Prime Day 2026 — what to do, what to drop

A read on Protekt's 2025 Prime Day performance using Business Reports for sales, the CLV cohort export for customer mix, and the 23-month cohort retention curves for downstream value. The goal is a clear answer on what to run in 2026 — and at what depth.

01 — Executive summary

The headline: Prime Day works. The 35% SNS coupon doesn't.

Prime Day 2025 lifted topline by 57% on the event week, brought in 49 net-new customers above baseline, and acquired them at a CAC only 7% above normal. The cohort projects to a healthy $115 LTV at month 23 — above the mature baseline. But the 35% SNS coupon that ran alongside the 20% PED was almost entirely wasted spend.

✓ Run again in 2026

20% PED on hero ASINs

Hydration + 30-pack · Immune Support · Lions Mane

This was the active driver of Prime Day 2025 sales. With ~$9,271 of unredeemed SNS budget freed up, consider adding Sleep ASINs to the PED list this year — they were excluded in 2025 and we'd expect comparable lift given Sleep's emergence as a category leader through Q4 / Q1.

✗ Don't run again

35% SNS coupon

15 products · $10,000 committed · $729 redeemed

7.3% of the committed budget redeemed. Only 37 units sold through the coupon over four days — about 4.2% of total Prime Day revenue. The 20% PED was already pulling existing-customer stock-up demand. The deeper SNS coupon was redundant — same buyers, deeper discount, almost no incremental conversion.

02 — How the data ties together

Two sources, one view of the customer

Business Reports gives us sales. The CLV file gives us repeat customer % per week. Multiplying those gives us NTB sales — the part of Prime Day revenue that came from genuinely new customers, vs existing customers stocking up.

Metric	Baseline week	Prime Day week	Δ	Source
Topline (Business Reports)
Total ordered product sales (B2C)	$10,984.42	$17,287.18	+57.4%	BR direct
Units ordered	418	686	+64.1%	BR direct
Sessions	1,448	2,602	+79.7%	BR direct
Buy Box %	88.6%	97.0%	+8.3 ppt	BR direct
Customer mix (BR sales × CLV repeat %)
Repeat customer %	29.6%	38.8%	+9.2 ppt	CLV column G
NTB sales (BR × NTB %)	$7,805.85	$10,574.57	+35.5%	derived
Repeat-customer sales (BR × repeat %)	$3,703.05	$6,712.61	+81.3%	derived
Acquisition economics (CLV)
NTB customers	66	115	+74.2%	CLV
Ad spend	$1,443.96	$2,583.84	+78.9%	CLV
CAC ($/NTB customer)	$20.96	$22.47	+7.2%	ad spend / NTB cust
CAC as % of NTB sales	19.0%	24.4%	+5.4 ppt	ad spend / NTB sales

03 — Sales lift, decomposed

Where the +57% came from

Total Prime Day sales beat baseline by $6,303. Of that, repeat customers contributed more incremental dollars than new customers did.

NTB sales lift

+$2,769

From $7,806 → $10,575

49 more new customers showed up vs a typical week. NTB sales grew 35% — the genuinely incremental piece.

Repeat sales lift

+$3,010

From $3,703 → $6,713

45 more existing customers came back. Repeat sales grew 81% — the stock-up effect, partly pulled forward from later weeks.

Total sales lift

+$6,303

From $10,984 → $17,287

Repeat customers contributed 48% of the incremental dollars. New customers contributed 44%.

What this tells us

Prime Day is doing two things at once: bringing in net-new customers and giving existing customers a reason to stock up. The repeat-customer surge is real demand — existing buyers re-engage when there's a reason to. That's part of why brand defense matters during these windows: if Protekt isn't visible, those same customers go to Liquid IV or Element instead.

04 — CAC on the new customers

What it cost to acquire the 115 new customers

Ad spend went from $1,443.96 (typical week) to $2,583.84 during Prime Day — a 79% increase. CAC per new customer barely moved.

CAC per NTB customer

$22.47

Baseline $20.96 · only +7.2%

CAC as % of NTB sales

24.4%

Baseline 19.0% · +5.4 ppt

NTB AOV

$91.95

Baseline $118.27 · -22.3%

Incremental NTB customers

+49

Above what we'd get on a typical week

CAC verdict

CAC per NTB customer barely moved (+7%) despite 1.79× the ad spend. The reason: NTB customer count almost kept pace with the spend increase (115 vs baseline 66, +74%). The slight drop in NTB AOV (smaller first-order baskets, typical of discount-driven first purchase) is the only real efficiency cost. This is a healthy acquisition window.

05 — Weekly trend

The Prime Day spike in context

Sales, NTB sales, and CAC across the full 38-week dataset. The Jul 6 event week is annotated.

Weekly sales — total vs NTB vs repeat

Stacked decomposition shows how much of each week's sales came from new vs returning customers

NTB sales (new customers)

Repeat sales (returning customers)

Baseline median total sales

CAC per NTB customer, by week

$ ad spend per genuinely new customer acquired. Lower is better.

