Work01 · Direct-to-consumer productTangible product
Anvil
Selling a tangible product without the race to the bottom
A small-batch coffee roaster competing against grocery-shelf prices and Amazon. A shop that sells the craft instead of the gram, a subscription that builds recurring revenue, and an agent that plays expert barista so nobody buys the wrong bag and walks away disappointed.
Try the product
Ask the roast-finder
Tell it how you brew and what you like — it recommends a bag, the way the real shop agent would.
A working agent, running live — not a mockup
Anvil
Roast-finder · live demo
ANVIL
Forge — dark roast
GRIND
$19.00
$16.15
Espresso · every 2 weeks, skip anytime
Interactive prototype — click around, it actually works
Strategy deliverable
The journey we mapped before designing anything.
Every stage from first spark to loyal return — what the customer does, feels, and exactly where the experience leaks. The dip in the curve is the moment we built the work around.
Customer journey map · concept
Marcus, 29 — finding their coffee
A home-brewing hobbyist tired of supermarket beans but intimidated by specialty coffee.
Customer journey map for Marcus, 29 (finding their coffee), across 6 stages. Sentiment dips lowest at the Consideration stage (the biggest drop-off) and peaks at the First brew stage.
- Awareness: Tired of the supermarket bag. The customer Decides he wants genuinely better beans, unsure where to look. Pain point: The world of single-origins feels like a club he wasn't invited to. Our move: A warm, unsnobby brand that meets a curious beginner exactly at their level.
- Consideration: Twelve bags that all sound the same. The customer Stares at a product grid and can't decide which to risk. Pain point: Choice paralysis and taste anxiety quietly kill the first order at the grid. Our move: Taste-first merchandising and an agent that recommends one bag with a barista's confidence.
- Purchase: Okay — the chocolatey one. The customer Adds a bag and meets the subscribe-and-save offer at checkout. Pain point: One-off buyers rarely return; the relationship resets to zero each month. Our move: The subscription is framed as the obvious default, not a pushy upsell.
- First brew: Oh. That's different.. The customer Brews his first cup — and it's genuinely, noticeably better. Pain point: A great product gets undersold by a flat, forgettable unboxing. Our move: Considered packaging and a brew tip make the very first cup a clear win.
- Reorder: Never runs out. The customer The next bag arrives before he's scraping the bottom of this one. Pain point: Manual reordering is exactly where direct-to-consumer loses people. Our move: A cadence he controls, so reordering never becomes a decision he can forget.
- Advocacy: You have to try this. The customer Gifts a bag to a friend and posts the packaging. Pain point: Most roasters never make the ask, so happy customers stay quiet. Our move: A gift and referral nudge turn a regular into a word-of-mouth channel.
Scroll the map sideways →
Engagement
Discovery + Sprint
Timeline
4 weeks
Surfaces
Shop, product pages, subscription, agent
Agent role
Guided product recommendation
The problem
A tangible product with no brand competes on price alone — and a small roaster can never win that fight against a supermarket. Anvil's coffee was genuinely better, but their site was a bare product grid that gave a shopper no reason to pay more and no help choosing between twelve similar-sounding bags. Choice paralysis sent first-timers away, and one-off purchases meant starting from zero every month.
The approach
We ran a behavioral analysis of specialty coffee buyers and found the real barriers weren't price — they were taste anxiety ('what if I don't like it?') and decision fatigue. So we merchandised around flavor and confidence, not origin tables and specs only a nerd would parse.
The brand and packaging system does the heavy lifting a small team can't: it makes a $19 bag feel like a considered object rather than a commodity, on the shelf and in the unboxing.
The shop makes the subscription the obvious default — recurring revenue is what lets a roaster survive — and the roast-finder agent removes the wrong-bag fear by recommending with a human's confidence and an expert's knowledge.
Key moves
01
Taste, not specs
Every bag leads with flavor and feeling — 'dark chocolate, low acidity, forgiving' — with the technical detail available but never in the way.
02
Subscription as the hero
The default purchase is a subscription, priced and framed so it's the obvious choice. One decision becomes twelve months of revenue.
03
An agent that removes the wrong-bag fear
A few friendly questions — how you brew, what you like — and it matches you to a bag with the confidence of a barista you trust. Fewer abandoned carts, fewer disappointed first orders.
Design targets
- first storefront, not an afterthought
- Subscription
- dead-end product pages
- 0
- from unsure to the right bag
- < 1 min
A self-initiated concept study. Anvil is a fictional roaster; the products and prices are illustrative, and the agent below is a working demonstration.
Next study · 02 Generative content tool
Plume →