Seth Grimes interviews Ipsos Loyalty SVP Trish Dorsey about her work, CX trends, and futures.
By Seth Grimes
Customer experience (CX) is the set and sum of perceptions formed through interactions with a brand, product, or service. CX starts with awareness created through advertising, word-of-mouth, and social buzz. It is confirmed in the course of online, in-store, and point-of-sale and point-of-service interactions. Central are perceptions of packaging, quality, and value, factoring in cost and relative to the competition. CX affects customer satisfaction, loyalty, and advocacy.
Ipsos Loyalty is the customer experience practice within Ipsos, a global agency that employs more than 16,000 people and conducts research in more than 100 countries. Ipsos Loyalty’s value proposition: “We specialize in all matters relating to measuring, managing, and improving customer relationships. We help our clients manage the experiences they deliver in a way that maximizes the value of customers to their organization.”
Ipsos Loyalty SVP Trish Dorsey
Measurement can get tricky. Modern CX involves dozens of touchpoints and feedback sources, among the latter surveys, reviews, social media, and contact centers, and analyses apply a suite of techniques that both optimize interactions and the overall customer journey, the path an individual takes from needs awareness to purchase and, in the case of a happy customer, loyalty, and a lasting relationship.
Given CX’s importance and Ipsos Loyalty’s prominence in the CX world, I’m delighted that Ipsos Loyalty SVP Trish Dorsey is speaking at the up-coming Sentiment Analysis Symposium, a conference I organize, and to have an opportunity to interview her about her work, CX trends, and futures. This interview, then, is a look at –
The Role of Measurement, Insights, and Loyalty in Customer Experience
Seth Grimes> You’re SVP East Region at Ipsos Loyalty, a global custom research agency. Would you please say a bit about your work?
Trish Dorsey> I’ve worked for Ipsos for 3.5 years and have 25+ years of experience in the research industry, across multiple industry verticals. Most recently, I have been supporting the customer experience measurement needs of the technology and telecommunications industry. This has been especially exciting for me personally, as companies in these verticals have been ‘reinventing’ themselves of late. It’s an incredibly dynamic space and that makes the customer experience challenges really fun to explore.
Seth> What do you mean by ‘reinventing’? How have customer experience goals, methods, and data sources changed in the years you’ve been working in CX?
Trish> All of us have seen the world of customer experience change. CX has become more important as evidenced by our own expectations as a consumer (thanks to brands like Apple and Amazon) and how the power of social media/the internet, etc. makes the importance of getting it right even more important. Consumers have access to more information and can make more informed decisions. Social media gives them a voice/power over brands they never had before, one customer’s feedback on a negative experience being able to influence millions.
At the highest level, I’ve seen the need for and reliance on customer experience measurement and management increase significantly in the 25+ years I’ve been in research – and at an especially aggressive pace in the recent several years. You don’t have to look far to find a CEO who will say that understanding his/her CX dynamics is one of his/her highest priorities! And the fact that many organizations are formally assigning a C-Level position to CX is even more testament to this fact. Moreover, the nature of the need and reliance has changed – no longer is it “as simple” as designing good surveys. Rather, organizations investing in CX are looking for partners who can provide research expertise, advisory services, and technology support.
Four stages and five outcomes: an Ipsos Loyalty customer experience conceptual framework (Ipsos Loyalty image)
Seth> And the data side?
Meanwhile, the numbers and types of available data collection tools have increased as well. Of course, we’ve managed the move from CATI (for most audiences) to online and now to mobile in the last 20 years or so. Add to that our continued evolution in use of other data collection capture methods, e.g., video, audio. Likewise, we also no longer restrict ourselves to structured survey responses for our CX programs. We’ve seen an increased interest in leveraging unstructured data sources – CRM behavioral data, social listening data, IVR, telemetry – to provide new data streams to provide insights around our client’s CX performance.
And all of this has happened as the availability and sophistication of technology platforms has exploded of late. No longer are we limited to back-office data collection platforms. Rather, clients increasingly require access to data in real-time; in highly visualized format; and in highly democratized ways.
Seth> Data analysis?
Trish> Of course, these dynamics mean that our analytic techniques have had to evolve as well. Certainly, we maintain analytic rigor in all that we do, but there is also increased focus and demand for real-time (or near real-time) analytics. Other than the traditional statistical frameworks, this has meant that we’ve had to invent new data science techniques and algorithms to accommodate this dynamic.
