Today’s customers are hyperconnected. They expect seamless journeys across every digital and physical touchpoint, tailored around their individual needs and goals. This means marketers need to leverage huge volumes of data to deliver increasingly personalised experiences.
Over the course of this series on the customer experience in the financial services industry (FSI), we’ve been exploring the components of delivering great experiences to our customers. In the first article, we spoke about the explosion of customer and transactional data. In the second article, we examined the need to streamline and deliver more content than ever before.
As we saw in those earlier pieces, it’s impossible to manually manage the amount of content necessary for all those experiences. The only way you’ll be able to deliver proper content velocity, at scale, with individual messaging, is with the help of machine learning.
In a recent Adobe survey asking FSI marketers which areas seem to offer the most exciting prospects for 2020, the majority agreed that the number-one most exciting prospect is using artificial intelligence (AI) to drive campaigns and experiences. IDC research found that European business spending on AI has increased by 40 percent since 2016, while even more organizations are currently planning to increase their spending in this area, with the goal of improving customer engagement.
This is because AI is no longer just an opportunity. It’s a requirement for any organisation that aims to stay connected with its customers in the coming decade. But this doesn’t mean the marketer’s role is becoming obsolete. On the contrary, machine learning is designed to handle the “heavy lifting” of rote, repetitive tasks (such as checking and tagging thousands of images), leaving the marketer free to provide more of the unique human touches that make a campaign brilliant.
There are many areas that AI and ML can help and support a marketer in delivering the optimal experience. In this blog, I am going to focus on only a few areas such as anomaly detection, contribution analysis and content personalisation. Other areas like subject line optimization, media mix modelling and recommendations can be just as effective.
So, how do you use the predictive capability of AI and ML to deliver better customer experiences? It all starts when you look beyond your basic analytics and begin to search for actionable anomalies in the data.
A predictive marketing workflow
The first step in an AI-centric marketing approach is to understand why your analytics fluctuate at specific times. For example, you might know your average conversion rate and average number of daily visitors to your website, which makes it very easy to get an average and upper and lower bound, but you might not know when or why those rates shift.
This is where anomaly detection comes into play. You can use machine learning to look at the historical data and recognise the trends, such as upward or downward fluctuations according to the day of the week or time of day. Once you’ve established those trends, you can use anomaly detection to tell you when something out of the ordinary has happened.
The detection of the anomaly leads to the second component of this workflow—contribution analysis. Say, for example, you’ve seen a spike in traffic and you want to know what caused it. This task would ordinarily require your analysts to scour hundreds of reports. It may take them days to pinpoint the cause of the spike. But with AI and machine learning, you can search through many factors that may correlate with that spike, and quickly zero in on the most likely root cause.
Proactive content personalisation
Now that you have the ability to detect traffic anomalies and identify their causes, you’re well-equipped to perform the next layer of analysis to understand why certain audience segments are drawn to certain touchpoints at certain times.
This is where propensity modelling comes in. This machine learning technique automatically figures out which factors engage or convert certain audience segments at certain times. These factors may be specific types of content, specific messages or calls to action (CTAs), or other factors that vary throughout your A/B tests.
The final part of the predictive market workflow is activating these new intelligent segments with the right content. If you’ve got 10 different templates for your home page banner, 1,000 different background images, and 15 potential messages, that adds up to an enormous number of potential combinations of image, template, and message. There’s no way your design team could manually assemble all those final pieces for each individual customer—certainly not in real time. Nor could your Analytics team identify every potential audience segment – especially when it is a segment of one!
But AI can handle all that heavy lifting for you. Now that you know exactly what each audience segment responds to, you can use AI to generate thousands of personalised combinations of image, message, and CTA based on quantitative analyses of all your previous interactions with that particular segment. You’ll always deliver just the right content and message for each specific visitor.
In fact, this content personalisation brings you back full circle. You’ve gone from anomaly detection to contribution analysis to propensity modelling to activating the audience, and now, as you serve personalised content to each audience segment, you can continue to gather data on their reactions, and use that data to further sharpen your content for even stronger engagement in the future. The next step is to deliver those fluid experiences across all channels. That’s what we’ll be discussing in the next instalment of this series. See you there!