Use Text Analysis and Dashboards together!

Text / Comment / Verbatim analysis is a fantastic technology for understanding customer experience. 

Customer comments, whatever their source, contain lots of very useful information. They tell businesses about challenges and opportunities, frustrations and delights, poor processes and great. In fact text analysis systems are so effective now-a-days that their main challenge in the market is providing enough visualization capability so that their analysis results can be presented easily. And thereby, informing managements about actions they should be taking in response to the feedback that they are getting.

Visualization tools do a great job helping businesses drill down on complex data sets, then extract and present relevant salient facts in easy to understand visualizations.  

Text analysis provides the complex data sets, placing statistically valid structure around unstructured data (text and comments), in the form of topics discussed and sentiments expressed. Visualization tools drill down on that newly structured data and make it actionable. 

Visualization and Text Analysis should be used together. At Etuma we understand this. We provide some of our own visualizations.  But, importantly we have built in connectors to visualization systems that provide real-time data flow to dashboards built with them. Some examples of visualizations that can be built off of Etuma created data sets can be found in this excellent post on the Etuma Blog: 12 Text Analysis Visualization Tips. I encourage you to check them out.

Stewart Nash
GM Etuma USA
stewart@etuma.com

Smart Loop Closing with Text Analysis

I subscribe to Bruce Temkin's Customer Experience Matters blog. Its a good resource for folks who are "doing" customer experience. In a recent a post and report titled: "Use Customer Insight To Close Four Loops" Temkin made the argument that there are four distinct feedback loops that organizations need to evaluate and act on. As a proponent of Closed-Loop feedback management I am very much in agreement with Temkin's four feedback loops and approaches for "Closing" those loops. The Temkin article can be read here. He uses restaurant chain feedback as an example. 

Text Analysis can contribute to implementing loop-closing processes for each of Temkin's four loops. The four feedback loops include:

• Immediate Response
• Corrective Action
• Continuous Improvement
• Strategic Change 

• Immediate Response. Temkin - Rreach out to a restaurant customer who responded on a survey that "the bathroom was dirty" and help take care of her ongoing concerns.

How Text Analysis can help.

• In the immediate response scenario, feedback from customers of this type is typically delivered directly to the store manager, who then responses to the person providing the feedback or acts on solving the problem. Text Analytics only provides limited help in this instance documenting the issues a customer or customers are having.
• Corrective Action. Temkin -Get the manager or employee to clean the bathroom in that restaurant.

How Text Analysis can help.

• Though immediate corrective actions are likely being spawned by the immediate response process handled by store managers, taking broader corrective actions that prevent issues from re-occurring is something Text Analytics is very good at helping with.  For instance, if dirty bathroom is a regular occurrence at a particular restaurant or group of restaurants then there may be a training, process or staffing issue that higher level management needs to address. Text Analysis can also help understand what "dirty" actually means. i.e. is it full trash cans, dirty floors, dirty stalls, etc.. Text Analysis can often pinpoint what the most severe issue actually is, making it a lot easier for senior management to provide specific guidance to store managers on corrective actions to take.

• Continuous Improvement. Temkin - Create new process for restaurants to check and clean bathrooms on a regular basis.

How Text Analysis can help.

• Similar to corrective actions, Text Analysis can help a lot with continuous improvement efforts, by continuously monitoring and reporting changes in topic sentiment about various aspects of, in this case, bathrooms, Soap dispensers , hand dryers, cleanliness of sinks, floors and toilets.  All these topics that can be monitored for changes in sentiment (and using Etuma, can be filtered by region, district or restaurant as well) , enabling proactive management of "dirty" bathroom issues across the restaurant chain. 

• Strategic Change. Temkin - As part of new restaurant formats, design bathrooms so that they don’t require as much time from employees to keep them clean.

How Text Analysis can help.

• Here again, Text Analysis can be very helpful in the strategic domain, by providing insight about what "dirty" actually means for customers, which in turn informs how to strategically change for the better. For instance, if "dirty" often means un-flushed toilets, a strategic change could be to install self-flushing toilets in the restaurants.

Temkin's bottom line: Make sure to build out four closed loops.  

My point: using Text Analysis makes it a lot easier to determine which actions to take to "Close" each loop.  Also, actions taken will be supported by data, as text analysis is providing statistically meaningful support for the insights being generated. The result being that its much more likely that customer issues will actually be addressed in such a way as to improve VOC.

A couple of additional points....

Automated Text Analysis provides continuous and on-going evaluation of feedback.  When people have to do the evaluation of feedback without text analysis, they can be inconsistent, or if there is lots of feedback, only base their analysis on samples.  So, for organizations who want continuous improvement, continuous monitoring is necessary.

