Case Study: How Wise Built Localized Content That Compounds Across Markets

Wise, founded in 2011 as TransferWise and rebranded to Wise in early 2021, became one of fintech's most studied organic-search programs not through a single campaign but through a system. The company paired a small set of reusable programmatic templates with a custom, SEO-first content management system called Lienzo, then reproduced both across dozens of country subfolders and languages, a structure documented by Ahrefs (2021) and updated by Digital Rainmaker (2025). The public record also shows a rare migration outcome: Wise moved from transferwise.com to wise.com in early 2021 and transferred its organic traffic without noticeable losses within about four months, per Ahrefs (2021). The first lesson of this teardown is that none of the traffic figures that circulate for Wise are company-reported numbers. The second is what a smaller site can actually copy: the templates, the internal linking, and the hreflang and canonical discipline that keeps hundreds of thousands of near-identical pages from cannibalizing one another.

The numbers: three tools, one wide gap

The first thing to establish about Wise's numbers is that every traffic figure in circulation is a third-party tool estimate, and those estimates disagree with each other by a wide margin. Wise does not publish its organic traffic, so the story has to be assembled from crawlers and SEO suites that each measured the site at a different moment with a different methodology. Reading them as a single series is a mistake; reading them as a spread is the accurate way to understand the program's scale.

The earliest well-known snapshot comes from Ahrefs, which estimated about 6.4 million organic visits per month in mid-2021 and more than four million URLs indexed in Google as of that year, per Ahrefs (2021). The same teardown counted nearly 12,000 landing pages about SWIFT/BIC code combinations aimed at United States visitors alone. Ahrefs also put Wise's Domain Rating at 75, a figure worth keeping in its proper category: a tool's authority score at a point in time, not a number Wise reported about itself.

Later trackers produced much larger numbers. Digital Rainmaker estimates about 30.3 million monthly visitors from search, and separately reports Wise published 213,746 pages in the preceding year while its indexed pages grew from 173,781 to 387,527 and its estimated visitors climbed from 15.8 million to 30.47 million per month, per Digital Rainmaker (2025). Daydream, citing Ahrefs, estimates over 385,000 programmatic pages drawing more than 48 million monthly organic visits globally, and over 260,000 currency-converter pages drawing more than 46 million monthly organic visits, per Daydream (2024). Other write-ups sit between and beyond those figures: Omnius cites 16 million plus, while ViewEngine and Gracker claim 60 million, per Omnius and ViewEngine. The range, from roughly 6 million to 60 million monthly visits, is itself the strongest evidence that no single number should be quoted as Wise's official traffic.

The origins: a program built before a team

Wise was founded in 2011 and operated as TransferWise before rebranding to Wise in early 2021, per Digital Rainmaker (2025). The instructive detail about the early years is that the company did not have an SEO team until 2015, yet Ahrefs recorded a Domain Rating of 75, per Ahrefs (2021). That gap between the founding date and the SEO hire matters because it means the foundations of the program, the product pages, the brand, the early editorial content, predate any dedicated optimization staff. The authority was earned before the discipline arrived, and the discipline then compounded what was already there.

Fabrizio Ballarini is Wise's Head of Organic Growth and has led the company's organic growth program for roughly a decade, per Exposure Ninja (2026). His account of the early years is specific about pace. He says Wise built roughly 60% of the content infrastructure that still underpins its acquisition in the first few quarters, per Exposure Ninja (2026). He also called indexed-pages-per-month his favorite early-stage KPI, on the reasoning that feeding Googlebot more quality content makes it crawl, index, and serve a site more, per Ahrefs (2021). That is a volume KPI, but it is a specific kind of volume: pages that earn indexing, not pages that merely exist.

What they built: Lienzo, the SEO-first CMS

Wise's content engine runs on a custom content management system named Lienzo, built in-house to solve the problems off-the-shelf platforms leave behind for a site operating at this scale. Ahrefs describes Lienzo as SEO-focused, with features for templates, internal linking, meta tags and structured data at scale, hreflang and canonical handling, and built-in SEO split testing, per Ahrefs (2021). The feature list is itself a summary of the whole strategy: the CMS exists to generate pages in bulk, link them to each other, mark them up correctly, tell search engines which regional variant belongs where, and then test which version performs.

The split-testing feature is the part most general-purpose tools miss. A conventional CMS stores and serves content; Lienzo treats SEO changes as experiments. That is the difference between a site that launches a template once and a site that keeps tuning it after launch. It also explains why Wise's program could scale without quality collapsing: the CMS encodes the canonical and hreflang rules as system behavior rather than as a checklist each author has to remember. When the machine enforces the internationalization rules, thousands of pages can be generated without each one requiring a manual review of its link tags.

For a smaller site the lesson is not to build a custom CMS. It is to recognize that the CMS is where the SEO rules live. If a team generates pages but leaves canonical and hreflang to author memory, it will eventually produce the duplicate-content problems a system like Lienzo is designed to prevent. A single page can be checked against those rules today with the on-page SEO checker, which reports the canonical and heading structure a page actually serves, and a multilingual deployment can be verified with the hreflang checker.

