Build in Public, Month One: Launching a Coffee Brand from a Terminal

If you are reading this post near the time it was published, you are among the first few hundred people who have seen anything HexRoast has put out. That is a nice place to be. These early build-in-public posts are the ones you will be able to reference later and say "I was there when the numbers were small." The numbers are small.

This post is the first in a monthly series documenting what we are doing, what is working, what is not, and what we think the next month should look like. No sugar. No marketer's voice. Actual numbers, actual decisions, actual mistakes.

/ 01The numbers, first

Waitlist-to-visitor conversion is about 5%, which is both a lot better than industry baseline and, honestly, lower than we want. We have some theories about why (below).

/ 02What worked

The hex naming system #4ADE80

The single biggest signal we got this month was from the hex code product naming idea. Every time someone shared the site, that was what they quoted. Every email we got back started with something like "love the hex thing." We spent about three hours on it during a late-night conversation, and it has done more brand-lifting work than anything else we made.

Takeaway: the idiosyncratic, specific, slightly-weird part of your brand is the part that travels. Not the polished copy. Not the logo. The thing that makes people want to explain it to a friend.

Writing the flagship article first #4ADE80

The Brew Stack piece has been 40% of our total visitors. Not because it ranked — it hasn't had time to rank — but because it was the one piece we could confidently point to when anyone asked "what are you doing?" The answer "go read this, it explains everything" is a very efficient answer.

Takeaway: a single good piece of content beats five mediocre ones by every metric we can think of, as long as the good one is genuinely good.

Not trying to go viral #4ADE80

We decided early not to chase initial launch attention. No Product Hunt. No Show HN. No paid influencer. Just write, ship, wait, repeat. This felt slow and bad in the first two weeks and quietly correct by the end of the month. The people who found us are the people we actually want on the waitlist.

/ 03What did not work

• LinkedIn. We put three thoughtful long-form posts there. Engagement was low and the right kind of reader was not there. We are going to stop for now.
• The Hacker News submission. One piece, one submission, one flameout in the "new" queue. We picked the wrong post (a cold brew recipe) for the wrong day. The right HN submission is a long, original, opinionated essay that can generate comments. We know what we're submitting next month.
• The nav structure on week 1. We shipped with "Manifesto / Journal / Shop / About" as nav items and changed it in week 2 when nobody was clicking Journal. The new nav names the actual pillars (Brew / Build / Culture / Community) and clicks doubled.
• Email capture below the fold. We moved the waitlist form to the top third of the homepage in week 3. Conversion doubled. Obvious in hindsight. We left signups on the table for two weeks.

/ 04What we are doing in month two

A short list of the experiments for May, committed publicly so we have to actually do them:

• Pick a roaster. This is the biggest open decision. Seven candidates, three of them plausible, one of them strongly preferred. We need to commit.
• Ship two more posts. Targeting one in Brew and one in Build. The Brew post will be a proper side-by-side tasting; the Build post is tentatively titled "The remote designer's kitchen."
• Figure out the crowdfunding platform. Kickstarter vs Indiegogo vs direct pre-orders on Shopify. We will ask the waitlist directly in a survey.
• Start Instagram. One post per week, all hex-themed. Low effort, high visual fit with our brand.
• One good Hacker News submission. Armed and loaded, we know what it is.

/ 05How to follow along

The next BIP will land in roughly thirty days. Subscribing to the weekly roast log is the easiest way to get it, since we always link to new BIP entries from the email. Alternatively, just bookmark the Community pillar and check in whenever.

Thanks to the 147 of you who signed up during month one. We will try to make month two worth your patience.