At the end of the day, all of these dynamics make the design and management of a CX program exponentially more difficult while at the same time making it more important to get right. And for those organizations who do, there is a financial gain – brands who provide a better CX have higher returns and outperform their competition in the stock market.
Seth> You’ll be speaking at the up-coming Sentiment Analysis Symposium on “Using Sentiment Analysis to Identify Emotions, Provide Insight, Enhance Customer Experience and Prevent Churn.” What emotions and what sort of insights are most relevant for your telecom clients?
Trish> I would say that we build our buckets from the bottom up by looking at which combinations of issues and emotional response have the highest impact on KPIs. I suppose that means that we don’t really use any specific model, per se.
Seth> Do the same measurement priorities apply in other industries, for instance, retail, financial services, and hospitality? I guess what I’m really asking is, how uniform are CX practices, and satisfaction, loyalty, and similar measurements, across industries?
Trish> At the highest level, the CX ecosystem we use to contextualize our measurement is the same regardless of industry. That is, this ecosystem suggests that we need to think about four levels of CX for measurement: relationship, touchpoint, transactional, operational. These levels exist regardless of industry. What might differ is how we think about the interaction among those different levels and how much measurement attention we give to each. For example, in those industries where the purchase experience tends to be more transactional (e.g., hotels, restaurants), more time is spent on event-based measurement while in industries where the experience tends more towards the relationship-level (e.g., investments, automotive) there might be more measurement at the relationship or touch-point levels.
Likewise, in terms of actual measurement, the conceptual frameworks we apply are largely similar across industries. NPS is NPS, regardless of industry, for example. The differences across industries come mostly in the types of corollary and diagnostic metrics we use or how we might design some of those. For example, in industries where there is multi-brand usage (e.g., hotels) one might use Share of Wallet as a way to measure strength of relationship, whereas in a market where usage is more monogamous (e.g., wireless phone providers), usage, total spend, or consideration might be appropriate metrics.
Measurement points and performance indicators in the CX ecosystem (Ipsos Loyalty image)
Seth> I interviewed an Ipsos Loyalty colleague of yours, Jean-Francois Damais, in late 2015. J-F is heavily into text analytics, the application of natural language processing (NLP) to produce insights from text-extracted data. What techniques are most important for you, in your work?
Trish> On any engagement, we leverage the techniques that best suits the objectives, budget, and timing constraints.
First, we take inventory of available data – what data does the client already have? – to answer the question at hand, and in what format do those data exist? If there are information gaps, we identify the best source (given budget and timing) to fill in those gaps. Who do we need input from (e.g., B2B or B2C customers)? Do we require structured survey data? Or can insights be culled from unstructured sources such as text analytics or social media listening or passive measures like online/offline shopping behaviors – this is the key to omnichannel – and so on.
More and more, however, it seems as though we are looking to unstructured, passively-collected data to address our business questions as survey response rates continue to decline and willingness to provide feedback in longer surveys wanes.
Of course, thinking about what kind of data are available is only part of the equation. We must also determine the analytic methods that are most appropriate to answer the business questions and we evaluate those methods within the context of what types of data are available. Linear regression? Logistic regression? CHAID analysis? Conjoint-based choice modeling? Text analytics mining? Etc.
Seth> Ipsos republished my interview with Jean-Francois under the title Don’t Kill the Analyst Just Yet, suggesting that human judgment and perhaps domain expertise are key assets. Has machine learning started to change your work? Or if not machine learning, are there other technologies or practices on the horizon that show promise of disrupting or revolutionizing the research and insights industry?
Trish> Absolutely, machine learning has provided new and unique tools for our analysts and the sophistication of these tools continues to improve at a rapid pace. AI is being used in many applications, for example:
- Many companies are using AI/machine learning to know their customers and predict their behaviors.
- Others are using AI/machine learning to predict demand for their products.
- Still others are using AI/machine learning to manage dynamic pricing based on changing market conditions.
That said, human judgement and domain expertise remain crucial to the development of hypotheses to be tested by these models; interpretation of model results; and activation of results towards effecting business outcomes.
Seth> Thanks Trish!
For further reading, check out The Essential Steps for Building and Maintaining a Best-in-Class Customer Experience Culture, written by Trish and Ipsos Loyalty colleagues. And to meet Trish and a few dozen other cool speakers, join us at the Sentiment Analysis Symposium, June 27-28 in New York. Use the ID code GREENBOOK for a 15% discount on your symposium registration.