Growth plan for Harney & Sons, 2 Oct 2026
Scale spend. Hold CAC.
One number to protect: a $15.36 target CAC. Harney & Sons already has proof, such as 87 reviews on the Holiday Duo Gift, and over 300 blends to test. The plan turns that into creator videos, landing pages and tested ads.
- Target CAC
- $15.36
- New ads a week at peak
- 20
- Creators live by week six
- 10
- Test spend, six weeks
- $24,000
Contents
10 chapters, then the gates and every number’s source.
01 The number
Hold CAC at $15.36 on tins that sell for $18.50
Target CAC is $15.36. It comes from a ceiling of $25.60: a $32 average order, a 40% margin and one repeat order, then 0.6 of that as a guard line. The $32 and the margin are my inputs until week one replaces them with your numbers.
Protect
Target CAC: $15.36 on a first purchase
Ceiling = average order × margin × (1 + repeat orders): $32 × 40% × (1 + 1) = $25.60. Guard line = 0.6 × $25.60 = $15.36. Across the ranges: $2.63 to $324.00.
Three moves
- Run tests on the six-week ramp and kill losers early, one variable per test, so spend only follows concepts that hold the line.
- Put proof next to the offer: the Holiday Duo Gift's 87 reviews and the $59+ free-shipping line go into the first landing pages.
- Report daily CAC against the $15.36 guard line and cut any concept that drifts above it.
- reviews on the Holiday Duo Gift
- 87
- Published
- reviews counted on one page of the site
- 14,058
- Published
- tea blends, gifts, treats and teaware in the home description
- 300+
- Published
02 Variety
Over 300 blends give every ad a different reason to buy
Each ad concept has to carry one reason to buy: a cup, a gift, a price or a proof point. Harney & Sons has the range for it, from an $18.50 tin to a $125.95 collection, so no two ads need to say the same thing.
| Reason to buy | Proof you already have | Ads (proposed) |
|---|---|---|
| Over 300 tea blends | Shop over 300 tea blends, tea gifts, treats and teaware online from Harney & Sons Tea. | 5 |
| One sachet, one cup | Each sachet brews one 12 oz. cup of tea. | 4 |
| Tins you can refill | Be sure to recycle your tin, or reuse it and refill it with one of our Bags of 50! | 3 |
| Ready-made host gifts | Perfect as a thank-you or host/hostess gift | 3 |
| Free shipping at $59+ | Free Shipping on Orders $59+ | 2 |
| A family company since 1983 | the company founded by their father, John Harney, in 1983. | 1 |
| Total | The ad concepts tab | 18 |
Each row is a reason to buy that the brand already has proof for. The count is how many of the ads carry it: a reason with a single ad is a bet, not a pattern, and the next batch of concepts moves toward the reasons that win.
03 Disturbances
Seven risks, from lower 48 seltzer shipping to holiday timing
| Disturbance | Likelihood | Impact | Owner | Gate (proposed) |
|---|---|---|---|---|
| Seltzer cases ship only to the lower 48 states, so ads can reach buyers the store cannot serve | Med | Med | Your team | Exclude locations outside the lower 48 before any seltzer ad goes live |
| Single tins at $18.50 sit below the $59+ free-shipping line, so first orders may be small | High | Med | Both | Day 14: replace the $32 input with the real average order; rebuild bundles if it is lower |
| The Holiday Duo Gift leads the first tests, but holiday sets have a season | Med | Med | Me | Move spend to tins and bundles if holiday ad CAC passes $25.60 for a week |
| Returns run 30 days with no return labels, which can add service load from new buyers | Low | Med | Your team | The claims sheet states the 30-day return rule before creators go live |
| Creator videos drift into claims the brand has not published about its teas | Med | High | Me | Every script is checked against the claims sheet before posting; no unapproved health claims |
| I have no tea experience, so early hooks may miss how tea drinkers talk | Med | Med | Both | Your team approves the hook library before the week 2 creators go live |
| Event tracking may be incomplete and hide which orders came from which ad | Med | High | Your team | Daily CAC must reconcile with store orders before spend moves past the week 2 ramp |
Scores are my read from outside the company, re-scored in week one with your data. Each risk has one owner and one gate that can be checked with a number, so nobody has to argue about whether it happened.
