Are you considering text analytics for your Voice of Customer programme? If so, you may already have discovered the minefield that is “text analytics”. If you’re bamboozled by the amount of information out there in the form of blogs, trade articles, conference presentations et al, we’re here to help you make sense of it all.
Text analytics is the buzz phrase of today and it can be difficult to know where to start. We’re here to try and guide you through the initial maze of jargon and pre-purchase conversations you may have.
Is Text Analytics for you?
Text analytics is a hugely powerful and fast solution for analysing high volumes of customer feedback comments. If you cannot effectively analyse your comments using manual resources, yes, text analytics should be considered as you may be “missing out on customer gold”, as one of our clients once said.
Do your comments contain valuable information that would help you improve customer experience?
By looking at a small sample of comments, you can gauge whether there are quality insights within them which, if analysed and effectively reported back to you, could be used to improve experience. Not all feedback merits using the technology but if yours does, text analytics may be the answer.
Do you need text analytics or insight?
Words such as ‘sentiment’, ‘emotion scoring’, ‘impact analysis’, ‘topic coding’, ‘natural language processing’ are commonly used to sell text analytics software. (Download our Text Analytics Product Guide for more information). Not only do you need to understand what these phrases mean, you need to understand whether these things will help you achieve your goals. If your goal is to gain insight into what drives customer satisfaction and loyalty, you need more than a pure text analytics tool. You need a solution that provides the insight in a manner that you can understand and act upon. Effective reporting tools are vital to gaining insight into the text analysis results.
Do you need an out-of-the-box solution or a proprietary text analytics solution?
Some suppliers use machine learning and Natural Language Processing methods for analysing text. Others use propriety software that they have designed and developed specifically to address the needs of their target market.
A proprietary solution such as Feedback Ferret uses humans for quality control and updating their Lexicon. This “human + machine” approach tends to deliver consistently superior results compared with out-of-the-box NLP solutions. It can also be customised to meet the requirements of different clients in different industries. For example, brand names, products, locations, outlet names, abbreviations, staff names, competitor names, etc – and all the variations and misspellings – can all be added into a proprietary Lexicon to enable more accurate categorisation of your company information.
What should you expect to glean from a demo with a supplier?
Most suppliers offer a very top level demonstration of their solution as a starting point. This is a good opportunity to get to know more about the software and the company’s approach to text analytics. However, in an initial demo, the chances are the data used won’t be yours. So at this early stage, it is almost impossible to judge whether the solution is the right one for your data.
Any solution provider worth their salt should be able to offer you a Proof of Concept using your own data. This is an opportunity for your data to be put to the test and run through the text analysis process. You will supply the data and see the results. If you can see value in the results, then the solution is worth looking at more closely. If for any reason your data is not ‘fit for purpose’, the solution provider should let you know once they have reviewed the data.
At this point, it’s important to consider that if using your own data in a Proof of Concept and the turnaround time is live or instant, the accuracy of the text analysis coding should be called into question. Either the levels of granularity are such that the results will not give you the insightful results you are looking for or you are going to have to do a lot of your company specific coding yourselves.
How much coding and data processing is involved on your part?
Some solutions require you to do the text coding, data management and data processing yourself while others do this work for you. These are tasks that must not be underestimated in their complexity and time required to undertake. Find out exactly what is involved, talk to other clients of the solution provider and make sure you fully understand this process. If you don’t have the resources in-house, consider an end-to-end solution where all this hard and complex work is undertaken for you and you are simply presented with the results.
Who analyses the data?
Some solutions simply present you with the raw post-text analysis data and leave you to interpret it in whichever way you need to. Other providers give you tools to enable you to carry out further analysis and data visualisation of the text analysis results in a format that is easy to digest and take action on (dashboards, regular PDF reports, etc).
It’s also worth asking if reports are standardised for all clients or whether they can be customised to suit your business requirements.
Automated Hot Alerts based on the text analysis results can be one of the most powerful applications for creating value from the Voice of Customer / Text Analysis project. Being able to quickly identify issues and act on them – based on what customers have said – can be invaluable for improving customer experience, staff behaviours and improvements to products, services and facilities.
Do you have to change your feedback surveys if you are going to use text analytics?
Some text analytics suppliers demand you use their feedback surveys, others are more flexible. If you have to use the supplier’s surveys, check the surveys they use satisfy your business needs. If not, look for a supplier like Feedback Ferret, who is flexible enough to analyse the data from your current surveys and other Voice of Customer feedback channels.
It is always worth asking advice from suppliers about your surveys and how you can improve response rates and increase the quality of feedback. Gathering feedback for automated text analytics puts a different spin on how you ask for the comments compared to traditional market research orientated surveys. The days of “Any other comments?” should be well and truly over!
Is Impact Analysis important?
Impact Analysis identifies the topics that customers talk about that have the greatest influence – positively or negatively – on customer loyalty and satisfaction metrics. Whether you need Impact Analysis modelling depends on your business needs. If you are looking to identify the key promoters and detractors of customer loyalty, Impact Analysis is an invaluable tool. It enables you to prioritise your actions and focus root cause analysis to stop or fix the problems that are dragging down your loyalty and scores.
Some providers include Impact Analysis as a standard part of their service, others may charge extra for this additional layer of complex statistical analysis. Ask to see evidence of how Impact Analysis has benefitted organisations and fully understand the way it can add value to a Voice of Customer programme.
Can text analytics be used if you have more than one feedback channel?
Consolidating multiple feedback channels is something text analytics does very well. This is a basic requirement of many of Feedback Ferret’s clients who operate several different surveys across different media (email, SMS, web, telephone etc). They want a 360 degree view of their customers’ experience – sometimes including feedback from social media – at a level of granularity and detail that is useful and useable.
Do you need to change software and systems used to accommodate text analytics?
Some providers insist you install their software to use their solution. Others are much more flexible and integrate with your CRM or other systems as required.
Navigating your way through the text analytics minefield without a compass or a map can be challenging. Here at Feedback Ferret, we offer sound, honest advice based on 15 years of experience gathering, analysing and reporting on customer feedback comments. We will work with you to fully understand your requirements and implement the best, fully customised solution to fit your organisation’s requirements. Our team of friendly, flexible and resourceful account managers and analysts will support you every step of the way. We won’t sell you something that can’t help you achieve your business goals.
By Nicola Douglas, Marketing Manager