TITAN Internal Memo — 2026-05-13
Prepared by: SCOUT
For: Harnoor Singh
Re: Twice-daily newsletter architecture for TITAN's 4-newsletter portfolio
---
The best-executed AM/PM programs in 2026 follow a tightly disciplined formula: distinct editorial mandates per edition, 9–10 hour minimum gap, and Mon–Fri only (weekends suppressed for 70%+ of top publishers).
| Publisher | AM Send (ET) | PM Send (ET) | AM Focus | PM Focus | Reported Open Rate |
|---|---|---|---|---|---|
| Axios AM / Axios PM | 6:00 AM | 4:00 PM | Overnight news + day-ahead | "3 things to know" wrap | AM 50–58%, PM 52–55% |
| Morning Brew / Evening Brew | 6:30 AM | 4:00 PM | Numbered-list briefing, fun tone | Snappier recaps, meme-lite | AM 56–58%, PM 46–50% |
| Politico Playbook AM/PM | 6:00 AM | 4:30 PM | Insider tips + Hill intel | Scoops + vote previews | AM 55%, PM 49–52% |
| Stratechery Daily Update | 7:00 AM | N/A (once-daily) | Deep analysis, one story | — | 60%+ (paid list, high signal) |
| The Information PM | N/A | 5:00 PM | Paid exclusive PM brief | Tech exec scoops | 45–50% (subscriber list) |
Key structural rules extracted from these publishers:
NEWSLETTER-DEDUP-RESEARCH-2026-05-13.md.---
Source: Omnisend 2026 (3.5B+ emails/week), ElectroIQ 2026, Mailchimp benchmarks.
| Time (ET) | Avg. Open Rate | Notes |
|---|---|---|
| 6:00–7:00 AM | 30–35% | Early adopters, commute; lower absolute but highest intent |
| 7:00–9:00 AM | 33–38% | Peak for newsletters; B2B professionals before deep-focus block |
| 10:00–11:00 AM | 33–37% | Strong alternative; inbox revisit after morning stand-ups |
| Time (ET) | Avg. Open Rate | Notes |
|---|---|---|
| 12:00–1:00 PM | 30–33% | Lunch browse; high click-through but lower opens than AM |
| 4:00–5:00 PM | 32–36% | Axios/Morning Brew validated; end-of-work inbox check |
| 6:00–8:00 PM | 25–30% | Evening browsing; lower for B2B, moderate for B2C |
Optimal for tech/AI audience (US Eastern):
Fatigue data point: ElectroIQ flags that sending more than 5 newsletters/week to the same subscriber risks a 22% engagement drop. 38% of subscribers say they want more-than-weekly frequency — but "more often" does not mean "twice daily from the same sender." The implication: AM+PM is sustainable only if the two editions feel editorially distinct, not like the same briefing split in half.
---
Recommendation: Option D — Daily morning sends as today + new weekly "TITAN PM DIGEST" Sunday evening (best-of-the-week)
| Option | Verdict | Kill reason |
|---|---|---|
| A — All 4 go AM/PM | Rejected | 8 sends/day = ~16 Perplexity calls/day at $0.04/call = ~$19/month added cost + extreme subscriber fatigue risk. Over-engineering for an early-stage list. |
| B — Agent-Stack AM/PM only | Tempting but premature | Agent Stack is not yet a standalone product with a distinct subscriber list. No proven demand for daily twice-sends on this topic yet. |
| C — Combined AM Brief + PM Wrap | Appealing architecturally | Collapses 4 newsletters into 2 but loses brand differentiation. OpenClaw subscribers don't want Agentic AI Weekly's stories in their feed. Cross-topic bundling is noise, not signal. |
| D — Morning daily + Sunday PM Digest | Selected | Low cron complexity, zero new daily Perplexity cost, high perceived value (weekend reads are a validated format), and it preserves each newsletter's identity. |
1. Cron complexity is zero. Sunday PM Digest is one new scheduled task firing 6:00 PM ET on Sundays. The four existing AM crons run unchanged.
2. Content source is free. The weekly digest pulls from the already-sent issues of the week — it's a best-of curation, not new research. No Perplexity calls required.
3. Subscriber fatigue profile is low. A single weekly PM send does not compete with daily AM issues. It arrives in a different mental context (Sunday evening, planning for the week ahead). Multiple studies show Sunday evening is a high-intent reading window for professionals.
4. Dedup is trivially handled. The digest reads from the sent_stories SQLite ledger (already designed in NEWSLETTER-DEDUP-RESEARCH-2026-05-13.md), selects top-scored stories of the week, and marks them digest_included = TRUE so they never repeat in Monday's AM issue.
5. It creates a new format without cannibalizing existing ones. The digest is a different product — retrospective synthesis, not live briefing. This is the Stratechery model: deep analysis as a complement to fast daily updates, not a competitor.
---
| Task | Schedule | Trigger | Content source |
|---|---|---|---|
| titan-sunday-pm-digest | 0 18 0 (Sun 6 PM ET = 23:00 UTC) | New | sent_stories ledger, Mon–Sat of that week |
Existing tasks (titan-agentic-ai-weekly Mon, titan-openclaw-weekly Mon, titan-claude-weekly Wed, plus titan-daily-newsletter) run unchanged.
TITAN PM DIGEST[TITAN PM] Best of the week — {date_range} · {top_entity}1. Opening header — "What TITAN tracked this week" (2 sentences)
2. Top story per newsletter — 1 pull-quote + link from each of the 4 newsletters (4 blocks)
3. Editor's pick — highest-scored story across all 4, with 2-sentence analysis
4. What to watch next week — 3 bullets generated from forward-looking signals in the week's research
daily-email-template.html base. Add coral section dividers for PM variant. Subject line badge [PM DIGEST] in teal.| Edition | Source | Freshness window | Dedup gate |
|---|---|---|---|
| AM (existing) | Perplexity + RSS, live | Last 24–48 hours | sent_stories table, 14-day lookback |
| Sunday PM Digest | sent_stories ledger | Mon–Sat of current week | digest_included flag, prevents AM re-use |
---
Subscriber fatigue: Low. One additional send per week (Sunday PM) is below any documented fatigue threshold. The 5x/week cap flagged by ElectroIQ refers to sends to the same subscriber on the same topic — Sunday Digest is cross-topic and retrospective.
Content cost (Perplexity): Zero incremental cost. Digest is assembled from the ledger, not new research calls.
Dedup failure: If digest_included flag is not set correctly, a story from Saturday's send can appear in both the Sunday Digest and Monday's AM issue. Mitigation: set digest_included = TRUE atomically at digest-generation time; Monday AM scraper filters WHERE digest_included IS FALSE.
Format fatigue over time: The weekly digest format has a shelf life of 6–12 months before it needs a redesign. Monitor open rates on the Sunday send monthly. If Sunday PM drops below 28%, either tighten curation (fewer, better stories) or introduce a new section (e.g., a podcast summary).
Option B as a future upgrade: If Agent Stack's subscriber list matures and shows demand for faster cadence, Option B (Agent Stack goes AM/PM) is the logical Phase 2. The Sunday PM Digest infrastructure (dedup flag, new template) makes that upgrade easier — PM edition reuses the same ledger query with a tighter time window.
---
Memo path: F:/TITAN/plans/audits/AM-PM-NEWSLETTER-STRATEGY.md
Recommended option: D
Total daily sends under recommendation: 1 weekday / 2 Sunday (AM + PM Digest)
Weekly sends: 6 (was 5)