Your leak is two gaps. We price both, then seal them.
Keep scrolling — the ledger builds as you read. You draw 10–15k visitors a month stated, but only ~2.2% buy — well under the bracket.
Close the conversion gap toward the 4.6% median benchmark and it alone is worth $3,236/mo derived. Incremental math uses the $30 single-bag AOV; blended is ≈ $39 with bundles.
Add lifecycle email — $526/mo derived — and the leak totals $3,762/mo, about 35% of your ~$10,750 revenue. Messaging, ads, social and the quiz push traffic to convert; email compounds it.
Why ads still wait a month: bought traffic landing on a silent hero, measured by no pixel, burns the margin that keeps it safe. Seal first — hero, DMARC, pixel — then ignite. That ordering is the plan.
Monday, in three moves — your first deliverable
What your rivals say instead — checked on their homepages measured
HARBOR ROASTERS
“Single-origin, small-batch”
Origin plus process — nothing a shopper can taste at checkout.
MERIDIAN COFFEE
“Coffee, elevated”
A vibe, not a reason to choose.
DUSK & DAWN
“Wake up better”
Aspiration with no mechanism behind it.
Button: “Taste the difference”. Voice rule: talk like a roaster who cares when you brew, not a brand that sells beans.
One idea, six rooms
This sentence now runs everything below.
Voice rule, everywhere it travels: talk like a roaster who cares when you brew, not a brand that sells beans.
Your ads need a front door — a brew quiz, five questions in
You didn’t brief a funnel — we built one anyway, because “roasted for how you brew” begs a question only the shopper can answer. An ad pointing at a quiz converts better than an ad pointing at a homepage; all four ads at stop 07 link straight here.
For your developer — the full build brief
Q1 how you brew · Q2 roast level · Q3 milk or black · Q4 how much you drink · Q5 whole bean or ground.
Results map to a roast + grind and one bag, in the roaster’s voice from stop 02. Page, logic table, and the result copy for each branch are included.
Five fixes, pinned where a buyer meets them
Homepage, category, product — walked in buying order measured. Every fix specified; the biggest costs ten minutes.
The six-slot image order — and what we will not ask you to test
1
Bag hero — real bag, real roast color
KEEP2
Roast-date stamp, on-image — “roasted 3 days ago”
MISSING3
Beans close-up — the roast, up close
MISSING4
Testimonial — “smooth, never bitter”
MISSING5
Brew shot — a cup being poured
KEEP6
Origin / the farm it came from
MISSINGNo split tests. At your volume a trusted A/B test still needs months per variant to read clean — anyone selling you a testing program first is selling a coin toss with a dashboard. Change the hero, keep it, judge revenue over the quarter.
Your core flows are live. So we built the layer above them.
We checked your account before writing a word measured — no flow-building here. What you lack: deliverability, segments, and a monthly campaign engine.
PASS
SPF
Published and correct. Nothing to change.
PASS
DKIM
Signing correctly on every send. Nothing to change here.
FAIL
DMARC
No record published. Nothing tells inboxes what to do with mail that fails your checks — so some of it silently never arrives.
1. At your DNS host, add a TXT record named _dmarc.northbean.co 2. Value: v=DMARC1; p=none; rua=mailto:dmarc@northbean.co; fo=1 3. Leave p=none for two weeks and read the aggregate reports — Then raise to p=quarantine, and later p=reject, once your legitimate mail is passing clean.
Click any of the eight to read it — each ships with its own art, generated from your store’s URL and look. “Copy HTML” puts the send, image included, on your clipboard ready for your ESP.
Checked live — Meta Ad Library measured
search: “specialty coffee” · active ads · country US → 41 results — every one on price or origin. Nobody argues freshness-to-door.
