Roast profile
SEO for coffee roasters that sells the bean, not just the bag
SEO for coffee roasters has to work twice: once for a home brewer typing a tasting note, and once for a buyer sourcing forty kilos a month. Two audiences, two page paths, one coherent site.
- Origin, process and flavour pages built on real cupping language
- Product schema and variant handling that survives filters
- Wholesale path with volumes, samples and lead times
- Subscription funnel with retention copy, not just a signup box
- Brew guides as the top of a content cluster
Roastery operators only. No reseller network, no white label.
The bean list
A catalogue people can actually search
Tasting notes are keywords. "Chocolate and nuts" is a phrase every roaster uses. "Cacao nib, dried fig, low acidity" is a phrase someone types looking for exactly that cup.
Guji, Ethiopia
Natural · 1,950 m · Filter
Stone fruit, jasmine, bergamot.
- Floral
- Bright
- Tea-like
Huila, Colombia
Washed · 1,700 m · Espresso
Red apple, cane sugar, cocoa.
- Balanced
- Sweet
- Round
Cerrado, Brazil
Pulped natural · 1,100 m · Espresso
Peanut brittle, milk chocolate.
- Nutty
- Low acid
- Heavy body
Gayo, Sumatra
Wet-hulled · 1,400 m · Filter
Cedar, dark plum, tobacco leaf.
- Earthy
- Syrupy
- Long finish
Sample layout · tasting language shown as an example, not a live catalogue
Flavour wheel · buyer vocabulary
Five axes. Every product page needs at least three.
The problem
Good coffee, unsearchable pages
The product is rarely the issue. The issue is that the site cannot be read by a machine, and the language does not match how buyers describe taste.
One page, forty variants
Grind size, weight and roast level create a faceted navigation of hundreds of near-duplicate URLs. Crawl budget drains, thin pages get filtered and the shop wonders why nothing ranks.
Handled properly, facets are controlled with canonical rules, a sensible parameter policy and a category page that earns the traffic instead of the filters.
Wholesale buried in a footer
Cafes and restaurants want volumes, sample policy, lead time, packaging and invoicing. If that is one line of text under the retail shop, the buyer assumes you do not want the account.
Subscription with no reason to stay
A signup box with no explanation of rotation, roast date, pause rules or shipping cadence converts badly and churns harder. Copy is the product here.
Origin stories told on social
Farm relationships, processing experiments and harvest news live in reels, not on the site. That is exactly the material that earns links from coffee publications and builds E-E-A-T.
No brew guidance
Dose, ratio, grind and water temperature tell a buyer whether the coffee is for them. Guides also catch the informational searches that feed the commercial pages.
Supplier photography
Green bean shots borrowed from the importer rank in image search for the importer. Original roastery photography, named and described, brings the traffic home.
The build
Six stations along the search path
Each one has a different job, and none of them is "write a blog post about coffee".
Bean and origin pages
One page per coffee, with farm, altitude, varietal, process, harvest and a full tasting description, plus dose and ratio guidance. A short internal link to the brew guide and to the subscription page closes the loop.
Process vocabulary matters more than most roasters realise: washed, natural, honey, anaerobic, wet-hulled. Buyers search those words directly and they are cheap to own.
Product schema and variants
Product markup with offers, availability, weight and roast level, an aggregate rating where genuine reviews exist, and a variant strategy that does not spawn a URL for every grind setting.
On the commerce side this overlaps heavily with Local SEO For Small Business, since catalogue structure and faceted navigation decide whether a store can rank at all.
Brew guides as a cluster
V60, espresso, aeropress, moka, cold brew. Each guide names the coffees it suits and links to those product pages, which is how an informational cluster passes authority into the catalogue instead of sitting in isolation.
Wholesale path
A dedicated page covering volumes, sample boxes, lead times, packaging formats, training and account terms, with a short qualification form. Then a second page for the trade show or cafe group you already supply.
Subscription funnel
Rotation logic, roast date promise, pause and skip rules, shipping cadence, and what happens after month three. Retention copy sits on the page, not only in the welcome email, because the decision happens before checkout.
Roastery photography and local trade
Original images of the drum, the cooling tray, the cupping table and the bag, compressed and captioned, with ImageObject markup. For roasteries with a cafe or a collection counter, the local layer runs through Ecommerce SEO Services: opening hours, collection, tasting events and map accuracy.
Wholesale flow
From a sample request to a standing order
Four steps, each with its own page and its own proof. The fastest way to lose a cafe account is to make the buyer email for basic terms.
Sample request
A short form: volume, equipment, current supplier, taste preference. No essay required.
Cupping and fit
A sample box with tasting cards and brew recipes, so the buyer can compare on their own bar.
Terms agreed
Volumes, roast schedule, packaging, delivery days and invoicing, stated in writing before the first pallet.
Standing order
Recurring schedule, seasonal rotation and a named contact, which is what keeps the account for years.
Process
Five stages, in roast order
Green audit
Index coverage, facet bloat, page speed on mobile and the current share of non-branded clicks. Baseline recorded before any change.
Cupping the demand
Origin, process, flavour and equipment terms grouped into clusters, then matched to pages that can plausibly win them.
Profile build
Copy, schema, internal links and imagery shipped in one release per cluster, so a page never sits half finished.
First crack check
Technical QA: canonical logic on filters, Core Web Vitals, duplicate collection pages, sitemap hygiene and crawl budget.
Development
Sixty-day review of queries, then expansion into the next cluster. Wholesale and retail measured separately, because they behave differently.
Results
Measured in accounts, not applause
Anonymised roastery data, monthly organic enquiries.
A specialty roaster, 6 months: organic wholesale enquiries per month after a dedicated trade page replaced a single contact form.
A subscription roaster, 5 months: organic trial signups per month with retention copy on the page.
A roastery with a collection counter, 4 months: map listing calls and direction requests per month.
Is your catalogue readable by Google?
Send the shop link and the two coffees you want to sell more of. We will list the pages missing from the path.
Scope sets the number, never the other way round
A one-person roastery with eight coffees and a subscription needs a different engagement from a wholesale operation with forty SKUs, three grind formats and a cafe. We quote after the audit: catalogue size, platform, facet complexity, whether product data is clean and how much content already exists.
- Fixed setup for catalogue architecture and schema
- Monthly work sized by SKU count and season
- Copy, brew guides and photography brief in one scope
- Optional trade-show and launch support
Questions
What roasters ask us before starting
Do flavour notes really influence rankings?
They influence which queries you can match, and how well a page converts once it is found. "Blueberry, jasmine, tea-like" is a long-tail phrase with real volume across many small searches, and it is the language specialty buyers use.
Should wholesale and retail live on one site?
Yes, with two clear paths and different pages. Splitting into two domains splits authority and doubles the maintenance. Keep one site, separate the audiences at the navigation and the enquiry forms.
How do you handle hundreds of grind and weight variants?
One indexable page per coffee, with variants handled through selectors, canonical tags and a parameter policy. The collection page ranks, filters do not, and crawl budget stays where it earns.
How long does it take?
Technical fixes and product schema often move impressions within a month. New origin and guide pages need two to three months to settle. Wholesale enquiries usually lag retail by a further month because the buying cycle is longer.
Is this safe for a shop that already ranks?
We work in releases and keep a rollback for every structural change. No mass redirects in week one, no template swap without a staging check. If something drops, we can trace it to a specific release.
How is it reported?
One page a month: non-branded clicks by cluster, wholesale enquiries, subscription starts and any technical regression. Written plainly, with next month's queue attached.