Destin's email programme has a healthy top of funnel and a broken middle. List growth runs at 5.5% monthly against a 2–5% benchmark, deliverability is 99.8%, and open rates are far above benchmark. But campaign click rate is 0.98% against a 2–5% benchmark, click-to-open is 2.08% against 10–20%, and the click rate has fallen for three consecutive months to 0.70% in July — its lowest reading in the series.
The more uncomfortable finding is about measurement. The campaign programme generated 162 orders across 37 sends in twelve months. At that volume a typical campaign expects five orders with a 95% range of one to ten purely by chance. Most campaign-to-campaign revenue differences at Destin are noise, and ranking creative by revenue mostly reads that noise.
I tested three claims from the first version of this report. One survived, one failed, one was never supportable:
| Claim | Verdict |
|---|---|
| Plain text beats designed | Confirmed — 1.81× conversion rate, p = 0.0014 |
| Collections beat single-product emails | Retracted — p = 0.85, no effect |
| Click rate predicts campaign revenue | Not supported — r = 0.12, p = 0.48 |
That leaves two levers with evidence behind them: use plain text, and don't go dark. The calendar below is built on those, not on creative theme selection.
Acquisition is the bright spot, with one live defect. The homepage popup converts at 6.56% against a 2–5% benchmark and drives 79% of all list growth. Its A/B test has been running since 2 April and shows that adding an SMS capture step costs nothing in email signups (p = 0.895) while adding 18% more step submissions — so it should be rolled out. But those phone numbers go nowhere: both toll-free sender registrations failed on 20 April, so Destin has spent four months collecting SMS consent into a channel that cannot send. See section 9.
Last 12 months. Benchmarks are standard email-marketing ranges; ✅ at or better than benchmark, ⚠️ marginal, ❌ below.
| Metric | Campaigns | Flows | Benchmark | |
|---|---|---|---|---|
| Delivery rate | 99.76% | 99.14% | 95–99% | ✅ |
| Open rate | 47.03% | 36.62% | 15–30% | ✅ (see MPP note) |
| Click rate | 0.98% | 2.34% | 2–5% | ❌ campaigns / ✅ flows |
| Click-to-open | 2.08% | 6.38% | 10–20% | ❌ both |
| Unsubscribe rate | 0.516% | 0.451% | <0.5% | ⚠️ campaigns / ✅ flows |
| Conversion rate | 0.123% | 0.732% | 1–5% | ❌ both |
| Revenue per email | $0.0522 | $0.3444 | — | flows 6.6× better |
| List growth | +5.51% monthly | — | 2–5% | ✅ |
On the open rate. Treat 47% as inflated. Apple Mail Privacy Protection auto-opens messages, which lifts the numerator and mechanically depresses click-to-open. Some of the CTOR gap is measurement artifact, not audience behaviour.
Click rate is the clean signal. Clicks per delivered are unaffected by MPP. At 0.98% campaigns sit below the floor of the 2–5% band, while flows sit inside it at 2.34%. Same list, same brand, same products — so this is a campaign-execution problem, not an audience problem.
100 delivered → 47.0 open → 0.98 click → 0.12 order
▲ far above ▲ below floor
benchmark of benchmark
Awareness and deliverability are solved. The leak is at Interest → Consideration: people open and then don't click. Per the optimization framework, that points at email body content, hierarchy, CTA placement and offer clarity — not at subject lines, sender reputation, or list quality.
162 orders across 37 campaigns. Median campaign: 4 orders. Range: 0 to 15.
| Delivered per send | Expected orders | 95% range by chance alone |
|---|---|---|
| 4,000 | 4.9 | 1–10 (±104%) |
| 10,000 | 12.3 | 6–20 (±63%) |
| 25,000 | 30.6 | 20–42 (±37%) |
| 50,000 | 61.3 | 46–77 (±26%) |
A campaign returning $327 and one returning $55 differ by roughly five orders. That is inside the noise band. A chi-square test across all 37 campaigns lands at p = 0.049 — campaigns differ, but only barely, and not enough to rank them individually.
What this changes:
The first version said "every winner is a collection, new arrival or seasonal edit; every loser is a single product category." That does not hold:
NEW! Anklets is a single product category and was the top full-price earner at $327; I had filed it under "new arrival" to preserve the pattern.It was pattern-matching on six orders. It has been removed from the plan.
Plain text converts 1.81× better than designed. 43 orders on 22,009 delivered (0.195%) versus 119 on 110,200 (0.108%), p = 0.0014. It also clicks better — 1.19% vs 0.94%, CTOR 2.59% vs 1.98%. This is the single strongest, best-evidenced finding in the account.
Caveat carried forward: only 5 plain-text campaigns have run, and the two best went to the widest 720-day audience, so format and audience breadth remain partly confounded. The effect is large enough to act on and large enough to be worth deliberately de-confounding — see the test plan.
