OpenDM

BYO Labs set out to test an idea about attention. Anyone with a following has the same problem — a public inbox fills with spam, bots and cold outreach, so they hide their email address and become unreachable for the messages that actually matter. OpenDM turned that around. You publish one OpenDM address on every social profile you have, and you set a price. Someone who wants to reach you agrees to your fee before their message is forwarded, and you are paid for each message you read within 48 hours of it arriving. Your real address is never exposed, and you can send your fee to charity instead of keeping it. The pitch on the home page put it plainly: make people pay you to reach you. The audience they were aiming at was the influencer crowd, the people whose inboxes are worst affected and whose attention has a market price.

BYO Labs did not want a finished company. They wanted an answer, which meant getting a working product in front of real people quickly enough to learn something from it. Bitcube built it and kept it running.

Nothing existed at the start — no cloud account, no domain, no code — so the entire foundation had to be laid before a single feature could be tried, and it had to be laid in days rather than months because the point of the exercise was speed to market. The thing being tested was a money flow, which meant it could not be mocked: strangers had to be charged for real and creators had to be paid out for real, with identity and tax handled by somebody qualified to hold that risk. The mechanism itself sat in the hardest part of email, receiving mail sent to arbitrary addresses, holding it back until payment cleared, then forwarding the original message intact and proving afterwards that a human had read it. Real people's private correspondence would pass through that pipeline from the first day it was public, so privacy and spam handling could not be deferred on the grounds that this was only an experiment. And the audience's main network was outside anyone's control: Instagram would not permit sign-in through its API, cutting off the verification route the product was designed around. Finally, an experiment is only worth running if it reports back, so the product had to be measured from the moment it went live.

Alongside the build, the product was wired to report on itself: Google Analytics 4 on 4 August, Microsoft Clarity on 30 August, load tests against the live API, and health checks, error reporting and Slack alerts. That instrumentation set what got built next. The client could see dead clicks piling up on the home page, which became an A/B test of four headline treatments, and a slider people fought with was replaced by a dropdown. Through September the common components were pulled out one at a time into a library documented in Storybook. In October, when Instagram refused sign-in, the team prototyped plain URL entry rather than lose the network.