Weeks 9-13: The Directory Ate August. So I Shipped Something Else.

This journal went quiet for five weeks, and on a build-in-public site, silence is the one genuinely bad look, so here is the honest catch-up. Short version: the directory build ate August, real money landed anyway, and week 13 ended with a new product shipping on a Friday, BuiltByBadge. Longer version below, ugly parts included.

What the directory did to the cadence

Week 8 picked a directory as the next build, and I wrote at the time that the pattern had earned it: structured data survives the AI-answer era in a way articles do not. What I underestimated is what ThrowSpot should have taught me: a directory is a dataset wearing a website, and the dataset is the product. Sourcing it, structuring it, and getting it clean enough to be genuinely the best resource in its niche is slow, unglamorous work that does not fit inside a Monday-to-Friday, and rushing it would ship exactly the kind of thin directory the whole bet is against. So after weeks in the mines, I made the call this project has made once before: when a build refuses to become a week-sized product, pivot to the sharpest adjacent idea and ship that instead. The dataset went on the shelf, and what the directory project actually produced is the build that shipped this Friday, more on that in a second. The streak metric still bent, twelve ships in thirteen weeks instead of thirteen, and I will take that trade. But I should have said all of this out loud in week 9 instead of going quiet, and that is the actual lesson of this post.

The money that landed while I wasn't posting

August put up the project's best month so far, which says more about the baseline than the month: $45.19, bringing the total to $79. Two events inside that number matter more than its size. TextScrubr sold the first copy of its $29 Claude skill, sold directly on the tool's own site, which lit up the second revenue source on the dashboard: Products, next to Affiliate. And the Amazon affiliate line kept trickling without a single new hour of work, $33.81 to $50, which is the first evidence of the thing this whole portfolio is betting on: assets that earn while you are busy building other assets. Neither number is impressive. Both curves point the right way.

Week 13: what the directory became

Here is the pivot, named. Somewhere in the dataset mines, staring at twelve sites that could not see each other, it hit me that the portfolio's connective tissue was a sharper product than one more directory. So the directory build became something else, and week 13 shipped it: BuiltByBadge went live Friday, August 28, and it is the most personal product in the portfolio, because it solves the exact problem this site's own builds have. Twelve live sites, and until now they were twelve strangers: nobody who finds one has any way to find the rest. BuiltByBadge is the fix, generalized: a skill for Claude, ChatGPT, and coding agents that sets up your badge and hub in one conversation, then asks, every time you ship, whether you want your badge on it. Three lines of static HTML, about 200 bytes, zero JavaScript, all pointing home. Free for the core, $29 once for Pro, and the FAQ admits the paywall is the honor system. The full teardown is on its build page.

The scoreboard, updated honestly

As of this Friday: $79 earned, twelve builds live, 0 subscribers, and the X account finally moved, 12 followers to 29, which at this scale is not a growth story, it is 17 real people, and I know which I would rather have. The streak bent this month and the numbers grew anyway, which is either a coincidence or the entire thesis, and the only way to find out is more weeks. Week 14 starts Monday, and its build is already scoped. Back to the weekly logs from here.