Don't go dark. Nothing has been sent since 26 July — 16 days. Last year the same drift ran 84 days (30 Jul → 21 Oct 2025) and August 2025 finished at $8,459 with zero campaigns. This is a structural finding from operational history, not a creative one, and it is the highest-confidence item on this page.
Campaign click rate, by send month — declining.
| Oct | Nov | Dec | Feb | Mar | Apr | May | Jun | Jul | |
|---|---|---|---|---|---|---|---|---|---|
| Click rate | 1.07% | 1.49% | 1.13% | 0.90% | 1.02% | 1.11% | 0.98% | 1.00% | 0.70% |
| Delivered | 1.6k | 7.4k | 4.3k | 11.7k | 27.7k | 11.9k | 30.7k | 12.2k | 24.7k |
July is the lowest reading on record, on the second-highest volume. The pattern across the year is that the biggest-volume months carry the lowest click rates — consistent with reaching further into less-engaged audience as send size grows. Watch this; if August lands below 0.70% the audience expansion strategy needs revisiting.
Unsubscribe rate is marginally hot at 0.516% for campaigns against a <0.5% benchmark, and 0.559% for full-price sends specifically. Not alarming, but it removes the case for pushing beyond 6 sends this month.
Seasonality is unreadable. Only one August exists in the data and it had zero campaigns, so there is no basis for an August seasonal adjustment. Any month-shape claim here would be invented.
Six sends, weekday mornings 08:05–08:35 CT, matching the account's established slot. Ship Summer's Best Accessories; retire the third Anklets run (already sent 21 Apr and 16 Jun to overlapping lists) and fold anklets in as a product block.
Format assignment is the deliberate part. Four of six are plain text — up from 5 in the whole prior year — because that is the only creative variable with statistical support. Themes below are reasonable and on-brand, but the honest position is that theme choice is not evidenced to matter at this volume; they are there to give each send a reason to exist, not because the data ranked them.
| # | Date | Campaign | Format | Audience | Why |
|---|---|---|---|---|---|
| 1 | Wed 12 Aug | 081226 - Summer's Best Accessories |
Designed | 365D Engaged (~3,900) | Existing draft — reschedule and ship today |
| 2 | Mon 17 Aug | Never Take It Off | Plain text | 365D Engaged (~3,900) | Waterproof proof, peak late-summer relevance, review-led |
| 3 | Thu 20 Aug | The Gold Question | Plain text | 365D Engaged (~3,900) | The 40%-of-negative-reviews objection, never addressed |
| 4 | Mon 24 Aug | Back to School, Back to Real Life | Designed | 720D Engaged (~5,800) | Gift framing justifies the widest list |
| 5 | Thu 27 Aug | How to Stack It | Plain text | 365D Engaged (~3,900) | Multi-item sets lift units per order |
| 6 | Mon 31 Aug | New for September + Labor Day preview | Plain text | 720D Engaged (~5,800) | Closes the month, teases 7 Sept without discounting |
Send 3, The Gold Question, is the one I would build first. Gold tone drives 40% of negative reviews — the "brassy in person versus the photos" gap — and no campaign has ever addressed it. Its value lands in returns and review scores as much as in attributed revenue, which is exactly the kind of return this account's campaign-revenue data is too noisy to measure anyway.
Every send carries one click-focused change, since clicks are where the funnel leaks and where results are measurable: a single primary CTA above the fold, product links in the first 100 words, and no more than four link destinations per email.
Derived from the measured campaign conversion rate and AOV ($42.58), with ranges from Poisson sampling variation rather than percentiles of history.
| Scenario | Expected orders | Expected revenue | 95% range |
|---|---|---|---|
| 6 sends × 4,000, historical format mix | 29.4 | $1,252 | $809 – $1,703 |
| 6 sends × 4,000, all plain text | 46.9 | $2,078 | $1,507 – $2,703 |
| Plan as specified (4 of 6 plain text) | ~40 | ~$1,750 | — |
The format decision is worth roughly $830 in expectation — more than any theme choice on this calendar, and the only lever here with a p-value attached.
Confidence: moderate. Ranges reflect sampling variation only. They do not account for August seasonality (unmeasurable — one prior August, no campaigns), audience fatigue, or the declining click trend, any of which could push results below the stated band. Per forecasting practice, treat a 3-week projection built on 37 prior data points as directional.
Keep it in proportion. Flows produced $19,106 of last year's $26,187 attributed and will carry August regardless. Six campaigns move the month by one to two thousand dollars, not ten. The real argument for sending is that August 2025 ran zero campaigns and finished among the weakest months on record.
Ranked by impact against effort.
