Cheap first year. Then three times the price.

The advertised price of European shared hosting is about €5.15/month (median measured offer, ex-VAT). That is not the price. It is the first-year price — and the storefront usually says so, quietly, in the "then €13.45/mo" small print. We extracted both sides of that sentence, storefront by storefront, across some eighty brands in two dozen European markets: 2361 offers in 22 markets state both the intro price and the renewal price on the page. 91.8% of them renew higher. The median shared-hosting offer renews at 3.0× its intro. That book spans several product categories. On shared hosting alone — the category every distribution below measures — almost every offer that states both prices renews higher; the flat ones are a rounding error.

Hosting churn debates usually argue about loyalty. The cliff suggests a simpler frame: hosting companies advertise a cheap first year, then charge about three times as much every year after — and how steep that jump is, is a choice each company makes, measurable per market and per operator.

The cliff is a national norm, not a European one

Merchandising conventions cluster by market. The median renewal multiple for shared hosting runs from about twice the intro price in Switzerland and Austria to more than four times it in Denmark — the same product category, merchandised under entirely different local norms:

The medians understate how aggressive the deep end is. At the 90th percentile the Netherlands reads around twenty times the intro price, and Poland, Germany and the United Kingdom all sit around ten — recruitment offers priced near zero against a full-price year two. The deep end is thin, so single storefronts move it from one capture to the next; the exact percentiles live in the figures panel below.

Genuinely flat hosting pricing — an offer that states both prices and renews at the same number — remains rare, and it is worth being precise about where the flat rows actually are. The offers that do renew at or below their intro price are a small minority of the book — and every one of them is a domain registration rather than a hosting plan, sitting on a handful of storefronts' flat domain price lists. On shared hosting the picture is the one the rest of this brief describes: flat offers are fewer than one in a hundred, and none renews lower. A flat price list is a real merchandising posture; on this book it is a registrar posture, not a hosting one, and counting it as a hosting segment would be reading a few storefronts' product catalogues as a market.

What the measured median cannot see

There is a storefront the measure above structurally excludes: the one that displays a single plain recurring price — no crossed-out number, no discount language, nothing to renew up to. Those storefronts state no renewal, so they never enter the measured median; the measured median therefore samples the teaser crowd and can overstate the market. We now carry a companion figure for exactly this: counting never-discounting storefronts at their displayed word — no jump at renewal — the structure-adjusted median — and, weighting every storefront by the customer relationships it serves, the customer-weighted cliff. Grain matters more than the number, because the two can invert. In Denmark, about half of the tracked storefronts display one plain recurring price, which pulls the storefront-grain median down by more than a point — but those storefronts are small operators, and the ones that do publish a renewal serve more than nine in ten captured relationships: the cliff the average Danish customer actually meets is about four and a half times the intro price. Norway inverts the other way: the measured median is around four times, but the flat-displaying market leader carries most of the footprint, so the average Norwegian customer meets ~1× — taken at the storefront's displayed word, since no published renewal exists to verify it. Neither companion is a measurement; both are stated with their basis and never as the headline. All three figures, per market, are on the market pages and in the connector (market_pricing: median_renewal_cliff, structure_adjusted_median_renewal_cliff, share_weighted_median_renewal_cliff, renewal_disclosure_weight_pct).

Both archetypes follow the convention. One prices it steeper.

Fold the measured shared-hosting offers by ownership (curated operator registry; aggregate-level) and the picture splits: brands owned by multi-brand consolidator groups renew at about three times the first-year price (371 measured offers across 57 brands), independents at about two and a half (39 offers across 13 brands - a thin base, read it as a direction). In the three markets where both archetypes field enough offers to compare, the independent cliff is the lower one every time. The cliff is the market's convention; consolidators price it a little steeper. Even the tail is close to proportional rather than a portfolio signature: most of the deepest teasers do sit on consolidator-owned storefronts, but consolidators are about nine in ten of the measured offers to begin with, so their share of the deep tail is barely above their share of the book. Which archetype runs the deeper tail has traded places between captures — the independent side of that comparison is a handful of offers, not a strategy — and that instability is itself the finding. On an earlier and much smaller book the tail looked like a portfolio signature; on this one it does not, and we would rather say so than keep the better story.

Within a single market the comparison is thin and does not point one way. Four markets field at least five measured offers on both sides, and they disagree: in Sweden consolidator-owned brands sit roughly twice as steep as independents and Norway runs the same direction, but Finland and Poland run the other way — independents steeper than consolidators, in Poland markedly so. Two dozen offers a side is a storefront sample, not a strategy, and the reader who needs this at operator grain should take it there rather than from an archetype.

How the cliff is merchandised

Across the several thousand priced offers we track, the promotion grammar around the cliff is consistent: the intro teaser is the workhorse — on about three in ten offers. A bundled free domain adds switching cost precisely when the customer is cheapest to acquire; an explicit percentage off frames the year-one price as a discount rather than the year-two price as an increase; a money-back guarantee reverses the perceived risk at signup while the real economics sit twelve months later. Each device is legitimate on its own. Together they are a machine for making year one feel cheap and year two feel like fine print.

Read device counts against the book they come from rather than against an earlier one: about three quarters of the tracked offers are now domain registrations, and the money-back guarantee is almost entirely a hosting device. A device count that falls while the book grows is usually telling you what the book is made of, not what merchants changed. Exact counts, per device and per market, are on the market pages and in the connector.

Who should care, and why

For anyone comparing books: the cliff is a book-quality lens. Two hosting books of identical size can carry very different renewal risk: one recruited at 2.0× in a restrained market, one at deep double-digit teasers. The deep-teaser book meets a price shock at every anniversary — its churn, its realized ARPU and its LTV assumptions are all functions of a merchandising decision visible on the storefront. The question to ask of any book is what it was recruited at, not only how big it is. The same logic describes a consolidator's playbook: buying a book and steepening its cliff harvests year-two revenue at the cost of measurable churn exposure. Whether a given group runs that playbook is an operator-level question, and the archetype split above is too coarse to answer it — which is the argument for asking it per operator rather than per category of owner.

Operators and corp-dev: the cliff is a strategy benchmark. Your merchandising norm is set by your market, and deviating from it is a choice with two edges: teasing deeper than the local norm buys volume and renewal-shock churn; pricing flatter is a retention story your competitors' year-two customers can hear. Knowing the market's actual median — not folklore — is the difference between a pricing strategy and a guess. Promotion depth is also a leading indicator worth watching on rivals: a competitor suddenly deepening teasers is buying growth; the bill for it arrives in their renewals.

Reproduce this — or ask it yourself

Every figure here is queryable through the HostingBrain connector. In Claude or any MCP-compatible assistant: