Challenge

Personalisation with privacy

As an independent, member-funded and not-for-profit group that regularly holds businesses to account for shonky practices, CHOICE is acutely aware of leveraging new technologies without compromising consumer privacy or data security.

Ashwin Sridhar, Chief Data & Technology Officer at CHOICE explains: “We have higher expectations from our members and the public than many commercial businesses. As a mission-driven organisation, we’re uncompromising when it comes to data privacy and security. Firstly, our policy is to only collect the data that we need to add member value, and secondly to ensure any data we do collect is used responsibly and aligned to our users’ expectations for privacy and security.”

“Having said that, data is crucial. Our members increasingly research and shop online, so we need to evolve with the digital environment to provide customised experiences across our platform.”With legacy IT systems having evolved over several years, it had become difficult for CHOICE to build a unified view of each member and integrate all their touchpoints across the organisation. While much of the data needed was ‘in the system’, it was split across different platforms within the organisation.

The product reviews platform is separate from the advocacy campaigns system making it difficult to join up the experience for their audiences. For example, if a user had purchased a fridge after reading a CHOICE review, they’re unlikely to want more fridge reviews but may be interested in advice on how to make their home more energy efficient, or a campaign to make products more reliable.

CHOICE partnered with Merkle Australia and Salesforce to work towards creating a unified view of its members to help better understand their needs and deliver more value for their membership.

Solution

Bringing all the pieces together
The first step was to integrate CHOICE’s member data into a unified and accessible format using a customer data platform (CDP), but that was only the beginning.

“Many CDPs can collect and centralise data, but then what? A connected view of members is great, but it’s just a means to an end. Translating that information into a personalised member experience that delivers more value is where we achieve the return on investment,” said Sridhar.

Having analysed the objectives, Merkle recommended running a proof-of-concept using Salesforce Interaction Studio (IS). While IS is relatively new in Australia, Merkle’s consultants – over 12,000 globally and 400+ in Australia – have been implementing it in other markets for many years, says Merkle’s Head of Technology Justin Grant.

“While it has origins as a CDP at its core, Salesforce IS is better described as a personalisation or digital experience platform. On top of stitching together customer identities from disconnected channels, it has artificial intelligence features to streamline customer service, content, marketing and member acquisition.”

The first trial was to personalise content recommendations on CHOICE’s home page specific to each member using machine learning powered by Salesforce Einstein. Engagement and click-through rates were AB tested against editorial selections over a period of 6 weeks.

Results

Human versus machine
Almost immediately the results showed that content selections automated by artificial intelligence outperformed editorial content. But, interestingly, this result flipped a few weeks into the test following a lot of media attention about health insurance.

“Our editors started covering the insurance issue closely, recognising that our audience needed advice on this issue. This new content resonated more strongly than the automated selections. Then, in a matter of days, the algorithm identified health insurance as a hot topic and began prioritising that content with great results,” explains Sridhar.

While the results ebbed and flowed, over the full 6-week trial the automated recommendations outperformed editorially selected content by 36% in terms of clicks and engagement rates.