High impact, low effort — do now 1. Reschedule and ship the stale draft today. 16 days dark, and last year's blackout started exactly this way. 2. Shift the format mix to plain text. Worth ~$830 this month; costs less production time than designed. 3. Cut link destinations to four per email, primary CTA above the fold. Directly targets the one broken funnel stage.
High impact, higher effort — plan and resource
5. Fix the SMS toll-free registration. Both registrations returned registration_failed on 20 Apr 2026. Phone numbers are being collected right now into a channel that cannot send — this is the most wasteful defect in the account.
6. Diagnose the footer form. 22,061 unique viewers, 141 signups, zero qualified_form events. At popup conversion it would be worth ~1,400 signups a year.
7. Investigate the popup's seven-month submit-rate decline (10.81% → 5.32%). It is still above benchmark, but it drives 79% of list growth, so the trend matters more than the level.
8. Build The Gold Question. Addresses a defect affecting 40% of negative reviews, with downstream value in returns and ratings.
Lower priority 9. Theme and product-mix optimisation — deprioritised. The data cannot currently distinguish good themes from lucky ones.
Two questions worth designing for, given every campaign-level result is noise-dominated.
Test A — de-confound plain text from audience breadth. Format is currently entangled with list size. Run one plain-text and one designed send to the same audience definition within the same week, judged on click rate, not revenue. At ~4,000 delivered a click-rate difference is detectable; a revenue difference is not.
Test B — does audience expansion cost click rate? The pattern suggests the largest sends carry the weakest click rates. Send the same creative to 365D and 720D segments and compare click rate directly. If the wider list clicks materially worse, the "go broad for revenue" habit is buying volume at the cost of engagement and long-term deliverability.
Both are one-variable tests judged on a well-powered metric. Document the result either way — a test that confirms the current approach still buys confidence.
Homepage - Multistep (VABSSs) is the main popup — 20% off, static code UNLOCK20, 5-second delay, 45-day re-show, suppressed on the bundle page. Steps: email → preference question ("Tarnish Free / 100% Waterproof / Affordable") → coupon reveal.
Popup (VABSSs) |
Footer embed (SG8NCa) |
|
|---|---|---|
| Unique viewers | 73,291 | 22,061 |
| Submits | 5,634 | 141 |
| Submit rate | 6.56% ✅ vs 2–5% benchmark | 0.40% ❌ |
The popup is the acquisition engine. Popup + footer produce 5,775 signups against 7,303 new email subscribers — 79% of all list growth. At 6.56% it is comfortably above the 2–5% ecommerce benchmark. This is the healthiest asset in the account.
Popup funnel: 73,291 viewers → 25,385 qualified (35%) → 5,634 submitted (22% of qualified, 7.7% of viewers). 39,148 closed without submitting — 53% of everyone who saw it.
Submit rate has halved since December.
| Dec | Jan | Feb | Mar | Apr | May | Jun | Jul |
|---|---|---|---|---|---|---|---|
| 10.81% | 9.66% | 7.93% | 6.80% | 7.16% | 5.44% | 6.02% | 5.32% |
Seven months of near-continuous decline, on rising traffic (July had the highest unique viewers of the year at 9,002). Still above benchmark, but the direction is wrong and it is the largest single input to list growth. Likely candidates: offer fatigue on a static 20% code, or a shift in traffic mix. Neither is confirmed.
The footer form is broken. 22,061 unique viewers produced 141 signups. At the popup's rate it would have produced ~1,400. It also records zero qualified_form events, which suggests a targeting or eligibility misconfiguration rather than weak creative.
A test has been running since 2 April 2026: CONTROL - Homepage - Multistep (25228365) versus NEW SMS STEP - Homepage - Multistep (25228366), which inserts a "PREFER TEXTING?" phone-capture step with full SMS consent disclosure.
| Last 90 days | Control | SMS step |
|---|---|---|
| Qualified | 12,466 | 12,592 |
| Submits | 714 | 727 |
| Submit rate | 5.73% | 5.77% |
| Form-step submissions | 1,409 | 1,668 |
The SMS step is free. Email submit rate is 1.008× — statistically indistinguishable (p = 0.895) — so adding a phone step costs nothing in email capture, while generating 259 additional step submissions (+18%), which are the phone numbers. At current volume the test can only detect a lift of roughly ±10% relative, so "no difference" here means "no meaningful harm", not "proven identical".
Recommendation: end the test and roll the SMS variant to 100%. It has run four months with no downside and a clear upside.
But there is a live defect. Phone numbers from variant B are collected into list QPevK2 — and SMS opt-ins read zero across all 24 months, because both toll-free sender registrations failed on 20 April 2026, eighteen days after this test launched. Destin has been capturing phone numbers into a channel that cannot send. Every number gathered since April is consent collected and unused, and consent ages.
Both halves must be fixed: resubmit toll-free verification and roll the SMS step to 100%. Doing either alone leaves SMS at zero.