04 The ceiling
The $25.60 ceiling decides which tins and sets can scale
| Line | Value | Status |
|---|---|---|
| Average order | $32 (range $5.00–$240.00) | Assumption Range from the site product prices (Published); the midpoint is a guess |
| Gross margin | 40% (range 35–45%) | Assumption Replace in week one |
| Repeat orders after the first | 1 (range 0.5–2) | Assumption Replace with cohort data in week one |
| Margin per customer, all orders | $25.60 | Calculated Average order × margin × (1 + repeat orders) |
| Guard line for CAC | $15.36 | Calculated 0.6 × the margin per customer |
The math, with the assumed lines
Ceiling = average order × margin × (1 + repeat orders): $32 × 40% × (1 + 1) = $25.60. Guard line = 0.6 × $25.60 = $15.36. Across the ranges: $2.63 to $324.00.
Range across the assumptions: $3 to $324.
The ceiling is a formula, not a fact: a $32 order, a 40% margin and one repeat order are my inputs. Week one replaces them with your real numbers, and the $15.36 guard line moves with them.
The ceiling is a range, not one number: the most a new customer can cost before shipping and packaging, which I do not have. Replace the guessed lines with yours and every figure below recalculates.
05 Test ramp
Six weeks of tests, $24,000, before any winner gets more
| Week | New ads × budget / ad (proposed) | Weekly test spend | Cumulative | Creators posting | CAC target (proposed) |
|---|---|---|---|---|---|
| 1 | 12 × $250 | $3,000 | $3,000 | 0 | $15 |
| 2 | 12 × $250 | $3,000 | $6,000 | 2 | $15 |
| 3 | 12–20 × $250 | $4,000 | $10,000 | 4 | $15 |
| 4 | 12–20 × $250 | $4,000 | $14,000 | 6 | $15 |
| 5 | 20 × $250 | $5,000 | $19,000 | 8 | $15 |
| 6 | 20 × $250 | $5,000 | $24,000 | 10 | $15 |
Weeks 1 and 2 run 12 ads at $250 each, $3,000 a week. Weeks 3 and 4 run 12–20 ads, $4,000 a week. Weeks 5 and 6 run 20 ads, $5,000 a week. Total $24,000 of tests, Proposed, with a hook library growing the whole time.
The math. Week 1: 12 × $250 = $3,000; Week 2: 12 × $250 = $3,000; Week 3: 16 × $250 = $4,000; Week 4: 16 × $250 = $4,000; Week 5: 20 × $250 = $5,000; Week 6: 20 × $250 = $5,000. Total $24,000 (96 ads × $250). A range such as 12–20 counts as its midpoint.
Weeks one and two test what the brand already has, rewritten per reason to buy. From week three, creator videos enter paid only after they beat their own account median. A range of new ads counts as its midpoint in the spend column.
06 Volume
From 20 to 80 concepts: more winners out of over 300 blends
20 concepts give 2 winners and $18,000 added a month; 80 give 8 winners and $72,000. Hit rate and spend per winner are an Assumption. More concepts out of over 300 blends means more shots at a low CAC.
| Concepts tested / month | Hit rate (assumed) | New winners / month | Spend each winner holds (assumed) | Added spend at target CAC |
|---|---|---|---|---|
| 20 | 10% | 2 | $300 / day | $18,000 / month |
| 40 | 10% | 4 | $300 / day | $36,000 / month |
| 60 | 10% | 6 | $300 / day | $54,000 / month |
| 80 | 10% | 8 | $300 / day | $72,000 / month |
The math. 20 concepts × 10% = 2 winners × $300 × 30 days = $18,000 a month; 40 concepts × 10% = 4 winners × $300 × 30 days = $36,000 a month; 60 concepts × 10% = 6 winners × $300 × 30 days = $54,000 a month; 80 concepts × 10% = 8 winners × $300 × 30 days = $72,000 a month.
