Ecomma Email Analysis. Feb to Apr 2026

Top-line read

Four things that show up in the data before you read another row.

Opens are decent. Clicks are flat.

Average proxy-excluded open rate is 15.3% across 44,453 sends. Average click rate is 0.42%. The list opens at a healthy clip, then doesn't act. Click-to-open is 1.30% versus a Mailchimp all-industry average around 14%.

The Office Opening invite was the standout.

The April 6 invite hit 30.4% proxy-excluded open rate, roughly 2x the broadcast average. The two reminders (#8 and #9) regressed to ~12%. Event invitations work the first time. Reminders to the same list within 11 days do not.

A/B testing is set up but not learning.

All seven variate campaigns ran with test_size 100% and wait_time 0h. No winner_combination_id was ever declared. Three of seven tested the same subject in both cells. The variate engine is producing no actionable signal.

List is leaking faster than benchmark.

0.64% unsubscribe rate per send, roughly 2.5x Mailchimp's all-industry average of 0.26%. April alone: 313 unsubscribes plus 517 deleted contacts on the Buyers list.

Performance benchmarks

Against Mailchimp's published all-industry averages. Proxy-excluded open rate strips out Apple Mail Privacy Protection auto-pixels and is the truer engagement signal.

Total volume

44,453

9 broadcasts. ~5,000 per send. 60-day window.

Delivery rate

99.64%

Healthy. Bounce rate 0.36% sits well under the 2% threshold.

Open rate (raw)

31.9%

Below avg. Mailchimp all-industry: 35.6%. Marketing/Advertising: 30.07%.

Open rate (proxy-excluded)

15.3%

Truer signal after Apple Mail Privacy Protection. Use this number internally.

Click rate

0.42%

Critically low. Mailchimp all-industry: 2.62%. Even allowing for tracking gaps on variate sends, this is a real problem.

Unsubscribe rate

0.64%

Roughly 2.5x avg. Mailchimp all-industry: 0.26%. Indicates content/audience mismatch or cadence fatigue.

Campaign-by-campaign

Sorted chronologically. OR_pe is proxy-excluded open rate. CTOR is click-to-open rate. Times shown in Dubai (UTC+4).

Sent (Dubai)	Day	Title / subject	Sent	OR raw	OR_pe	Clicks	CTOR	Unsub
Feb 24, 08:02	Tue	Re-Engagement #1. "Still interested in ecommerce acquisitions?"	4,349	35.1%	17.0%	65	4.29%	1.20%
Feb 27, 09:20	Fri	Re-Engagement #2. "Why we acquired this eyewear brand at $900K"	4,404	31.7%	15.1%	8	0.57%	0.57%
Mar 02, 10:27	Mon	Iran-UAE Operational Update. "Iran–UAE Situation: Operational Update from Ecomma" / "While the world paused, our portfolio hit record revenue."	4,471	33.2%	16.5%	7	0.47%	0.69%
Mar 05, 17:00	Thu	WWP Highlight. "New Deal: USA Health & Wellness Brand \| $300K Feb Revenue (+119% YoY)"	4,539	27.9%	11.2%	8	0.63%	0.51%
Mar 12, 17:22	Thu	Cost Optimisation. "How to 33% in profits after an acquisition"	4,778	27.9%	11.4%	0	0.00%	0.82%
Apr 06, 19:21	Mon	Office Opening Invite. Best "You're invited: Ecomma Dubai Office Opening — April 24" / "Are you coming to our Office Opening Party? April 24"	5,354	44.6%	30.6%	40	1.68%	0.86%
Apr 14, 12:45	Tue	$1.8B / 0 employees. "$1.8B Ecommerce Company. 0 employees."	5,463	29.8%	12.4%	13	0.80%	0.29%
Apr 17, 18:45	Fri	Office Opening Reminder #8. "You're invited (again): Ecomma Dubai Office Opening — April 24" / "Additional spots released: Ecomma Dubai Office Opening (April 24)"	5,505	29.5%	12.1%	24	1.48%	0.44%
Apr 23, 17:15	Thu	Office Opening Reminder #9. Diminishing returns "You're invited for Tomorrow: Ecomma Dubai Office Opening — April 24" / "Tomorrow Is Our Office Opening Party - Are you coming?"	5,590	27.9%	11.3%	19	1.23%	0.48%

Subject line patterns

Ranked by proxy-excluded open rate. Patterns visible across nine subjects, not statistically conclusive.

Top three subjects

"You're invited: Ecomma Dubai Office Opening — April 24"

30.6% OR_pe Event invite. Specific date. First send of the series.

"Still interested in ecommerce acquisitions?"

17.0% OR_pe Direct question. Re-engagement framing. Drove the only standout CTOR (4.29%).

"Iran–UAE Situation: Operational Update from Ecomma"

16.5% OR_pe News tie-in. Implies urgency and relevance.

Bottom three subjects

"New Deal: USA Health & Wellness Brand | $300K Feb Revenue (+119% YoY)"

11.2% OR_pe Information-dense. Reads like a press release.

"You're invited for Tomorrow: Ecomma Dubai Office Opening — April 24"

11.3% OR_pe Third reminder. Same audience, third time hearing it.

"How to 33% in profits after an acquisition"

11.4% OR_pe Has a typo (missing verb after "How to"). Possibly hurt deliverability or click intent.

