Protein Coffee: what the launch actually produced, reconciled

1What the launch produced, verified

The arc is complete: eight slots, 21 production sends (9 email, 12 SMS) from 23 September to 13 October, every one read fresh from Klaviyo on 2026-10-20 with attribution settled (the youngest send is 7 days old; attribution settles in about five). Store figures come from every Shopify order in the window, read 2026-10-20 22:42 UTC. Nothing on this page is carried from the 8 October build except the six older launch days, which are labelled.

The eight questions, answered

Every finding in this report, worst first

Each finding carries a status against the 8 October build: CONFIRMED, REVISED, RESOLVED, DISPROVEN, UNVERIFIABLE or NEW. Click one to jump to it.

2Complete launch timeline and campaign inventory

Every production send between 23 September and 13 October (the last wave went at 20:00 ET, which is 14 October UTC), as named in Klaviyo, with its id. Read from GET /api/campaigns (both archived states, both channels) on 2026-10-20 22:16 and valued with campaign-values-reports against Placed Order. GlockApps seed sends are excluded. Times are Eastern; the Klaviyo dashboard shows the same sends in UTC.

RPR is Klaviyo’s revenue_per_recipient = attributed revenue ÷ delivered. CVR here is attributed orders ÷ delivered. SMS “Click” is link-scanner traffic, roughly 95% machine, and is shown only to document it. Every email row cross-foots: orders × AOV = revenue to the cent on 21 of 21 rows. Audiences: email sends include 3 lists/segments and exclude 14 suppression segments (15 on the last-chance main arm, which added the buyers segment XhKjwD); SMS sends include 7 audiences with no exclusion, except the two last-chance waves which exclude XhKjwD.

3Email and SMS performance by campaign and phase

All figures Klaviyo-attributed GROSS, settled. Revenue per recipient (RPR) is the house decision metric and is relative, so it survives the attribution halo that § 4 measures; dollar totals do not.

By phase

Did revenue per recipient deteriorate through the campaign? Re-assessed on settled data

Attribution maturity, measured on this arc

The 12 October column is the Klaviyo dashboard as read that day (the read that caught the 8 October figures). A send read on its own day is worth roughly a third of its settled value on launch day and about half on an ordinary send; by day 4 it is within a fifth of settled; by day 7 it is settled. This is now four observations, not one, and it is the reason no number on this page is younger than seven days.

Three launches, side by side — whole arcs, equivalent maturity

Sparklines are email RPR by send, each scaled to its own arc. Nattokinase’s arc excludes the Memorial Day promo ($121,621 on 29 May, $70,677 on 1 June) that cut it; Black Seed Oil’s excludes the 311-recipient VIP 70% variant on launch day ($1,226). All three are settled.

Revenue per recipient, beat by beat, against the house record

Open rate and list hygiene

The SMS waves: twelve sends, four of them same-day repeats

The last-chance buyer split

4Shopify actual sales versus Klaviyo attribution

Every Shopify order created between 22 September and 20 October (25,881 orders, of which 25,547 live after removing cancelled and test), read 2026-10-20 22:42 UTC. Definitions: store GROSS = total_price (tax and shipping in, before refunds); store NET = line revenue − discounts − refund line items, ex tax, ex shipping; coffee NET = the Ancestral Beef Protein Coffee lines (product id 15411323011299) net of their allocated discounts, ex tax and shipping. Cancelled and test orders are excluded from every money column.

Reconciliation by window

Campaign claims are assigned to the window of their send date; a send keeps attributing for about five days, so the “whole arc” row (sends to 13 Oct, store through 18 Oct) is the only one where the two sides are fully matched. Campaign and flow claims are not added as unique revenue anywhere on this page: Klaviyo does not expose the order-level join, so their overlap on a single order is [DATA MISSING: order-level campaign/flow attribution]. The “All Klaviyo” column is an upper bound.

The store, day by day

Accounting identities, every order

• Identity A (line revenue − order discounts + tax + shipping = total_price) holds to the cent on 25,547 of 25,547 live orders. Zero breaks; the 8 October audit had found one edited order in the Sleepy Cocoa window and none here.
• Identity B (total_price − refund line items incl. tax − order adjustments = current_total_price) holds on all but 22 orders, net residual $-383.17 across 25,547 orders. Refund transactions total $2,056.22 while refund line items incl. tax total $9,327.42: refunds issued as store credit or without a cash transaction, the same mechanism found on 8 October. Refunds are read from line items everywhere on this page.
• Line-level discount allocations vs order-level total_discounts differ by $194.73 in total (shipping discounts, which allocate to no line). Order-level discounts are used for store NET; line allocations for coffee NET.
• Attributed orders × AOV = attributed revenue on 21 of 21 Klaviyo rows.

What the SMS cost

Contribution margin and ROI are not computed: unit cost of goods and fulfilment cost per order are [DATA MISSING: COGS per unit and fulfilment cost on this product], and the SMS invoice allocation is above. Profitability is stated only as what can be read: credits, two unit prices, attributed GROSS and product NET.

5Launch performance versus historical benchmarks

Two instruments disagree, and both are right about what they measure. Klaviyo revenue per recipient says the coffee launch beats were the weakest on record. Shopify says the launch day was normal on product sales and the biggest store day of the three arcs. This section reconciles them.

Launch day, seven arcs, audited

The six older rows are complete store-local days audited on 8 October and are not re-pulled (their refunds may have moved by a few dollars since). Protein Coffee’s row replaces the two partial reads of 8 October (19:44 UTC: 990 orders / $73,692 GROSS / $23,058 product; 20:59 UTC: 1,101 / $81,478 / $25,521) with the complete day: 1,791 orders, $140,752 GROSS, $34,959 of coffee, 982 units. Klaviyo campaign claims on the four older launch days are email-only because no SMS campaign existed on those days.