The hit rate and the spend each winner holds are placeholders. The first thirty days of tests replace them with real numbers, and the table is recalculated the same day.
07 Allocation
$100,000 split so gift sets and tins get tested first
- Creator content for tins, gift sets and matcha kits
- $45,000
- 45%
- Paid tests on Meta, the six-week ramp and after
- $30,000
- 30%
- Landing pages for gift sets and bundles
- $15,000
- 15%
- Reporting, tracking checks and reserve
- $10,000
- 10%
Of the $100,000 Proposed budget, the largest share goes to creator content, because creators feed every ad and every landing page.
The math. 45% × $100,000 = $45,000; 30% × $100,000 = $30,000; 15% × $100,000 = $15,000; 10% × $100,000 = $10,000. Sum 100% = $100,000.
This split is a proposal. It moves toward scaling as winners prove out, and toward testing when fatigue shows.
08 Channels
Meta first, then where tea gifts and matcha kits get found
| Channel | Open when (proposed) | Why wait |
|---|---|---|
| Meta (Facebook and Instagram) | Week 1, with the first 12 ads | Fastest read on which hooks sell a tin, a gift set or a trio. |
| TikTok | Once creators are live, from week 2 | Brewing one 12 oz. cup on camera suits short vertical video. |
| YouTube Shorts | When a hook holds CAC under $15.36 on Meta | Reuses winning creator videos without new shoots. |
| Google Search | After the gift set landing pages are live | People searching for tea gifts or matcha kits are already close to buying. |
| Email to past buyers | When tracking shows repeat orders | The $25.60 ceiling assumes one repeat order, so repeat orders carry the plan. |
A channel opens when its condition is true, not when the calendar says so. Until then the budget stays where the CAC is earned.
09 Payback
Payback decides: a $30 CAC never pays back on a $32 order
Payback is the real constraint. At a CAC of $10 the first order pays back, with $15.60 left. At $15 it pays back after repeat orders, $10.60 left. At $30 and $35 it never does: stop, with −$4.40 and −$9.40 left. That is why $15.36 is the number.
Cumulative margin per customer, order by order, before CAC: $12.80, $25.60.
| CAC (scenario) | First-order margin | Year-one margin | Left after CAC | Payback |
|---|---|---|---|---|
| $10 | $12.80 | $25.60 | $15.60 | First order |
| $15 | $12.80 | $25.60 | $10.60 | After repeat orders |
| $30 | $12.80 | $25.60 | −$4.40 | Never: stop |
| $35 | $12.80 | $25.60 | −$9.40 | Never: stop |
The math. CAC $10: $25.60 − $10 = $15.60 left; pays back: First order; CAC $15: $25.60 − $15 = $10.60 left; pays back: After repeat orders; CAC $30: $25.60 − $30 = −$4.40 left; pays back: Never: stop; CAC $35: $25.60 − $35 = −$9.40 left; pays back: Never: stop.
Read each row left to right: the margin of the first purchase, the margin over the whole horizon, what is left after CAC, and the purchase on which the CAC is earned back. A scenario that never pays back is a stop, not a test.
10 Scope
Scope: ads and creators for harney.com, not the cart or tracking
Covers
- Meta ad tests, creative briefs and the hook library for harney.com
- Creator sourcing and briefs for tins, gift sets and matcha kits
- Landing pages for gift sets, tins and bundles
- Daily CAC, weekly learnings and monthly cohort payback
Doesn’t
- Event tracking build: I spec it, your team builds it
- Fulfillment, shipping rules and returns handling
- Product, pricing and claims your team has not published
- Email and site changes outside the landing pages
What this plan covers is what I can change inside the ad account and the creator program. Everything outside it is named on the right so nobody assumes it is handled.
