within 4-6 weeks, that addresses our current view of the problem-space.

What this will ultimately come down to is whether we think there is something we can launch in a week or two that genuinely provides value to users. The startup mantra is to launch quickly and early, but there’s a risk you erode user trust if you continually launch something that fundamentally doesn’t make their lives any better.?

So far, our feedback gathering process has been 1:1 based, mostly through our existing networks and cold outreach on LinkedIn. Launching publicly has the benefit of casting the proverbial net much wider and perhaps (hopefully) reaching people that resonate with our product concept and are able to provide feedback before having the real thing to play with.

On the other hand, I can see the benefit of deciding to scope and build a functional MVP over a month or two. It’s not as if we’re flying blind.

We’ve spoken to 25 potential users, some multiple times, and have spent a lot of time reading industry papers to deeply understand the problem space. Committing 4-6 weeks still allows us to be nimble and consider changing direction if the feedback suggests we’re not on the right track.

The bottom line is we need to make a decision and create a plan. Which if course we'll share in next week's edition.

Steve’s (Engineering Lead)

As Jon says above, this week was a bit more frustrating than last week. I made some more good progress on the prototype I was building - putting together a decently complicated web application, which linked together energy consumption & energy performance and sketched out some of the possibilities to combine and aggregate this based on postcode areas.

However, showing this to the rest of the team made me realise the potential wasn’t perhaps as obvious as I’d first thought, and the work that would be required to even put it in front of people as-is, would be non-trivial. When we took into account the features we’d need to really explain it, shoring up the data quality, setting up hosting and making the application performant enough to put out there in public, it really started to add up.

So, at the end of the week we took a step back and asked ourselves whether we were really working as smart as we could be. We primarily want to know if this data is useful, and if so, how people want to consume it. With some timely feedback from our friends as Open Climate Fix, we decided to explore another couple of angles - releasing it as cloud-optimised geospatial formats, or directly in cloud databases like Google Big Query. I think my former colleagues at Kaluza will find it especially funny that I’ve somehow managed to find myself figuring out how to put energy consumption data into BigQuery again!?

I’ve still got a lot to learn about this new approach, these kinds of file formats are pretty new to me (please feel free to reach out if you’re an expert and interested in helping us figure things out!), but it seems promising. As an example of why they’ve got some potential, here’s a quick map we whipped up of energy consumption at the substation level across the whole UK. The cool thing about this is there’s no server involved, it’s simply javascript in the browser, loading geospatially indexed data from a 300MB static file using HTTP Range requests.