Since 2015, the Japanese car brand has been working with Publicis.Sapient to migrate its digital properties across 109 countries onto this platform and has reached 105 today.
However, with this single platform came vast amounts of data that the team knew had to be analysed and acted on but did not have the time or resources to do so.
"The 105 countries were attracting 1.15 million visitors every day across 190 digital showrooms," Dév Rishi Sahani, global head of customer experience data and analytics at Nissan, told audiences at Adobe Summit London on 3 May.
The resulting data points were "overwhelming", Sahani said. "We didn’t just want a summary of all this data, we wanted to measure the global scale of the organisation, to drill down on a market and bring in all these nuances and have it make sense across all the brands."
To tackle this, the team drew up a goal-driven measurement framework for the data insights they were after. This was led by data that pertained to customer satisfaction, then came brand visibility and reach and ensuring that the work aligned with business goals.
"But we’re not an FMCG brand, we don’t have thousands of people working on customer insights," Sahani said.
So the team at Publicis.Sapient took advantage of the functionalities present in Adobe Experience Manager and Analytics Cloud – which hosts Nissan Pace – and built its own custom model that attributed scores on an individual market by market level to determine what was working and what was causing churn.
"Not every page is the same, and we don’t have enough people to control it all so we turned to machine learning to ask the data questions such as: ‘What factors drive conversion?’ or ‘Where do non-converted visitors drop off?’," Scott Ross, vice-president and executive client partner at Publicis.Sapient, explained during his portion of the presentation.
While on a small scale, these questions are easy to answer, Nissan’s presence across 147 markets in 105 countries made it a challenge to answer the question on a market-by-market basis, rather than as a global aggregate.
"We used historical data to train the machine-learning algorithms on conversion models and then looked at present-day data at scale," Ross said.
This resulted in some interesting insights for the car maker.
In one Asian market, we were seeing high rates of customers looking to book a test drive, Sahani shared. But, while nearly 27,000 customers visited the site to fill in the form, only 91 (0.3%) actually completed the form.
Machine-learning identified this problem and, after crunching the data, identified that 99% of customers were exiting the form at the very first question.
"You see, this market has an 84% mobile penetration rate and the form was not at all mobile-friendly," Sahani said. "The first time they had to select something, 99% left the form."
Taking that data, the team examined the form’s UX, and tested several different versions. This insight led to a 900-times increase in test drives booked and they could then roll out their learnings from this market to all of Nissan’s sites.
"Looking at the data globally, it wasn’t significant, but this when you examine it by market, you really learn things," Ross explained.
Not every market has the same attribution path, he continued. "One big example is testimonial content. Globally, it wasn’t significant so markets don’t focus on it. But when we introduced machine-learning and looked at the data at scale, we saw different markets had different activities and for one market, testimonials were an order of magnitude more important – and that local market had no idea," Ross said.
Mastering these minute data points matters so much to Nissan because "it’s not just a website, it’s our showroom and understanding the digital body language of our customers is critical to providing the relevant experience," Sahani concluded.