AI Layer Reduces Digital Agency Customer Churn
Product used: bespoke customer service chatbot.
Our client is a large digital agency based in the UK. They sell digital products like advertising, PPC, SEO, design, and others to thousands of small businesses.
In early 2017, our client came to us with a very specific problem: too many products, too many reports, not enough insights. It’s a typical situation most large businesses that sell goods and services find themselves in. As the business grows, products are added to the offering to generate more revenue through up and cross-sells. Unfortunately, this also makes things complicated. This is a perfect AI layer opportunity.
Immediate money-making actions
Applying a layer of artificial intelligence across all products, reports, and customers of a large business is no small task. It takes time to understand the products, centralise the data, play with learning models, run internal tests, etc.
This is usually a 12 month project, one that must be split into achievable chunks.
Finding quick wins is a big part of our work. As we start digging through a client’s data, our first goal is to figure out the immediate money-making actions we can make right now. With this much to do over so many months, where do we start?
Centralise reporting across 32 products
For this particular client, this action was to improve reporting. They have 32 services they sell to over 60,000 customers, each of which has its own data format, language, and reporting template. This is extremely confusing to most customers (and, as we found out, most sales reps). Numbers and customer feedback shows this confusion prevents them from renewing and purchasing more services.
This was a perfect first step.
We first sat down with the reporting rep of each product to understand their logic, the numbers to report on, how they calculate them and what their language means.
Armed with this knowledge, we were able to grab a sample export of each report and start working on centralising the data. Our aim was to produce one sensible reporting format any of their product (and combinations of them) could use.
The new reporting format we produced makes sense to their customer base. It uses as much ‘plain English’ as possible, is easy to read and display the data points that really matter to them. This is a huge win.
We ran the new reporting format on a test pool for 5,000 customers for three months. Our client noticed a significant increase in stickiness and reduced churn in their core customer segment.
As a byproduct of this improvement, we found sales staff now manage sell more. The ‘plain English’ reports give them a more solid ground to stand on when up-selling further services and able to keep their customers happy, engaged, and spending money.
Start improving your customers' experience today. Talk to chatbot experts.