The ability to analyse vast amounts of data and then automate decision-making processes based on the results is a huge opportunity that everyone in fintech should be exploring.
Technology in the service of people
At Simply Business, we’re integrating machine learning and attendant techniques into our everyday business. We’re major advocates not only for the use of new technology, but also for collaboration between this technology and humans.
We’re not interested in replacing people with algorithms, particularly when we’re dealing with clients as personal and unique as the UK’s small businesses. Instead, we want to find ways to put technology in the service of our customers and of our own employees.
We’re doing this in several distinct ways, and I’d like to explore one in particular. We have a machine learning guru who was the first data scientist we hired, way back in 2015. One of the most impactful projects she’s worked on is the use of a model to help us prioritise leads.
How can machine learning help customers?
We have a large contact centre based in Northampton, which is a key part of our relationship with customers. We understand that small businesses often want to talk personally with an expert, and our contact centre lets them do this.
But managing volume is a challenge for us. We originally wanted the new model to give us the tools to prioritise telephone leads based on every scrap of information we held about a prospect – the type of business, their revenue, the age of the business, and so on. Crucially, it also prioritises based on their existing interactions with us, such as their time on our site, the number of pages they visited, the channels they came to us through, and more.
But we came up against a problem – our telephony system wasn’t up to the task yet. So we brainstormed the best way to use the tech that we’d built, and we decided instead to experiment with it in tandem with our remarketing efforts.
Giving customers a tailored experience
We identify cookies using Google Analytics. But now, with our machine learning techniques, we’re using this technology to track and prioritise prospects in far more efficient ways. As soon as a visitor arrives on our site, we use machine learning to give them a priority score. This score changes based on their engagement with our site – it’s constantly updated, depending on how the user behaves. They’re then placed into one of eight groups, from ‘most likely’ to convert to ‘least likely’.
This tool has had a concrete impact on our marketing strategy. We’ve stopped marketing to leads with the lowest likelihood of conversion, and now instead invest that money in providing even more tailored messaging to those more likely to convert – and we’ve seen double digit improvements.
We’re also using machine learning on our content side. Our Knowledge centre is an important acquisition and engagement tool for us, but we were lacking a tool that could help us serve useful recommended articles to specific segments in a seamless way. We’ve built some tech that lets us do just that and with our in-house machine learning recommendation algorithm, we’re recommending the right article to each customer – and we’ve already seen positive results.
Where do we go from here?
That said, some of the technology isn’t quite there yet. We’ve been experimenting with providing tailored ad copy to specific remarketing segments using Google’s Ad Customiser. However, the product doesn’t work at scale – and that’s a problem when you’re trying to tailor copy for as many as 1,000 different audiences. We’ve been working with Google, asking them to expand their customisation options, and we hope this will happen soon.
Machine learning and related technologies are only going to grow in importance, especially for a tech-led company like Simply Business. We’ve seen real and measurable benefits from our investment in machine learning. If you’re looking for a way to understand data and automate tasks to deliver serious bottom-line results, this could be it.