Digital Transformation Agency
What are you looking for today?
Search
This document can be found at
Home >
Website personalisation and testing made easy
Website personalisation and testing made
easy
The Observatory Team – 13 May 2021
Tags:
Observatory (/taxonomy/term/412) Data (/taxonomy/term/17)
Service design (/taxonomy/term/64)
As user needs and community expectations change, agencies need to continually revitalise
their digital services to keep pace with these expectations. To help agencies meet these
expectations, the gov.au Observatory now provides a simple tool to test and implement
changes.
We've added a new tool called Optimize 360
To help gov.au Observatory subscribers meet changing user expectations, we have added
Google Optimize 360 to our service.
Optimize 360
(https://marketingplatform.google.com/about/optimize/) is the premium version of Google’s
website testing and personlisation product. It's an optimisation tool that can test different
website elements and create personalized experiences for specific user needs.
Out of the box it integrates easily with Google Tag Manager, Analytics and Firebase. This allows
you to use the data to create specific user cohorts and capture test data.
Using the
Optimize editor (https://marketingplatform.google.com/about/optimize/) you can
change the color of a button or move important content higher up the page. As your
confidence grows you can move from simple A/B testing to full multivariate testing. Once you
have validated your testing hypotheses, you can use the editor to deploy these changes as
targeted personalisations.
Better website testing
As you review your web analytics, you may start to identify parts of your service that need
improving. Building on this evidence, you will create hypotheses that a change may improve
your users’ experience. But how will you know if the changes work?
Starting with
A/B testing (https://support.google.com/optimize/answer/7012154#ab-tests) you
can create a simple experiment that compares the original website (A) with a variant (B/C/D/…).
Optimize allows you to choose how often variations appear and who sees the variations.
Audience segments can be based on what device they use, where they came from, or how they
were referred to your service. Alternatively you can create redirect tests
(https://support.google.com/optimize/answer/7012154#redirect-tests) that allows you to test 2
totally different pages.
As your hypotheses mature, you can expand your test to include multivariate experiments
(https://support.google.com/optimize/answer/7012154#multivariate-tests). These more
complex experiments allow you to test 2 or more elements of your website simultaneously.
Rather than seeing which variant is most effective, the multivariate testing lets you identify the
most effective element type and the best combination to present them in.
Website Personalisation
When you have identified new features and designs through testing, you can use website
personalisation to offer a specific service to tailored user cohorts. You can also tailor your site
to reinforce an agency campaign, such as support for rural and regional Australians. As users
access your service through various calls to action, such as an email, news elements can be
added to your service to emphasise the campaign’s call to action.
How do I get started?
Contact the Observatory team at
xxxxxxxxxxx@xxx.xxx.xx (mailto:xxxxxxxxxxx@xxx.xxx.xx) to
enable Optimize 360 for your services. We are providing this service at no additional cost for
the remainder of the year, to see if there is sufficient demand to keep it as a permanent
feature.
We are looking to run free training on how to make the most out of Optimize 360. Visit the
Observatory website and
subscribe to our newsletter (https://observatory.service.gov.au/get-
started/#email) for updates about this training.
The Observatory Team uses data science to learn how people interact with government, and
manage several whole-of-government data services.
© Commonwealth of Australia. With the exception of the Commonwealth Coat of Arms and
where otherwise noted, this work is licensed under the CC BY 4.0 license.
