I was stimulated to write this blog after I read an excellent post on LinkedIn by Dave Grow (COO of Lucid software) giving advice about the importance of doing things that didn’t scale to help grow his company. It’s apparently common advice given out to high growth companies and start-ups, but it had never resonated with me before, until now.
Dave Grow made his “un-scalable” investment with a commitment to read customer care trouble tickets every day. He managed 100,000 over a 7-year period! That’s an average of over 50 tickets a day, depending on your definition of a working week! By any measure that’s an enormous investment for a CxO to make, or anyone else in the company with a day job for that matter. He argued the investment paid back many times over and said it enabled him to gain a unique and deep understanding of his customer’s experiences, helping him to directly drive growth and improve his business.
It struck a chord with me and made sense, but I thought it highly unlikely many busy people would make that level of investment. However, what if it were actually possible to surface all the hidden unstructured dark data deep within trouble tickets without actually reading it. Imagine, in a couple of clicks of a mouse it could be made immediately available and accessible to many users across the business as insights? A system that could instrument the unstructured free-text data automatically with discovered topics, creating common themes and new categories. This sort of capability could “bubble” up new populations of potential problems, or clusters of auto-discovered keywords and topics. This is the sort of tool that would turn unobtainable, hidden data into a set of immediately consumable insights, for care managers and other interested stakeholders across the business. This is doing the un-scalable but in a scalable way! Think it might be useful? Read on.
There is a myriad of Helpdesk or Customer Support engagement solutions, more than can be spoken about sensibly in a short blog. Most of them are in some fashion trying to track and solve customer problems and help manage the overall customer experience as they come into contact with a brands products and services. These tools, and the people that operate them, represent an important source of data, information and knowledge about the customer across many touch points.
Unfortunately, a large chunk of the data associated with each customer transaction is stored away as raw data, in the form of free-text descriptions or voice transcripts and remains generally un categorized and un tagged. This sort of data is hard to extract and turn into information or gain knowledge from quickly and easily.
It is hard to bring to bear by a manager who is trying to identify clusters of new problems or patterns of different behavior, or just discover what is the burning issue of the day. It’s essentially dark data that needs processing to help it become useful customer information and knowledge. A perfect job for an AI solution providing both supervised and un-supervised machine leaning to improve information and knowledge.
Currently 2nd & 3rd line support experts who, knowing what to look for, dig deep into trouble ticket data, when solving a specific problem. However, a Customer Care manager, or COO or Marketeer for that matter, is not able to leverage the information hidden in the data and search it at an aggregate level or report on it using a BI tool easily. Only the data that has been previously encoded and tagged with meaningful topics or keywords is accessible like this. It’s the sort of hidden insight that Dave Grow was talking about benefiting from by reading all those trouble tickets.
Managing through KPIs
The industry has developed a number of important metrics to try and help understand and manage the customer experience and their advocacy. E.g. Volume trending and ratios of open and closed problems, Mean Time to Repair trends, Customer Satisfaction scores (CSAT), Customer Effort Scores (CES), Net Promoter Scores (NPS) to name a few. These are all great measures, but they still miss the mark in terms of revealing that extra hidden value in the underlying data. Recently there have been moves to improve the capability of tagging and categorization of support tickets both at creation time, and afterwards. Unfortunately, today up to 70% of trouble tickets still remain un-categorized or wrongly categorized! Let me repeat that, 70% of all your customers conversations and problems are stored away without any simple way to extract insight gained during the care interaction. The overused iceberg picture springs to mind!
Hidden care opportunities
So what sort of value is hidden in Customer Support tickets? What can Support Managers achieve by properly analyzing this hidden customer data?
- Insights, Metrics and Trends for Management: Getting a snapshot of what is actually happening with the ability to help visualize, prioritize and manage issues and problems. Identify what’s hot right now? Why are your customers calling? What issues do they report? What are they asking for? What are the emerging issues? What needs escalating? Where should the teams focus be going? A capability like this analyzes 100% of all your ticket to bring its content to the surface, automatically tagging tickets with relevant topics, product keywords, and other categorization. Making it useable and actionable.
- Training agents: Agent monitoring to improve training and knowledge transfer. With a highly granular break down of topics against each ticket, it becomes fairly obvious which agent is performing, what problems they are identifying and solving and where the focus or need for training is required.
- Digital Transformation: Self-Service and call diversion to a Bot, FAQ, or other self-service content, is a key part of becoming more lean and digital. A system like this provides detailed breakdowns of new topics, categories and themes and leverages machine learning to identify potential self-service candidates. These specific trouble tickets can be automatically tagged back into the support systems for automation and self-service handling.
Artificial intelligence can now be used to automatically analyse and trend your customers problems using natural language processing to understand the context within which a sentence is used and match this with semantically similar sentences to create knowledge and insights within your care data. The future of better understanding our customers and their issues lies within AI and will be a vital tool for Support managers and the wider business to improve the overall customer experience and make our customers happier.
But of course, if you just like reading support tickets and have 7 years to spare then best stick to it!