The Trouble with Hreflang

Tristan Pirouz
Tristan Pirouz

On 11th March 2016 • 4 min read

Google’s solution for hreflang seems relatively simple to understand, but the technical challenges to implement the solution are not obvious from the outset.

Producing hreflang tags at a page level, or in Sitemaps, requires a specific set of data to be available in the databases, which can be queried quickly and efficiently.

Pages have to be loaded very quickly, which means it’s probably necessary to cache the page level hreflang tags.

Sitemaps can contain 50,000 URLs, each of which may require 10s of hreflang tags – so a lot of work may need to be done to calculate everything for all of the pages in a single process.

The requirements for Sitemaps are generally simpler than page level tagging. They have a much longer time-out, and it’s relatively easy to generate them automatically at specific periods to create cached versions that can be downloaded quickly.

Both systems require the same data structure to create; a mapping of every shared product between each regional site, URLs for each variant, and the country-language setting of each URL.

Below, find the most common reasons why this might not be easy, or possible at all.

Different Content/Inventory per Region

The content or inventory is rarely identical for every region. Sometimes, there is a heavy overlap with some differences. In other instances, little to no overlap (in this case hreflang might not make sense at all).

Lack of Shared Key Fields for the Same Product in Different Regions

It’s quite common to lack any shared fields to identify the same product in different regions. This makes it impossible to query to the database to produce the correct hreflang tags.

Mixture of Platforms for Each/Some Regional Site

If the regional sites are hosted on multiple platforms, they are unlikely to have a single connected dataset necessary to create the hreflang tags.

Language Translated URLs

Descriptive URLs are desirable, and this requires them to be in the local language for the region. However, this makes mapping the same page between regions much harder, as the translated URLs have to be available. A simpler URL structure that just uses IDs might be much easier to generate the URLs for the alternatives.


Hreflang Solutions


The Platform Solution

Hreflang support needs to be considered when redeveloping a new platform. Get the necessary data structures in place, shared product IDs, etc.

Partial implementation is better than nothing, so if just a few sites are on a new platform which supports hreflang, then migrate other regions over time.

The Manual Effort

If your development cycle is going to prevent the necessary changes within a reasonable timeframe, a manual exercise of mapping the URLs is possible, depending on the number of pages involved.

You can then create a manual XML Sitemap with all the hreflang information. If a site is relatively static, this could be a worthwhile exercise.

The Partial Solution

Google is much more likely to work out the right pages for language specific content without any hreflang tags, so you might want to focus on regions with shared languages to ensure the differentiation.

Read More About Hreflang

Lean more about international SEO

DeepCrawl's Ultimate Guide to International SEO

To learn even more about the ins and outs of hreflang and the other elements of international SEO such as geotargeting and content localisation, make sure you read our comprehensive internationalisation white paper.

Read the International SEO Guide Now


Get the knowledge and inspiration you need to build a profitable business - straight to your inbox.

Subscribe today