Changes in topic sentiment are difficult for people to identify unless the change is dramatic.  Automated Text Analysis solutions (such as Etuma) are able to assess and document changes to topic sentiment over short periods of time based on changes in the relative volumes of topics discussed in received feedback.

So, by all means, close Temkin's four feedback loops. But, by using automated Text Analysis, closing loops will be quicker, smarter and less effort than otherwise.

Etuma's 5 rules (and some comments) for creating a powerful customer feedback gathering system

My colleague - Matti Airas - published this on Etuma's Blog (Click Here to read the original). I liked the thrust of the post.  So, I thought I'd re-post it here and add my own commentary to it.  So full credit to Matti and Etuma on this. My comments in Red.

At Etuma, we have analyzed hundreds of different feedback processes and formats and seen what works and what doesn’t work. For a feedback analysis company, we have become surprisingly expert in the process of gathering feedback. We have learned how to design a survey process that both maximizes the volume of open-ended feedback and provides concrete actionable insights.

1. Make it easy and ubiquitous to submit spontaneous feedback  

Give customers the possibility to choose the channel they prefer. This includes text messaging, web forms, twitter, Facebook and email. Remember, the customer chooses the feedback channel, not you. Put the feedback URL everywhere. Don't be afraid to receive feedback, embrace it!  Needless to say, I agree with all of this. But, I feel its particularly important to trigger "outbound" feedback gathering at key touch points, even if there is not a transaction involved. As a personalized appeal to provide feedback is more effective than a feedback link customers glance at in passing. 

2. Run transaction-based surveys for key touchpoints

We like the transactional Net Promoter Score system (TNPS) but it can be any format as long as it is short and relevant for the experience (touch point) you are tracking. In transactional surveys the two most important things are timing and brevity: conduct the survey soon after the event using sms or email, and keep the survey as short as possible. Remember that NPS is not suitable for every transaction. For example, in customer support context, the Customer Effort Score (CES) is a better format.  Again, I'm in complete agreement with this approach. Having an e-mail or text based process as your feedback gathering core is important as it allows personalized and follow-up-able feedback to be gathered.

3. Dig deeper when you don’t have enough information

Customer experience management (CEM) platforms enable you to run sophisticated rule-based surveys. You should use this functionality to find out e.g. why people stopped using a certain product or when their survey response didn’t explain the reason for their reaction (e.g. NPS score) or provide sufficient information for root-cause analysis.  Amen to this. There are multiple ways to approach the need for additional "drill down" data. One is through rules based design. Another approach is to send follow-up surveys that incorporate original survey responses plus additional background data from your databases.

4. Run periodic relationship surveys on a representative sample

Spontaneous and touch point specific surveys often fail to get a comprehensive view on brand, competitor, pricing and marketing related issues. It is important to keep the relationship survey format as similar as possible to the transaction survey (although the periodic survey can have more questions). This gives you the ability to analyze customer feedback as a whole across all channels. Unless you have really large numbers of customers, say over 100K, today's feedback management systems make it easy to survey all your customers whenever you need to (typically 2x per year for relationship surveys).

5. Connect your company’s and competitors' Facebook pages and Twitter handles to the analysis service

More and more of the brand, product and service discussion is moving into social media. You need to connect your main social media channels into the feedback analytics process. Again, Amen to this. But, these channels are often not given the kind of serious attention that other processes are given. Gathering feedback from social media is important. Tying this type feedback into all the other feedback is critical.

6. My personal most important Rule about feedback gathering. Never ask a customer to give you information via feedback which you already possess (and which they know you possess).  

Nothing turns off a customer faster than a survey seeking redundant information.

Don’t try to get an answer to every question in one survey. Create a continuous high-volume communication process, in which the complete picture is formed from many small fragments. This is the best advice of all.

Follow up Boosts Survey Response Rates

This is a topic I've written about several times over the last 3-4 years. It's been my contention that response
rates for customer surveys are much higher when there is an expectation of  follow-up action on the part of the company issuing the survey.

Clearly, there are ways to boost survey response rates that don't include follow-up. Survey length, survey design, number of reminders and availability of mobile device support all effect response rates. But, in my experience doing customer surveys, about the best response rate that can be expected for customer survey processes not employing direct follow-up is 25%. In other words, if a business surveying customers does everything else right and doesn't follow up, 25% is a reasonable expectation for maximum response. Though, I expect few businesses attain that level of response rate very often on their survey processes.