Patterns worth testing more rigorously

Direct questions (re-engagement) and event invites with concrete dates outperformed information-stacked subject lines. Hard numbers in subjects ("$1.8B", "$300K +119%") did not produce above-average performance. Reminder sends to the same audience saw a sharp drop on the third pass. The first invite hit 30.6%, the second-week reminder 12.1%, the day-before reminder 11.3%.

Send time and cadence

Times in Dubai (UTC+4). Caveat: most slots have n=1, so treat as descriptive rather than directional.

Day (Dubai)	Sends	OR_pe (avg)	CTR	Unsub	Note
Monday	2	24.1%	0.48%	0.78%	Includes the Apr 6 19:00 invite (the standout).
Tuesday	2	14.3%	0.79%	0.69%	Tue 08:00 Dubai (Feb 24 re-engagement) drove the click rate.
Friday	2	13.4%	0.32%	0.49%	Lowest unsub of the four days. Mid-day Friday Dubai may sit better than expected.
Thursday	3	11.3%	0.18%	0.60%	All three sent at ~17:00 Dubai. The lowest performing slot in the data.

Cadence: 9 sends in 60 days, accelerating

Average gap between sends is 6.6 days. The pace tightened in April (4 sends in 17 days for the Office Opening sequence). The sharp drop in opens from 30.6% to 11.3% across the three-send invite arc, combined with the high unsubscribe rate, suggests the Buyers list is being hit too frequently for the content depth on offer.

Segment performance

Of 9 broadcasts, 7 went to the entire Buyers list with no segmentation. Two used a defined segment (the February re-engagement sends).

Audience approach	Sends	Total recipients	CTOR (avg)	Unsub rate	Read
Defined segment (Feb 24 + Feb 27 re-engagement)	2	8,753	2.51%	0.88%	The Feb 24 send carried the entire CTOR. The follow-up regressed.
Full Buyers list (all other sends)	7	35,700	0.99%	0.58%	Higher volume, lower engagement. Diluted intent.

N=2 segmented sends is not a trend

The CTOR gap between segmented and unsegmented sends is meaningful but the sample is too thin to claim segmentation alone explains it. The re-engagement campaigns also had different content (direct question, lower-friction CTAs to Calendly + teasers.ecomma.co), which likely contributed.

Operational gaps

What the API surfaced beyond the campaign-level metrics.

A/B variate engine is misconfigured

Every variate campaign was set up with test_size = 100% and wait_time = 0h, which means both cells are sent simultaneously to the entire list. Mailchimp's "test then send winner to the rest" logic never engages. Result: zero declared winners, no compounding learning across the seven A/B campaigns. Three of the seven tested only one subject in both cells (same subject_idx in both combinations), so they were not subject tests at all.
No automation flows are live

The Mailchimp `/automations` endpoint returns zero items. There is no welcome series, abandoned form, post-event nurture, or winback flow. All 22 entries showing as "automation-email" in the campaign list are individual flagged sends with no parent workflow attached.
No e-commerce store integration

The `/ecommerce/stores` endpoint is empty. Mailchimp cannot attribute revenue or product views to email opens, which means no campaign in the dashboard has a revenue number attached. For an account named "Ecomma" running a portfolio of Shopify DTC brands, this is the single highest-leverage missing integration.
Click tracking is sparse on variate sends

The `click-details` endpoint returns zero unique URLs for all seven variate campaigns, despite the report endpoint showing non-zero `clicks_total`. Only the two regular (non-variate) campaigns have functioning per-URL click data. This makes it impossible to know which CTAs are working in the larger sends.
List health is moving the wrong direction

April growth-history: 313 unsubscribes plus 517 deleted, against 0 imports and 0 optins. The list is contracting through unsubscribes and cleaning, not growing. At the current 0.64% per-send unsub rate over an 11-day cadence, the Buyers list shrinks by roughly 2% per month before any acquisition.

What to test next

Ranked by expected leverage. Pick at most two for the next 30 days.

Fix the A/B configuration before sending another variate campaign

Switch test_size to 20% and wait_time to 4 hours. Make the two subjects genuinely different (not minor wording tweaks). Let Mailchimp pick the winner and send to the remaining 80%. This produces actual learning per send instead of nine sends of zero-signal data.
Pull the cadence back and re-segment the Buyers list

9 sends in 60 days with a 0.64% unsub rate suggests the audience is being mailed faster than it tolerates. Test cutting volume to one send every 10 to 14 days for the next 6 weeks. Pair with a 30-day no-open segment and ship a winback before they unsubscribe themselves.
Diagnose the click problem

0.42% click rate against a Mailchimp benchmark of 2.62% is roughly 6x off. Two parallel actions: (a) verify Mailchimp click tracking is on for variate templates and that links are not raw http (some teasers.ecomma.co clicks worked, the rest tracked at zero), (b) audit the actual CTA placement and clarity. Front-load one clear CTA above the fold rather than three equal-weight links.
Connect the Shopify store to Mailchimp

Without store integration, none of this analysis can answer the question that matters: did this email drive revenue. Once connected, every future campaign report carries an attributed revenue number, and the difference between "Iran-UAE update" and "Office Opening invite" stops being a directional read on opens and becomes a measurable read on dollars.
Build the welcome and post-event automations that should already exist

Welcome series for new Buyers. Post-event nurture for the Office Opening attendees. Winback for 60-day inactives. These produce compounding value with no recurring effort and would cover a meaningful share of the volume currently going through manual broadcasts.