Barber cosmetics e‑commerce · Germany
Barberkosmetik sells men's grooming and barber cosmetics to the German market, from one catalog, at two sets of prices. Eleven of its pages turn up in ChatGPT answers — 42 responses in an Ahrefs check, plus three in Perplexity — on a domain that was new the day we launched it. The store runs on an engine we wrote, which is how a barbershop account and a private account can see different numbers on the same product page.
At a glance
The starting point
Barberkosmetik sells from one catalog to private customers and to barbershops, and the two groups do not buy on the same terms. A barbershop registers as a trade customer and expects its own prices on the same products. A hosted shop template treats both the same way, and every rule the business needs on top of that becomes a plugin sitting over someone else's checkout. The store also had to be correct for the German market on day one: consent handling before any analytics tag fires, a written returns policy, delivery through an integrated courier service, and a domain with no history in search at all.
What we built
Barbershops register as a customer group. From the moment their account belongs to that group, the prices shown to them in the catalog and at checkout are that group's prices. Same site, same cart, same checkout, different numbers. It is one of the rules that pushed this project onto a custom engine in the first place.
Google Consent Mode governs what fires and when, so statistics and marketing tags respect the visitor's choice in the cookie banner. Conversions are configured against real checkout events, and Search Console holds data going back to the launch window, which is why the first eight weeks can be shown as a chart rather than described.
Checkout is integrated with payment systems, and orders hand off to a courier service for delivery. The returns policy is written into the store itself, because in German e‑commerce the return terms are part of the buying decision and buyers look for them before they pay.
The store sends its own email: newsletters and automatic notices when an item comes back in stock. Product data also reaches external marketplaces through integrations, so the same inventory can earn orders on channels the site does not own.
Pricing rules bend to the way this shop sells, because the code belongs to the store rather than to a platform. Group pricing, marketplace integrations and stock automation are part of the same application, so each one behaves identically in the catalog and at checkout. A business rule that no shop platform ships with is a change to the code, not a search for a third app.
Design and devices
The store had to work as a German shop for two different buyers without splitting into two sites. One catalog, one checkout, and a customer account that changes what a logged‑in barbershop pays. Around that sit the parts a German buyer checks before paying: the consent banner, the return terms, the delivery handling. All of them are built into the store, so they behave consistently instead of living as pages bolted on after launch.
What came of it
The strongest numbers on this project come from a third-party tool. An Ahrefs overview for barberkosmetik.de counts the assistant citations — 42 responses in ChatGPT across 11 pages, 3 in Perplexity — and that is the part of its own performance a young store almost never gets to see. Google Search Console tells the slower story alongside it: clicks and impressions sit flat near zero at the end of June 2026 and climb steadily through mid-August. The client reports that the store began taking orders from organic search three to four weeks after launch, and began to recoup the build inside that window.
ChatGPT responses citing the siteAhrefs AI visibility overview, across 11 cited pages
Pages cited by AI assistantsAhrefs records all eleven as gained in the measured period
Perplexity responses citing the siteAhrefs, up by two in the same period
From launch to first organic ordersClient-reported.
Ahrefs' AI visibility overview for barberkosmetik.de records 42 ChatGPT responses citing the domain across 11 pages, and 3 responses in Perplexity. AI Overviews shows none. By classical measures the domain is still young: 2 organic keywords, an estimated organic traffic of 1, and 3 backlinks from 3 referring domains at the time of the snapshot. Assistants are quoting a store that Google has barely begun to rank, which is exactly why the measurement went in during the launch week — the citations are counted by a third party and anyone can check them.
How we ran it
Every project runs on a written plan. This is how this one ran.
We map the store before writing it: structure, products, customer groups and which prices each group sees. This is where the retail and trade pricing rules get defined, because retrofitting them later touches every page.
We build the store itself: product and cart logic, customer accounts, and the group pricing that follows a registered customer through the catalog and into checkout.
Payment systems, the courier handoff, the returns policy, and EU consent handling through Google Consent Mode. Nothing tracks a visitor before the banner is answered.
Conversions are configured against real checkout events, analytics and Search Console are connected, and the first weeks are read as data. This is where we learn which pages search picked up.
We keep watching what search and AI assistants do with the pages. On this project the client's own analytics surfaced assistant referrals about six weeks in, and an Ahrefs overview later put numbers on the citations. We treat that as something we measure; we do not sell it as a method with a guaranteed outcome.
Newsletter and back‑in‑stock email, marketplace integrations, and whatever business rule comes next. On a custom engine each of these is a change to the store the client already owns.
Straight answers
Yours is probably close to one of these. If not, ask it directly.
Ask about your projectSame team, same method
Tell us what you sell and where you sell it. You get a plan like the one above, with the first step priced out and dated before anything starts.