The launch phase, not just the day — three arcs on the same definitions

Why the RPR is the lowest on record while the store day was the largest

Every prior launch warmed the list with an offer. This one did not. — still true

Settled launch-day conversion: Nattokinase 0.461%, Black Seed Oil 0.314%, Protein Coffee 0.194%. The coffee figure is now complete and no longer understated. The association stands at six arcs by name and a conversion comparison across three; it is still not a causal claim, and the Nattokinase promo was a calendar holiday.

What this still does not prove

• Three arcs is not a sample. Product-level launch-phase comparisons exist for three arcs; launch-day comparisons for seven.
• Arc length and offer cadence are confounded. Coffee ran 15 days and 5 pre-launch sends with no offer; Nattokinase 21 and 7 with a holiday promo; Black Seed Oil 34 and 14 with a BOGO qualifier. No single variable isolates.
• The in-house rival. Longevity Coffee ($42.96) sits in the same format. Two commenters asked whether the new coffee replaces it; no send answered. Whether it split demand is [DATA MISSING: Longevity Coffee units in the same window were not isolated on this run].
• Price. The most common unit price paid in each launch window: coffee $35.96, Black Seed Oil $29.96, Nattokinase $26.96 (read from the order lines); launch-day basket size is similar across arcs (coffee $70.67, Black Seed Oil $68.42), so unit price does not explain the gap.

6Customer acquisition, opt-ins and pop-up performance

The founding list RunR7E “Ancestral Coffee Launch SMS Opt-In” re-read in full on 2026-10-20, every member with join date, consent and profile age; Klaviyo form reporting for the whole coffee window (20 Sep – 20 Oct); the Submitted Form event stream; and a per-profile sweep of what every member received and bought.

Opt-ins are list joins stamped on the send date. The page reads a ?src= parameter; none of the 21 sends carried one (every SMS link was checked), so same-day attribution is the best available. Launch-phase joins came from people who were already being messaged about the offer, so they are not acquisition in the pre-launch sense.

What the founding list was worth

Prior spend, carried from 8 October: of the 1,078 members then, 754 (70%) had bought before, $332,540 across 4,119 lifetime orders (GROSS). This sweep reads events from 15 September only and cannot refresh it. The founding list is the best of the file raising its hand, not a cold audience.

Every form with activity, whole coffee window

Tracking gaps named on 8 October, status today

7Blog engagement and customer demand

Hyvor Talk comment data for the three coffee articles and the three busiest articles of each prior arc, re-pulled 2026-10-20. Totals are exact; the comment-body sample caps at 50 top-level comments, which truncates the busiest Black Seed Oil and Nattokinase threads but none of the coffee ones.

The objections readers raised, and where each stands

8Operational execution and the innovations, assessed

Read from both hubs and every slot brief (fetched 2026-10-20), the ten ClickUp tasks live, and the campaign definitions themselves.

ClickUp: nine tasks, nine closed, and a hub that says otherwise

The hub, measured against Black Seed Oil’s

Which innovations produced measurable results

9Data quality, limitations and the reconciliation audit

What was re-pulled, what reconciles, what does not, and what changed against the 8 October build.

Reconciliation and QA table — previous value, updated value, verification

What changed from the 8 October build

Blind spots and traps this report is exposed to

Structural blind spots

10Wins, failures and what to change before the next launch

Every finding carries its evidence and its status against the 8 October build. Severity is the triage.

What worked

What did not work

Operational execution

The five highest-priority changes for the next launch

Also worth doing, cheaply

• Stand the countdown pop-up up with the arc, not two days before launch; it converted at 14.3%.
• Answer the sample-size question in public: the SAMPLE code already exists.
• Say that the coffees stack. Two existing customers asked whether to stop the cocoa.
• Keep the later SMS wave and drop the noon one if a day runs two: the evening wave won on all three days.
• Reserve one article handle per send; two live articles still share a title.

11Method, sources and extraction timestamps

Builders in automation/analysis/launch-retro/: pulls/ (the five pull scripts), prep.py (derives findings.json), build_retro.py (renders this page). The 8 October builders and data are kept beside them as *_2026-10-08.*. Page source: reports/protein-coffee-launch-retro.html. All read-only; no campaign, list, profile or order was changed.

Formulas

• RPR = Klaviyo revenue_per_recipient = attributed GROSS ÷ delivered (verified: every row cross-foots).
• CVR = attributed orders ÷ delivered.
• Store GROSS = Σ total_price; store NET = Σ (line price × qty) − total_discounts − refund line items (ex tax); coffee NET = Σ coffee line price × qty − its discount_allocations.
• Identity A: line revenue − order discounts + tax + shipping = total_price, per order. Identity B: total_price − refund lines incl. tax − adjustments = current_total_price.
• New customer = customer record created within 24 h of the order.
• Credits = Σ $usage_amount × count on Relayed Text Message events whose $message is the campaign id, in a window from the day before to three days after the send.
• Ask-to-buy = sampled non-brand comments matching the purchase-intent pattern ÷ sampled non-brand comments; question = top-level comment containing “?”; answered = has a reply.
• Opt-ins per 1,000 = RunR7E joins on the send date ÷ email delivered × 1,000.

Companion reports

• Launch history — 17 arcs, the per-beat benchmark map.
• Launch blog engagement KPIs
• Opt-in routing audit
• SMS reachability — the 8,207 consented-but-unreachable finding.
• SMS credit audit — where the credit ledger method comes from.
• Protein Coffee launch hub · Black Seed Oil briefs hub