A Berlin machine-parts supplier asked me to "just add Dutch" to his Elementor site, as if a language were a plugin you switch on. Six weeks later we had three half-translated menus, a checkout mixing German nouns with English buttons, and a cookie banner that promised privacy in a language the page itself did not speak. Adding a language is an architecture decision, not a toggle, and for EU and UK businesses it drags in hreflang, translation workflow, cookie consent per language, and real upkeep cost. Here is the decision tree I now walk every client through before a single string gets translated.
Path one is a lightweight translation plugin on a single site, which suits exactly one situation: two languages, modest page counts, and one person owning the translations. TranslatePress lets you translate from the front end, keeps one Elementor design shared across languages, and its free listing on wordpress.org makes it easy to trial. For a 15-page brochure site targeting Amsterdam and Berlin, budget around 800 to 1,200 EUR for setup plus professional translation of core pages, then 50 to 100 EUR a month in upkeep as content changes. The trap is machine-translating everything and publishing it: German engineers notice, Dutch buyers bounce, and your landing pages stop converting in the language you translated worst. Machine output is a draft, never the publish button.
Path two is full multilingual with WPML or TranslatePress Pro, which is what I recommend once you pass two languages or run a shop. Separate translation management, per-language menus and widgets, and proper hreflang output so Google serves the German page to google.de and the French page to google.fr. Setup for a 30-page site plus WooCommerce strings typically lands between 2,000 and 4,000 EUR including translator fees, with 100 to 250 EUR a month in ongoing translation maintenance. The honest trade-off is editorial burden: every offer, every banner, every legal line now exists N times, and someone must own that calendar. I tell London clients to appoint one language owner per market before we build anything, because orphaned French pages rot faster than untranslated ones and confuse buyers more.
Path three is structural separation, subdirectories like /de/ and /fr/ with distinct content per market, or fully separate sites per country. This fits businesses whose offers genuinely differ by market, different pricing, different stock, different sales teams. It costs the most, think 5,000 EUR and up per market site, but each market team edits freely without breaking the others, and local SEO compounds per market instead of fighting itself. I only recommend it when the business case is proven in revenue, not ambition. Two of my clients started here on instinct and merged back into one multilingual install within a year, paying for the lesson twice. Start unified, split when a market earns its own site.
Whichever path you take, the compliance layer is non-negotiable and it must speak every language you publish. Under the EU data protection rules, consent must be informed, which means a Dutch visitor must understand the banner in Dutch, with the same opt-in standard and the same granularity as the English one. The UK position in the ICO guidance points the same way. I build one consent configuration with translated strings, audited per language at launch, and I re-audit every time a market is added, because a compliant English banner next to an untranslated French one is the exact gap I documented in my GDPR Elementor guide. Regulators read the whole site, not just the homepage language.
My verdict after a dozen of these builds: most EU small businesses should start on path one or two, budget translation as content rather than tech, and revisit structure yearly as traffic splits. The 800 EUR setup that ships with a translator beats the 4,000 EUR setup fed with machine output every single time, because buyers trust precise words more than fancy architecture. Verify every number in your quote against your actual page count and string count, appoint a language owner per market, and translate the boring strings, checkout, errors, emails, banners, before the marketing ones. Nobody ever lost a customer to an untranslated hero headline. Plenty have lost one to a checkout they could not understand.
Look through my portfolio for Elementor builds structured for growth, then contact me to plan your language setup.