CAC ($/NTB customer)

Baseline median CAC

06 — 23-month cohort tail

What these new customers are worth long-term

The Jul 2025 acquisition cohort (which contains the Prime Day NTB customers) is now 8 months mature. Comparing its retention curve against fully-mature baseline cohorts (Feb 2023 – Apr 2024, all with 23+ months of data) lets us project where it's heading.

Cumulative LTV per customer — 23 months

Built from each cohort's actual retention curve × baseline avg revenue per repeat purchase ($30.65). Projection from M9 onward uses baseline post-M8 slope.

Baseline (15 mature cohorts, Feb 2023 – Apr 2024)

Jul 2025 cohort (Prime Day) — observed M0–M8

Jul 2025 cohort — projected M9–M23

Baseline M23 LTV

$104.92

Average customer value across 15 mature cohorts at month 23.

Jul 2025 projected M23 LTV

$115.72

Above baseline. Prime Day cohort starts higher (M0 $71.54 vs $57.96) and tracks parallel.

Incremental cohort value

$5,670

49 extra NTB customers × $115.72 M23 LTV. Long-term return on the +$1,140 incremental ad spend.

07 — The lever question

SNS coupon vs PED vs leave it alone — what worked in 2025

Protekt ran two stacked promos during Prime Day 2025: a 20% PED on hero ASINs and a 35% SNS coupon on 15 products. Here's what each one actually did.

20% PED · Hydration / Immune / Lions Mane

The active driver

The PED was visible to all Prime members, drove the +57% topline lift, and acquired customers at a healthy CAC. This is the lever.

35% SNS coupon · 15 products

Underperformed

The 20% PED was already pulling existing-customer demand. Stacking a 35% SNS coupon on top discounted the same buyers further without adding meaningful new conversions.

The "double-discount" problem

When PED and SNS coupons run simultaneously on overlapping ASINs, existing SNS subscribers are already getting their normal 5–15% S&S discount, plus the 20% PED visible to all Prime members, plus the 35% SNS coupon. That's three layers of discount on the same purchase. The data suggests we hit a point of diminishing returns — incremental conversion from the third layer was effectively zero.

08 — Recommendation for Prime Day 2026

What to run, what to drop, what to test

✓ Keep

20% PED · expanded

Rest/Sleep ASINs added · Hydration, Immune, Lions Mane retained

The 20% PED was the active driver in 2025. The Sleep / Rest category has emerged as a category lead since (March REST marketing month, M205% MoM growth on GABA L-Theanine). Adding Sleep to the PED list captures the category we excluded last year.

✗ Drop

35% SNS coupon

Replace with: leave SNS at typical 5–10% subscriber discount

The 35% SNS coupon delivered 4.2% of Prime Day revenue against $$10,000 committed. Drop it. The standard always-on S&S discount (5–15% by ASIN) is enough — especially when stacked under a 20% PED.

? Test

Targeted 15% SNS coupon — Sleep only

If Sleep is the category we're trying to convert into recurring revenue

Sleep buyers retained worse than Hydration in early data. A modest SNS coupon (10–15%, not 35%) only on Sleep ASINs could help convert first-time Prime Day Sleep buyers into subscribers — addressing the retention gap we'd otherwise see in the Jul 2026 cohort.

Bottom line

Run Prime Day 2026 with the 20% PED expanded to include Sleep. Drop the 35% SNS coupon entirely. If we want to push S&S subscription conversion on Sleep specifically, run a targeted 10–15% SNS coupon only on those ASINs — not the brand-wide 35% from 2025. Expected outcome: similar topline lift to 2025 (~$6–7K incremental), broader hero ASIN coverage, and the $$9,271 of unredeemed SNS budget gets reallocated to ad spend during the event window — which is where it converts.

09 — Next steps

What to do before June

Confirm 2026 Prime Day date with Quartile and lock the PED submission window. Amazon's PED submission window typically opens 8–10 weeks before the event. Submit Hydration parent + 30-pack, Immune, Lions Mane, Sleep parent + 30-pack at 20% off.
By early June
Drop the 35% SNS coupon. Decide on the targeted Sleep-only coupon test. Either run no SNS coupon, or run a 10–15% Sleep-only test. Don't run brand-wide 35% again.
By early June
Reallocate the SNS budget into ad spend during the event window. The $$9,271 we previously committed to SNS coupon depth is better spent on Sponsored Brand video, Sponsored Brand collections (now live for Hydration), and DSP retargeting during the 4-day event.
June planning with Scott / Quartile
Confirm inventory coverage on all PED'd ASINs through end of July. The 2025 Jul cohort was 658 customers. If the 2026 cohort is similar or larger, we need 4–6 weeks of buffer post-event to capture the halo demand without OOS gaps.
June with Gus on the depletion side
Set up post-event cohort tracking on the Jul 2026 cohort. Same retention measurement we used here. By Q4 we'll know whether expanding PED to Sleep moved the cohort retention curve up vs the 2025 baseline.
August onward