Case studies Specialty coffee subscription
Diagnosis & Plan · subscription e-commerce · details changed
Buying time to find a model that pays.
A product its customers loved, losing $2.1M a year. What the numbers showed, what we recommended, and what the plan projects.
Based on a real analysis. The product, the numbers and the timing are changed to protect the brand; the diagnosis and the plan are ours. Outcomes shown are projections, not results.
01What the numbers showed
The read
A specialty coffee subscription with a product its customers loved: four in ten subscribers had stayed a year or more. And about $4.8M in sales a year, losing $2.1M.
The ads weren't broken; the account even responded well to cost caps. The model was: a first box that cost more to win than it would ever return.
▲ VerdictNot with today's model. With no changes, losses would pass the owner's $750K limit in month 8.
The offer gave too much away, then looked expensive
First box
$18
two bags and a brewer, worth $62
Every renewal
$34
for the same two bags
Renew once
~40%
a new starter box costs less than renewing
02The plan, first
Stop the bleeding
Cap spend at payback and fix the leaks, so the business stays inside the owner's limit whatever happens next.
- Acquisition capped at what a customer pays backcap ~$22
- A third of paid clicks never loaded the page: tracking checked firstfunnel
- The welcome offer was landing in spamemail
- A new single-origin range launched to existing customers firstlaunch
Projected year-one loss
−$1.40M−$750K
At the limit, not past it, even if every test fails.
03Decision dates, not hope
Buy time
Cuts alone hold the line for a year, but the customer base shrinks and year two gets worse.
So the plan set two dates when the owner would know, and what each answer would mean.
Months from the start of the plan
Month 3
Does a one-time offer pay for itself?
Yes: scale it under its cap. No: stay on the floor.
Month 6
Commit, or change the conversation.
If nothing works: funding or the shape of the business. The owner's call, with the numbers.
04Find an offer that pays
Turn it around
Strangers buy before they commit. So lead with an offer that pays back on the first purchase, then sell the subscription to people who already like the coffee.
Not a promise: a test, capped so that even failure stays inside the limit.
- Lead ads with a one-time tasting set, priced to pay back on the first order~$64
- Spend capped at that first order's contribution~$38 a customer
- Offer the subscription by email as the coffee runs out, to buyers who already like itabout 5 weeks in
- Priced tests first, each capped at its own break-even3 versions
Each new customer, first 12 months, projected
−$13+$24
If the offer works, the year-one loss falls to about $240K on $5.0M in sales: $1.16M less than doing nothing. One number decides it: what a tasting-set buyer costs. 10% better and year two is profitable; 20% worse and it isn't. Every case beats cutting alone.
Year two:Do nothing−$1.31MStop the losses only−$1.02MStop, then the new offer−$0.17M
Every future still loses money in year one. The plan's job was to stay inside the owner's limit while finding out whether a model that pays exists.
Find out what's holding it back.
20 minutes, free, no pitch deck. Bring your numbers or your worry: you'll leave with an honest read and one thing to do next. If we can't help, or you don't need us, we'll say so.
How it works
- 01Pick a timeAny 20-minute slot that suits you.
- 02Four quick questionsYour revenue, where you sell, what worries you most and, if you have it, last year's P&L, so the call starts with your numbers.
- 03A call with Hamed, the founderAn honest read and one thing to do next. No pitch deck.