What you need to know

• Even after a decade, MYOB is still racking up big wins from its CDP, the latest being strong improvements to abandoned cart rates.
• The accounting software provider is also able to build experiments for its marketing team much faster, with months turned into days.
• With over 50 different development teams, the CDP is an essential part of taming data complexity, a task made easier now that the firm has addressed the issue of identity resolution.
• Thanks to these efforts, MYOB was able to finally tackle the challenge of abandoned cart re-engagement, and has now seen a 20.1 per cent improvement.
• The next big ticket item is personalisation at scale, says its technical product manager.

Even after a decade's experience with its Customer Data Platform (CDP), MYOB is extracting big wins from the software, realising double-digit improvements in cart abandonment rates, and reducing the time taken to design experiments down from months to just days.

"We were on Segment before Segment was even a CDP," said MYOB technical product manager,  Peter Yanny, who connected with Mi3 in between sessions at last week's Twilio Signal conference in Singapore.

The CDP is an essential part of Yanny's job of taming MYOB's data complexity, which is a natural function of the company's scale.

"We have about 50 development teams, each of whom is responsible for things like apps, services, or websites that capture some kind of customer telemetry, and often we are talking to those teams. They have their own product management structure so we would be talking to those stakeholders, and putting in place data contracts, such as: 'This is the type of data that you would send, these are the identifiers we need," he explained.

Those conversations happen at both a technical and marketing level.

"From a product vision, or marketing vision perspective, it would be, what do you want to do with this customer? You want to have an experience where, when a customer does XYZ, they have a personalised email at the outcome of that [for instance]," Yanny said. "We would put into place the telemetry we need to capture at these points, then we'd be able to create an audience based on that. And we'd be able to do the activation as a result."

Marketing Ops

Marketing operations is another important stakeholder in the process, Yanny continued.

"Our martech team is very tech savvy. They get requirements from marketing and then brief us such as, 'This is the type of trait or type of person or type of customer, or these are the product aspects or product features they would use. Can you create an audience around that?'"

"We enable them to self-serve, create an audience, then be able to push that to an integration like a marketing platform or email sending platform."  

Finally, Yanny's team also consults with the data engineering teams around data warehousing. It is a complicated environment but it delivers strong gains for the business when it all comes together.

Recent successes with experiment design and cart abandonment are both examples of the ongoing return the Australian accounting platform has been able to realise.

"We had been struggling for quite a long time as a company with abandoned cart and how you do that with accounting software," said Yanny. But in the last two years, the company has tackled the issue of identity resolution, which provided the bedrock on which to address the issue of cart abandonment.

"By addressing the data layer we were able to instantly get the benefits of putting an abandoned cart experience into that shopping cart. That included things like personalisation, so we would offer you this type of discount on this type of product in the next few days, or, offer concierge help as a follow-up to help set up your software."

As a result of the work done by the team, "The abandoned cart re-engagement is 20.1 per cent," Yanny said. "It grew over time. The way it happened was through experimentation and rapidly being able to experiment.

"[Questions like] what about this type of wording in the email? Let's do A and B tests. Did that work? How many people engaged and opened? We need an engaged signal when someone opens the email. So we got that and fed it back in, then automated the experience. Do we want to follow them up with a phone call if they're logged in on the trial on day 29? Do we want to do an in-app pop-up?

"There are all these different types of engagement that drive growth. I think that's the important story [of the CDP]. It's not just, I use Facebook marketing and I got X amount of conversions'. It's 'We tried this, and we tried this and tried this, and we learned over time. This approach works for this customer, and this one doesn't'. Language changes make a big difference. Image changes make a big difference. There are all of these dimensions we've learned over time [help] to achieve that result."

Personalisation at scale

MYOB is now turning its attention to providing personalisation at scale - a key theme at the Twilio Signal conference this year - although Yanny acknowledged it is still early days.

"Almost every company at this show is probably trying some sort of initiative like this. But this is something we've only really ever dreamed of with an organisation as big as ours, with as many products as we have, on as many platforms as we have, with so many distributed teams, and with such a huge range of customers as we have," he commented.

To get there, Yanny told Mi3, the team again started small, picking a specific vertical with a specific product. "We got the data right [across] all of the stages, then started building that out. I can't really go into too much of the details at the moment, because it's still in progress. But one of the things I can mention would be that the speed at which we're able to experiment has been phenomenal, from roughly a three-month end-to-end experiment to hours."