Chapter 09 — The daily cadence: twenty minutes that run the business

Products don't fail from missing features; they fail from absent operators. The antidote isn't more hours — it's a cadence: a small, fixed routine you can keep on your worst day. Ours takes about twenty minutes and runs the entire go-to-market. This chapter is the routine, the tools that automate each step, and the philosophy underneath: machines do the scanning; you do only the judgment.

The loop

flowchart TD
    START["☕ Daily start (~20 min)"] --> TRIAGE["1 · Triage the queues
(5 min)"]
    TRIAGE --> FB["Feedback console:
anything New?"]
    TRIAGE --> LEADSQ["Lead queue:
anything Assigned?"]
    TRIAGE --> ERR["Error console:
anything new & frequent?"]
    FB & LEADSQ & ERR --> HUNT["2 · Hunt where users complain
(10 min)"]
    HUNT --> HN["HN radar
(automatic, hourly, strict phrases)"]
    HUNT --> REDDIT["Reddit search kit
(bookmarked searches, new/past-month)"]
    HN & REDDIT --> JUDGE["3 · Human judgment
(5 min)"]
    JUDGE -->|"genuine demand"| REPLY["Draft reply / file lead
statuses: Responded → Won/Lost"]
    JUDGE -->|"noise"| SKIP["Skip — never spray"]
    REPLY --> LEARN["4 · Feed the machine:
recurring questions → FAQ, templates, product"]
    SKIP --> LEARN
    LEARN --> TOMORROW["🔁 Tomorrow, same time"]

Step 1 — triage (5 min). Three inboxes, in order of human urgency: the feedback console (users took time to tell us something — Chapter 10), the lead queue (today's radar sweep, pre-scored), and the error console (crashes the app self-reported). Anything actionable gets assigned; everything else waits unapologetically.

Step 2 — hunt (10 min). Your future users are publicly complaining about the problem your product solves — today, in threads you can read. Two channels, both automated to the edge of judgment:

• Hacker News radar: an hourly sweep (via Algolia's API) for exact-phrase matches on problem language. One hard-won detail: strict phrase matching — quoted queries plus a post-filter. Without it, a search for "wave alternative" matches any comment containing "alternative," and the queue drowns in noise.
• Reddit search kit: bookmarked searches (sorted: new, past month) across the subs where our segments actually live. We originally wanted the Reddit API; its policy approval sat in a ticket for weeks, so we switched to public search RSS and lost nothing. When an API gate blocks you, ask whether the public page already shows what you needed.

Step 3 — judgment (5 min). The radars score; only you decide. Genuine demand → draft a reply in the lead tool (so it's tracked), then post it where the person is, human and specific. Everything else → skip. One honest reply a day beats fifty spray-and-pray comments; the latter also gets you banned, which is a growth strategy with an expiration date.

Step 4 — feed the machine. Every recurring question becomes product: a FAQ entry, a template, a UX fix, a chapter of this series. The cadence's real output isn't replies — it's compounding sculpting of the product toward real demand.

Why "daily games" is really "operator habits"

The brief for this series asked about daily games and player engagement — gamified streaks, invites, nudges. Here's the honest sequence we landed on: engagement mechanics for users only work after engagement habits for the operator. You can't design a retention loop for customers you don't have yet; you can keep a daily loop that guarantees you'll find, hear, and serve the next ten. The user-facing loops (streak-like usage meters, founding codes, reminders) arrive in Chapter 12 — built on top of a cadence that made their data meaningful.

The metrics glance (the extra two minutes)

Dashboard: signups this week, invoices created, feedback count, crawler hits from the access logs (Chapter 11 made reach measurable — GPTBot visiting your sitemap is a real thing you can see now). No charts-for-charts; four numbers, honestly counted, enough to know if the week moved.

Recap. Triage three queues, hunt two channels with machines doing the scanning, spend your five judgment minutes on one great reply, and let every recurring question sculpt the product. Twenty minutes; the compounding is the point.

Next: Chapter 10 — Feedback → tickets → shipped: closing the loop while they remember.