Can you roll out successful split tests on your domestic site to other countries?

You’ve had a win on your domestic site. Will it work on your international sites, or do you need to adapt it? That’s the big question.

If the answer’s yes, you’ll save a ton of time and money by not researching each territory individually. But if the answer’s no, you won’t get the revenue increases on the international sites, and you may even damage the brand.

The good news is, there is an easy way to tell.

The answer is to do a simple piece of analysis to discover whether your domestic and international sites are tightly coupled or loosely coupled.

What do we mean by tightly coupled and loosely coupled optimisation?

Quite simply, if you run successful split tests on your domestic site and find that they are similarly successful on one, or more, of your international sites, you have a tightly coupled optimisation process.

This is excellent news as it means you can confidently cut back on the research and analysis of foreign sites, to speed up your conversion programme and free up resources.

When your domestic site split tests don’t deliver similar results or successes on your international sites, you have a loosely coupled optimisation process. This is more likely to occur when there are cultural differences between the visitors to the two websites in question.

How to identify whether your optimisation is tightly or loosely coupled

To identify whether your optimisation is tightly or loosely coupled, we recommend you take a batch of 7-10 split tests, which have been declared statistically significant ‘winners’ on your domestic site, and run them on your international site(s). Select split tests which cover different areas of your site to be sure that this coupling is site-wide and not just specific to a certain page or stage in the buyer’s journey.

Log the results of each split test in a table and you’ll be able to see if they are moving in the same general direction.

Below is an example where there is a high degree of similarity, which means these two sites are tightly coupled.

  Domestic site results International site results Similarity?
Test 1 +4.1% +3.6% Yes
Test 2 +3.1% +5.2% Yes
Test 3 -3.2% -4.5% Yes
Test 4 Inconclusive Inconclusive Yes
Test 5 +3.0% +2.4% Yes
Test 6 -4.2% -2.3% Yes
Test 7 -0.2% +1.5% No
Test 8 +8.9% +6.7% Yes

Table 1: An example of a tightly coupled optimisation analysis

In the scenario below, however, there is much less similarity between the two results with tests which were successful on one site often being unsuccessful on the other. This suggests that the optimisation of the international site in question is only loosely coupled with the domestic site

  Domestic site results International site results Similarity?
Test 1 +4.1% -2.3% No
Test 2 -3.2% Inconclusive No
Test 3 +3.0% -4.5% No
Test 4 +3.1% +3.6% Yes
Test 5 Inconclusive +2.4% No
Test 6 -4.2% +5.2% No
Test 7 -0.2% -0.5% Yes
Test 8 +8.9% -6.7% No

Table 2: An example of a loosely coupled optimisation analysis

It’s quite likely that not all of your international sites fall into the same category.

For example, if you have a website that trades in the UK, the Netherlands and China, you may find that split test results on the UK site and the Netherlands site are tightly coupled (as research has shown similarities between online shopping behaviours between UK and Dutch people), but the results on the UK site and Chinese site are loosely coupled (as research has shown wide cultural differences between UK and Chinese people).

We recommend looking at individual pairs in isolation so you can identify the strength of the coupling between individual sites. Comparing pairs of international sites, you may find that some sites which are loosely coupled to your UK site are actually tightly coupled to each other

How to set up your optimisation programme on sites where your optimisation is tightly coupled

If you discover that your optimisation is tightly coupled, your job becomes a lot easier, because it means that you can quickly and easily replicate wins you get on your domestic site, on your international site(s). Thus maximising the impact of your successful results.

However, it is still worth doing research and analysis on the international website(s) as well. It is possible that cultural differences, although small, are influencing user needs and context. This independent research can help you to identify and test hypotheses that could surpass the results generated from your domestic site optimisation efforts.

How to set up your optimisation programme on sites where your optimisation is loosely coupled

If you observe that results of your split tests tend to go in opposite, you will need to optimise the two sites separately and run independent research and analysis on the international site(s) to identify the root differences between your two audiences. Split testing hypothesis will need to be developed separately, and specific for each site.

Also be sure to get feedback on split test wireframes or designs from local representatives of your target audience to be sure that it really does address their concerns. Given that these international visitors haven’t reacted to your current designs in the same manner as your domestic one, this second opinion will ensure that your (domestic) designers have understood and addressed the opportunity in a way which is culturally appropriate to the international site in question.

Can we help?

As one of the few genuinely international CRO agencies, with offices in three continents, we have optimised a number of international websites. We’ve worked on projects where the international sites are tightly coupled and loosely coupled with the domestic brand.

Read our ebook below to find out how a focus on three power metrics can double your CRO success.


Discover how businesses have made the shift from CRO to experimentation – and you can too

Download your copy today >

Posted in: Conversion Rate Optimisation

If you enjoyed reading this, you might also like

Request a free CRO consultation

If you’re serious about initiating change within your business, we’d like to offer you a 60-minute consultation.

“We’ll share what we’ve learned from decades of experience working with businesses using CRO, innovation and experimentation to achieve business goals like yours”

Johann Van Tonder, COO, AWA digital.

This is an opportunity for ambitious online businesses who are already generating significant revenue… and know they can bring in more… but are not sure exactly how.


We partner with the leading CRO technologies