Ahead of her presentation at CDAO UK, we spoke with Quantum Metric’s Marina Shapira about predictive analytics, why companies should embrace a culture of experimentation and how CAOs and CXOs can work together effectively
What is behavioural research? And what role should it play in an organisation's data and analytics strategy?
Behavioural research seeks to understand what motivates people, how they perceive the world, make decisions and form habit.
One of the main discoveries of behavioural research is that people tend to make snap judgments based on rules of thumb (heuristics) and partial information. This has a huge impact on the way people interact with digital products. For example, merely changing the order of words on a webpage can have an impact on a person’s purchase decision.
Essentially, every digital product or service asks people to perform a certain behaviour. For example, signing up for a newsletter, opening a bank account or purchasing a flight ticket. Therefore, behavioural research can create a competitive advantage by empowering businesses to understand how to influence behaviour and persuade people better than the competition.
The most important takeaway is that behavioural research is essential to understanding customers, and customers are what grows the business. This ties behavioural research directly to the bottom line.
How far would you agree that embracing this kind of experimentation is an important thing for organisations to do if they want to use analytics to predict and influence customer behaviour?
Behavioural science can provide organisations with the theoretical framework needed to best utilize predictive analytics.
Currently, businesses often overemphasise analysing short-term behaviours such as conversion or click-through rate and place too little emphasis on understanding long-term behaviours such as customer retention and loyalty.
Machine learning modelling is a great tool to understand which customer behaviours shape future outcomes. Such predictions are very powerful because they facilitate intervention before a customer churns.
For example, at Quantum Metric we use these models to figure out if customers are having trouble on websites or apps that would lead to a major drop in their likelihood to return to the site. To remedy the situation, we can trigger a live chat with a representative or send a follow-up email with an apology and even a compensation.
Our research shows that these interventions work, and therefore I strongly recommend organisations to focus on long-term customer metrics and measure them using predictive analytics.
How should a CDAO go about getting buy-in to embrace this kind of experimentation?
It all comes down to the CDAO’s ability to prove ROI.
From what I’ve seen, it is recommended to start small and specific. Select a concrete customer metric to focus on and improve. (For example, increasing customer retention by 5% in the next six months.)
Through predictive modelling, it is possible to estimate the revenue growth if the project is successful – and for it to be successful, a cross-team collaboration is essential. Product, IT, marketing and analytics teams all play an important role in any effort to improve the digital experience and reach a goal like increasing customer retention.
It sounds like there's a lot of overlap between the responsibilities a CAO and a CXO might hold. How should people in these related roles work together?
I think that these roles must have a strong collaborative relationship. The CAO brings the data analysis expertise, and the CXO brings the understanding of the customer. I often see that customer experience issues originate from poor communication in the organisation, so it’s clear that one cannot fully prosper without the other.
What would you say have been the greatest changes in the UK and Europe’s data and analytics industries over the past 12 months?
Something that I see more and more is companies starting to use advanced research methods such as machine learning and AI. They are not regularly used yet to understand customer behaviour. But eventually they will become ubiquitous.
I recently fielded a survey among CDOs and CAOs, and most executives reported that incorporating machine learning and AI into the organisation’s data practices is a top imperative for 2020, and I believe it’s a very strategic path.
How can data and analytics leaders ensure that they get the most out of their AI investments throughout 2020 and into the future?
Something that many CDOs struggle with is ‘so many tools’. Many agencies are popping up trying to provide machine learning services. But these are just more vendors that might create more noise than signal.
So, something that I really want to advance is to bring AI, machine learning and predictive analytics to where we're already used to seeing data and consolidating everything into one place.
Most companies use traditional analytics tools like Google Analytics, Adobe Analytics or more advanced tools like Quantum Metric's platform. Imagine how powerful it would be to bring everything to one place – how much time and resources it would save. This is our vision as Quantum Metric – one that will provide companies with a competitive advantage.