A category giant runs 100+ active ads on generic beans — the category prints money. The freshness-first argument has no owner yet. benchmark
The 8-hook library, with the pattern behind each
| Pattern | Hook |
|---|---|
| Callout the staleness | If the bag won’t tell you when it was roasted, it’s already stale. |
| Taste letdown | If it needs sugar to go down, it wasn’t roasted for you. |
| Cost-of-cheap math | A $9 supermarket bag is stale before you open it. The real cost per cup. |
| Bitterness myth | If your coffee needs sugar, it’s the roast — not you. |
| Insider mechanism | Roast date is the number nobody prints. Here’s why days-off-roast is everything. |
| Identity mirror | You bought the $900 machine — then feed it months-old beans? |
| Objection flip | The “expensive” bag is the cheap one — once you count the cups you actually enjoy. |
| Competitor negative space | Their label tells you the origin. It never tells you the roast date. |
Retargeting and destination links, ready to use
Four in ten of your customers come back — the second look is the cheapest revenue in the whole ad account.
| Audience | Window | Suggested topic |
|---|---|---|
| Site visitors | 30 days | The freshness proof — their own review language |
| Video viewers | 14 days | The brew-quiz angle → five questions to your roast |
| Past buyers | 60 days | The tasting-flight bundle (store fix #3) |
northbean.co/pages/roasted-this-week?utm_source=meta&utm_medium=paid&utm_campaign=freshness-launch&utm_content=ad1-roastdate
Hold the budget until the store can see
Your site runs no Meta pixel today — we looked. Of every $30 single-bag order, about 45% is actually yours after beans and shipping — roughly $13; spend past that on one purchase and the sale loses money. Flipped: 1 ÷ 45% = 2.22× break-even — the number the KILL rules run on. Install the pixel first (30 minutes). We’d rather you spent nothing than spent it blind.
Send back five numbers. We’ll tell you what won.
Not a document that ends — the first turn of a loop. Ten minutes of your numbers; the read back is free.
The next batch built on it — new ads off the winning angle, the next month’s emails and theme — is the next plan, at the same $497. No subscription, nothing charged unless you ask.
Priced as five separate briefs
Judge the price like everything else here: against evidence. Scroll — the market quotes stack up, then yours lands.
Things only someone who looked would know
Your DNS publishes SPF and DKIM but no DMARC record at all. Your core flows are live. Your reviewers keep writing the same two words — smooth and fresh. None of that came from a template.
And one judgment call, owned: no revenue forecast on this page — a number nobody can defend costs more trust than it buys. The results section is day 30, measured, not page one, promised. — Itai, The Grow Program
Everything you own, on one screen
Each piece forwards on its own — the DNS block to whoever holds your domain, the quiz brief to your developer, the ads to whoever runs the account. Nobody has to ask you a single question.
Screenshot this — the plan on one card
Diagnosis: great coffee, a leaky funnel — conversion under bracket, email thin (~$3,762/mo).
The one message: roasted the week it ships — smooth, never bitter, never stale.
Monday: paste the hero · publish DMARC · install the pixel.
Day 30: send five numbers, learn what won — free.
Did this change how you see your store?
One tap. Nothing follows this.
Template demo — northbean.co is a fictional store; every figure and screenshot on this page is sample data showing what your plan looks like. Figures marked benchmark are category averages. Prices in the comparison are going market rates for each piece hired separately. The Grow Program · $497 per turn of the loop.
This is the demo. Yours runs on your numbers.
Same ten stops, built on your real store — the leak priced, the biggest lever first. Six questions, about two minutes, no card.
Start the free audit →northbean.co is a fictional demo store; every figure on this page is illustrative and does not describe any real business. Plan figures for real stores are estimates from provided ranges and curated category benchmarks — not guarantees. Results vary by store and execution. Terms & disclaimers.
Eight posts, one monthly theme, from what already works
We read your feed, your reviews and your rivals’ pages measured. One gap; month one’s theme — Roasted This Week — closes it.
Your feed leads with
Product flat-lays
Your buyers’ reviews lead with
Freshness & smooth, never bitter
Reel — week 1 · Freshness math
Beans die weeks after roast. Most shops won’t print the date — so the guess comes back in a flat, bitter cup.
Talking-head reel: one bag, the roast-date stamp shown, a fresh cupping at home. Same argument your welcome email opens with.
Reel — week 2 · From the roastery
We roast to order in small batches. Here’s what that lets us change when a customer says it tasted flat.
Founder on camera at the roaster. Nobody drop-shipping can film this.
The other six, with pillar and hook
Two of these hooks run again as paid ads — deliberate: organic is the free test, paid scales the winner. At day 30, engagement picks next month’s theme.