What they built: the programmatic templates

The page inventory that Lienzo produces is built from a small number of reusable templates, not from individually authored pages. Ahrefs and Daydream describe the same set: currency converters, SWIFT/BIC codes, routing numbers, IBANs, sort codes, and provider comparison pages, all generated from a handful of template patterns, per Daydream (2024). Each template answers a query type that already exists in money transfer: a person converting currency wants the rate, a person wiring money wants the receiving bank's code, and a person choosing a provider wants a side-by-side comparison.

The currency converter is the largest of these. Daydream reports over 260,000 currency-converter pages across dozens of languages and more than 60 country subfolders, drawing over 46 million monthly organic visits, a third-party estimate, per Daydream (2024). Digital Rainmaker narrows the lens to one region and counts 32,802 currency-converter landing pages in the UK subdirectory drawing about 356,634 visits per month, also a third-party estimate, per Digital Rainmaker (2025). The two figures measure different scopes, one global and one United Kingdom, but together they show the same pattern: one template, reproduced by currency pair and by market, becomes tens of thousands of pages.

The code pages work the same way at a smaller scale. Ahrefs counted nearly 12,000 landing pages about SWIFT/BIC code combinations aimed at United States visitors alone, per Ahrefs (2021). The routing-number pages show how far the long tail extends once a template is indexed: Ahrefs data cited by Daydream puts Wise as the featured snippet for about 5,700 routing-number-related keywords in the US, and finds Wise ranking for 18,136 US featured-snippet keywords in positions 1 through 10, against 3,062 for Western Union, per Daydream (2024). The Western Union comparison matters because it shows the gap is in breadth of coverage, not just raw authority: Wise holds the snippet on thousands of narrow code queries where the template supplied the only complete answer.

What they built: the 2021 domain migration

The most cited episode in Wise's SEO history is the 2021 rebrand and domain move from transferwise.com to wise.com. The timing is reported slightly differently by different sources: Ahrefs dates the migration to March 2021 while Digital Rainmaker dates it to February 2021, so the safe description is early 2021, per Ahrefs (2021) and Digital Rainmaker (2025). The disagreement over a single month is a useful reminder that even the basic facts of a large migration get muddied in secondary reporting.

What made the migration a case study was the sequencing. Wise did not redirect the entire domain on day one. It ran wise.com alongside transferwise.com and tested redirects on a smaller part of the domain first, per Ahrefs (2021). Only after that smaller test did it roll the redirects out across the rest of the site. Ahrefs reported that Wise transferred its organic traffic to wise.com without noticeable losses within four months of the migration, per Ahrefs (2021). That phased approach, testing on a subset before committing the whole domain, is the transferable lesson, and it matters far more than the exact month.

One risk surface emerged in the same period. Ahrefs reported Wise had close to 9,000 affiliate-related backlinks without link attributes as of 2021, per Ahrefs (2021). A backlink without a rel attribute marking it as sponsored or nofollow is a compliance exposure on both the editorial and the search side, and it is the kind of detail a migration and a growing affiliate program can quietly accumulate. The figure is a tool count, not an admission by Wise, but it shows that even a disciplined program carries legacy links that need auditing.

The internationalization layer: subfolders, languages, hreflang

The internationalization layer is where Wise's program differs from a single-market content engine. Digital Rainmaker reports Wise uses 61 country subfolders and writes in 33 languages across 57 markets, with about 37 content writers, per Digital Rainmaker (2025). Daydream describes the currency-converter footprint as spanning dozens of languages and more than 60 country subfolders, per Daydream (2024). The counts are close but not identical because they come from different crawls, and the number of markets should not be conflated with the number of subfolders. Exposure Ninja's framing of more than 160 markets, per Exposure Ninja (2026), refers to a different and larger unit than the 61 subfolders, since one subfolder can serve several markets and markets can be served without a dedicated subfolder. The language count varies the same way: 33 per Digital Rainmaker, "dozens" per Daydream.

The technical discipline that makes this breadth safe is the hreflang and canonical handling built into Lienzo, per Ahrefs (2021). When one template produces a page for the same currency pair or the same bank code in several languages and countries, those pages are near-duplicates by construction. Without explicit signals, search engines can treat them as duplicate content and pick the wrong regional variant for the wrong searcher, which is the cannibalization the international layer exists to prevent. Wise's answer is to make those signals systematic: every regional page declares its language and target market, and every localized variant points back to the canonical version of the page. A site that runs even two or three language versions can verify the same setup with the hreflang checker, and the full logic is laid out in the hreflang guide.

Why it worked

Three conditions made Wise's program work, and each one is worth checking before anyone copies the surface. The first is that the templates attach to demand that already exists and to a problem the searcher does not yet know they have. Ballarini frames the currency converter as a programmatic SEO "masterclass" built to educate customers about hidden bank markups, and states the core problem plainly: "people don't know they're being charged by their provider," per Exposure Ninja (2026). A person searching for a currency rate is about to learn that their bank's rate is worse than the mid-market rate. The page answers the query and delivers the product's reason to exist in the same visit. That pairing of informational intent with commercial conversion is what makes the currency converter more than a traffic grab.