Gates, written before spend, so stopping isn’t a negotiation.
| Day | Keep going if (proposed) | Stop or change if |
|---|---|---|
| Day 14 | Two creators live, 12 ads tested, and CAC readable against the $15.36 guard line | Tracking cannot match orders to ads |
| Day 30 | One concept holds the $15.36 guard line and creators are on pace for 10 live by week six | No concept sits under the $25.60 ceiling after the tests so far |
| Day 60 | Winners stay under $25.60 as spend rises, and repeat orders show up in cohort payback | CAC stays above $25.60 for two straight weeks |
| Day 90 | Daily CAC sits at or under $15.36 while spend scales from the $100,000 budget | CAC above $30 with no concept recovering |
Each gate is checked on its day with the numbers in the daily and weekly reports. If the stop condition is true, the spend stops and nothing renews until we talk.
What exists, what’s missing, and the order that’s forced.
| Piece | Exists today | Missing | Forced order |
|---|---|---|---|
| creators | Gift, tin and matcha products creators can show on camera | A roster and briefs | 1st |
| creative | Over 300 blends and 87 reviews on the Holiday Duo Gift | Tested hooks and ad formats | 2nd |
| claims sheet | Literal product lines, such as one 12 oz. cup per sachet | One sheet creators copy from | Week 1 |
| landing pages | Gift sets and bundles with struck-through prices | Pages matched to each ad concept | 2nd |
| reporting | A $59+ free-shipping line to measure order size against | Daily CAC tied to orders | Week 1 |
The order is forced on purpose: creative and creators come first because they are what moves CAC, and everything else waits for that data.
Every number, and where it came from.
| Figure | Where it came from | Type |
|---|---|---|
| 87 reviews on the Holiday Duo Gift | https://www.harney.com/products/holiday-duo-gift | Fact |
| 14,058 reviews counted on one page of the site | https://www.harney.com/ | Fact |
| 300+ tea blends, gifts, treats and teaware in the home description | https://www.harney.com/ | Fact |
| Product prices $5.00–$240.00 | product prices in the site product data (160 products), read 2026-10-02 | Fact |
| $32 average order · 40% margin · 1 repeat order | Placeholders, replaced in week one | Assumed |
| $15.36 target CAC | 0.6 of the $25.60 margin per customer | Calculated |
| $25.60 margin per customer | Price × (margin − discount), summed over the orders | Calculated |
| $24,000 of tests (96 ads × $250) | Danilo’s plan, matches Month one | Calculated |
| 0→10 creators in 6 weeks · 7 videos per creator a week | Danilo’s plan, matches Month one and the Creator engine | Proposed |
| $100,000 first budget split 45% / 30% / 15% / 10% | Danilo’s plan, re-set with the team in week one | Calculated |
| 10% hit rate · $300 a day per winner | Placeholders, replaced by the first 30 days of tests | Assumed |
| 303 creator videos a month at 10 creators | 10 creators × 7 videos a week × 52 ÷ 12 | Calculated |
| $5.74 per 1,000 views against a $1 target | Creator cost over the views the assumptions give | Calculated |
| $1,000 base pay · $50 bonus past 100,000 views | Danilo’s plan | Proposed |
| 1,500 median views · 4% breakout (10×) · 0.5% viral (750,000) | Placeholders, replaced by the first month of posts | Assumed |
| Pay bands $300–$600, $500–$1,000, $800–$1,500 | Danilo’s plan | Proposed |
Every figure on this page is listed here with where it came from. Fact means read on the site. Assumed and Proposed are Danilo’s own numbers, to be replaced in week one. Calculated is plain arithmetic on those, with the formula shown next to it.
Month one is where it starts.
In week one I read your product pages with your team, write the claims sheet, brief the first creators and launch 12 ads against the $15.36 guard line. We also replace the $32 guess with your real average order.