In my experience, for businesses employing customer survey processes where follow-up is included, the maximum response level can be much higher than 25%.  I've worked with businesses for several years implementing customer survey processes built around follow-up. What I have seen is response rates that tend to rise over time (versus declining over time as is generally the norm in surveying). And, I routinely see response rates above 40%, with some even above 50%.

For me, the evidence is in that expectations for follow-up on the part of customers drives response rates higher over time. It would seem obvious that this would be the case. But, most businesses in my experience aren't likely to act on feedback often enough to give their customers this expectation.

The chart below shows response rates from one of my clients who I've conducted eight surveys for over the last five years.  All the surveys used an NPS approach, employed what I've described above as "good" design practices and included multiple reminders plus follow up. Follow-up was facilitated by the "Notifications" process embedded within the QuestBack Essentials platform I use to deliver the surveys (meaning that follow-up is triggered by response profile). You'll see the early surveys tended to have lower response rates and the later surveys higher response rates.

Two things have been going on with this particular client. First, customers have developed the expectation that their feedback is going to be followed-up. Second, the business has actually implemented follow-up processes and responds to people giving feedback. If you notice, in the early surveys, customers did not have the expectation of follow up (they had not yet developed it). And, frankly the business really wasn't all that committed to following-up, at least at first. But, between survey 3 and Survey 4 this client began to have lots more competition. They reacted by paying a lot more attention to survey responses, acting on them individually and thereby improving customer experience across the business.

Beyond higher response rates, this business has improved customer retention, overall profitability and survived the entrance of major competition into its local market. And, in many ways has become a stronger competitor by paying attention to customer experience through the survey feedback / follow up process.

A last note.  In this customers current survey process, response rate is at 53% and still rising.

- Stewart Nash
LinkedIn: www.linkedin.com/in/stewartnash

Verbatim Analysis - A crucial technology to use in business

My friends at Etuma recently put out a post titled "9 Reasons why you shouldn't wait to implement a verbatim analysis solution. Click Here to read the blog post in its entirety. I liked the article enough that I thought I should reiterate the points and expound upon the underlying needs for actionable data that companies today have. Summarizing the Etuma post:

1. Leading companies in many industries have well developed customer listening skills. Most customer listening remains survey driven, but increasingly it is becoming omni-channel (meaning surveys plus social media) are quite well. Your competitors are learning and transforming how they react to the customer's voice.  Your company needs to as well.

2. Customer needs are becoming harder to predict. You need to have data that can feed analytical and predictive analysis tools in order to more rapidly detect emerging trends. Verbatim feedback is one of the data sources predictive analysis relies upon.

3. Paying people to decipher and report on customer interactions provides expensive and inconsistent data.
Feedback enters the call center from phone, emails, web forms and chat logs. Once calls are transcribed it's all unstructured open-text data and can be fed to verbatim analysis systems.

4. Front-line staff turnover is often quite high in contact centers - compounding the issues in #3 above by requiring constant training and monitoring to provide some consistency of results.

5. Many companies do not get customer feedback directly. As most of their products are sold indirectly, feedback is only as good as the filtering mechanisms that exist in their distribution networks. i.e. It is usually not good. Verbatim feedback lets product vendors monitor 3rd party review or web sites and detect issues and trends without feedback being filtered by the distribution network.

6. New products and services are introduced constantly. Detecting the presence of a new product or its impact on existing offerings is a lot easier when that information is extracted from customer comment automatically.

7. Just like new products, new competitors are much easier to detect when data sources are being analyzed and regularly.

8. Omni-channel marketing and sales makes the customer journey complex. Shopping has become more complicated. The customer journey can now involve many interactions with the company. Capturing those interactions and understanding them consistently and quickly is important and much easier with automated verbatim analysis.

9. Social media complicates communications and crisis management.  Getting quicker understanding of issues makes reacting to them a lot easier and the reactions can be much better planned.

Etuma makes the general point in the post that topic / issue detection and sentiment change detection are critical capabilities for organizations to have these days. And, especially so where a company doesn't get direct access to feedback.  The point I would make about the urgency of implementing text analysis capabilities is the "Don't know what you don't know" factor.  Meaning that without tools to detect issues and sentiment changes (Things you don't know), businesses Don't know What they Don't know.

In today's world, not knowing something important for any length of time tends to have associated costs. By the time a business learns what it needs to know about something, it may be too late to fix a problem, create a new product, add a new service or otherwise react to customer needs in effective ways.

The urgency of adopting automated text analysis solutions is clearly high.  Hopefully more businesses will do so sooner rather than later.