The second condition is that the template produces genuine differentiation rather than thin duplication. Every currency pair and every bank code is a real, distinct data point, and the page exists to present that specific data. Two SWIFT code pages share a layout but not a fact. That is the line between a programmatic page that earns rankings and a doorway page that does not, and it is why the volume works where a wall of spun paragraphs would not. The featured-snippet data supports this: Wise holds the snippet for thousands of narrow routing-number keywords, per Daydream (2024), because the template supplied a complete, correct answer where generalist pages did not.

The third condition is the deliberate absence of a link-building program, which sounds counterintuitive until it is read correctly. Wise says it does not do link building, and Ballarini describes the approach as being "just in the mood to build more pages," per Digital Rainmaker (2025). The links then arrive because the pages are useful, referenceable assets. A currency converter or a SWIFT code lookup is the kind of page other sites cite when they explain how to send money, which compounds the authority without an outreach team. The catch is that this only works when the pages genuinely deserve the citation, which loops back to the first two conditions.

There is also a structural reason the play worked: internal linking at a scale that matches the page count. A system of currency, code, and comparison templates is a ready-made internal graph, where each currency page can link to the related code pages, comparison pages, and market pages around it. When the CMS enforces that linking, no generated page sits orphaned, and authority flows across the thousands of pages instead of puddling on the homepage. That is the difference between a pile of URLs and a lattice.

What could break it

The model's largest vulnerability is the same one that threatens every programmatic program: Google's long-running effort to demote thin, templated, or auto-generated content that exists only to rank. Wise's pages survive because the data behind them is real and useful, but that is also the ceiling. If a template's data goes stale, or if the pages drift toward boilerplate, an entire category of pages can lose trust at once, because Google evaluates the pattern as well as the individual page. The scale that compounds on the way up also compounds on the way down.

The second risk is the ambiguity of the numbers themselves. Every traffic figure for Wise is a third-party estimate that conflicts with the others by a wide margin, from about 6.4 million monthly visits per Ahrefs in 2021 to 60 million in later write-ups, per Ahrefs (2021) and ViewEngine. That ambiguity does not threaten Wise's business, but it does threaten anyone who copies the model on the strength of a borrowed number. A team that builds 50,000 pages expecting a figure it read in a teardown is planning against an estimate, not a fact.

The third risk is the one Ballarini himself flagged: some search traffic will shift toward LLM recommendations, even as he notes Wise's traffic is at an all-time high and continues to grow, per Exposure Ninja (2026). If a user asks an AI assistant for the mid-market rate or for a provider comparison, the assistant may answer without sending the user to Wise's page. The currency converter was built for a search box; the next interface may not have a search box in the same sense, and a program optimized for featured snippets has to ask whether the snippet itself is the terminal surface.

The fourth risk is crawl and index bloat at this scale. Hundreds of thousands of pages, per the third-party counts, require a clean sitemap, disciplined canonical tags, and a willingness to prune pages that no longer earn their keep. The same generator that fills the index can also flood it with stale or near-empty pages, and once the low-value pages outnumber the useful ones, the whole domain's crawl efficiency degrades. A site can check its own inventory with the on-page SEO checker and audit its sitemap hygiene before it scales, because the cleanup is far cheaper at a few thousand pages than at a few hundred thousand.

How to apply it

The copyable core is the ordering of decisions, not the raw page count. First, find the queries in your niche that already exist and that carry a hidden insight, the way a currency-rate query carries the hidden markup. Use the keyword research tool to list the head and long-tail variations, and sort them by the specific data each page would need. Second, decide what unique data every page will carry before you generate anything: a live rate, a bank code, a fee comparison, a regional fact. A template page without a unique data source is a doorway page, and it will not last. Third, build one template, get it indexed, and verify it before you multiply it, which is the same test-on-a-subset logic Wise used for its migration.

The technical layer is where imitations fail, and it is where the free tools here save the most time. Before generating pages, inspect a reference page with the on-page SEO checker to see the heading outline, canonical, and internal-link structure a page like Wise's serves, then make sure your own template enforces the same rules. If you run the template in more than one language or country, set up the hreflang checker so each regional page signals the right audience and you do not compete with your own localized variants, which is the exact cannibalization Lienzo is built to prevent, per Ahrefs (2021).

The maintenance loop matters as much as the launch. A generated inventory decays unless someone keeps the data current, the internal links coherent, and the low-value pages pruned. Run a periodic pass to confirm the sitemap lists only live canonical URLs, that the internal links still route from broad pages to specific ones, and that no orphaned template page is accumulating. Read this play next to its closest relatives before committing: the NerdWallet comparison content case study shows the same comparison-template logic in a financial context, and the hreflang guide covers the regional-canonical rules that make a multi-market template safe.