Combining qualitative and quantitative methods for data-savvy, faster and more reliable insights for your Customer Journey project
CUSTOMER JOURNEY: THE RESEARCH
SHOULD I USE A
CUSTOMER JOURNEY MAP?
A customer journey is a very helpful tool to empathise with customers and plan ahead how to manage customer experiences.
With a proper Customer Journey Map you are able to promote a culture of customer-centricity in an organisation.
In practice, it means different teams start looking at the same direction: the customer's needs, wants, fears and motivations towards the problems that your organisation is able to solve.
In a step further, a Customer Journey Map serves as a high-level, bird-eye compass for operations. Once you attach Key Performance Indicators and live metrics to each stage of the journey, your governance model is achieved: lean, broad and in real-time.
METHODOLOGY
I have used this methodology to map experiences in some of the most iconic companies in the Nordic countries. I have used it in fashion, railways, airlines, retail and more.
My Academic background in culture and technology and my experience with Analytics has helped me to combine quantitative and qualitative data for richer and more reliable insights.
I usually start with an internal investigation in the organisation. I then proceed to sample the essential profiles of your customers, so we create a journey that represents most of your customers. I then train these users into agents for reporting what we need from their journeys, and send for a service safari. I use a minimum of n=6 to work the discoverable problems. We will look also into search behavior and analytics to corroborate information and gain insight on market sizes and opportunities.
Here is a comprehensive step by step for building maps that help you to design and orchestrate remarkable customer experiences.
TL;DR
YOUR METHODS TO GATHER INSIGHTS ON WHAT HAPPENS DURING YOUR CUSTOMER'S JOURNEY WILL BE STAKEHOLDER INTERVIEWS, SEARCH BEHAVIORS, AND SERVICE EXPLORATION.
STAKEHOLDER INTERVIEWS
BE THE SNOOPER
This is when you will learn how the people you are working with see the project. You will be a detective, trying to understand what salesmen are trying to push, what product owners want to achieve and what management is after on the long run. You will learn their intentions, and help them to align and understand one another.
Points to consider
You will be their shoulder to cry on.
Understand you are in a political discussion.
Trust the statements, but verify.
STAKEHOLDER INTERVIEWS
Methodology
In-depth interviews
Nature of data
Qualitative
Highly interpretative
Goal
Ground recognition
Optimal number of interviews
Interview all decision-makers involved.
Until answers start repeating (saturation).
Usually less than 10 subjects.
Resources
A week planning and recruiting
A week conducting interviews
2-3 days reporting data
USER INTERVIEWS
BE THE PSYCHOLOGIST
Now it's time to select your first segment of potential customers and gain insight on how they feel about your ideas, products or prototypes. You will truly empathise with them here, understand their context, their way of living and how they generally behave around the concept.
There are plenty of guides on what to ask users. The most important takeaway here: understand the goal of in-depth interviews.
In-depth interviews are perfect to understand common characteristics of a customer segment. They let researchers look into the common sense of a specific segment of clients.
These interviews reach general, established truths. They bring clarity to ethnographic profiles, socio-economical contexts, prejudices, general impressions and so on.
They shed light in what tends to be the general impression of a segment concerning a service, and answer a lot of why-questions about the service and new solutions proposed for discussion.
Points to consider
Don't let the end-users create solutions. Listen to what their problems are. Be contextual: understanding what is not said is as important as what is properly said.
PRO TIP
SELECTING INTERVIEWEES
If you want to interview existing customers, go for the best possible customers. That will give you the perspective of people who really love your product.
If you are exploring a new segment or product, go for representatives of this segment that spends significant time or money with a competitor or substitute solution.
Never recruit based on demographics alone. The best sources are the ones who are, somehow, engaged with what you can possibly offer.
IN-DEPTH USER/CUSTOMER INTERVIEWS
Methodology
Interview saturation
Nature of data
Qualitative
Highly interpretative
Goal
Insights used to form a hypothesis about the user's problem or early possible solutions
Optimal number of interviews
Until answers start repeating.
Usually less than 10 subjects.
Resources
A week planning and recruiting
A week conducting interviews
2-3 days reporting data
WHAT DESIGNERS WON'T ADMIT
User interviews are fairly easy to setup and it is even easier to get carried away by them. Clients tend to get involved and the designer's life gets easy. That is dangerous, because confirmation bias steps in: nobody wants to be the dissonant voice, and the real goal is missed.
Let's get it right: interviews are wonderful to understand if services are usable. They are great methods to generate hypotheses.
When it comes to certainty, they fall flat. "It seems this could be a good idea given a few circumstances", you may state after these interviews. Not much of an insight.
On the other hand, good interview processes get the designer excited and saying "hey, we now have many ideas to work with!".
Qualitative interviews are very unreliable when used to confirm if users would actually buy a product. You simply cannot ask users to emulate a decision. Unless they open their wallet and place a pre-order, you cannot take in account success market based on a few discussions. The sample being ridiculously small when compared to the market is also a big problem in this respect.
In this respect, interviews are heavily biased. And paradoxically, the bias comes as a side effect from empathising with the interviewer. Generating empathy is the purpose of the interviews — so better let interviews answer the right questions.
SEARCH BEHAVIOR
BE THE DATA SCIENTIST
This is where you can really differentiate from the vast majority of service designers out there.
Marketing people are very good with search behavior tools (they call it SEM, or Search Engine Marketing, or SEO, Search Engine Optimization). They use it to understand the exact keywords customers use to look for or discuss a product.
As Service Designers, we are not looking into search behavior for these purposes. We are looking for insight on the way people think, search and refer to a product, service or industry.
Access to this data is unprecedented, and greatly unbiased access to the collective human brain!
This is the step where you leave the 1-6 users sample and access much larger samples. So you are not evaluating contexts or common sense, but validating the common sense assumptions you made after the interviews.
Furthermore, this is what gets you a seat around the strategy table. You are able to tell how many people per month look for that kind of problem, product or issue.
These are clear indications that there is, possibly, a market for your client's idea — or not. How valuable is that?
EXAMPLE CASE
LAUNCHING OUR OWN
LIQUID HAND SOAP
Here's a fictional case where a client wants to launch a new line of products, starting with liquid hand soap.
If we check Demographics.io (a free tool), you will know that in the U.S. this is a popular search among 45+ women. We don't have to restrict ourselves to this audience, but isn't that an amazing insight, attained in 5 seconds and free of charge?
Now, let's check how many people actually look that up in the U.S. Any market is possible to check, but let's stick to the example.
Using SEMRush (about €99 per month) you are able to understand how many people search Google for the term, and if there's a fall or rise in trends.
12,000 people searching for your product category every month may not be bad at all!
The rise in trends is very sharp, and you may inform you client that they are actually onto something: the rise starts on March, 2020, and it may be one of those products that Covid lifted up. So let's check, also for free, in Google Trends:
It's pretty hard to argue with Google Trends!
Now, a good market analyst may gain the same insight by surveying thousands of supermarkets. But isn't it great that a service designer can give you these lean metrics so quickly and at the right time?
But we can dig much deeper. Here we can understand from SEMRush how people actually search these products, and reflect your user interviews:
This examples brings a few insights from Search Behavior to reflect with your previously conducted user interviews:
12K persons look up this product online. That tells that they are not yet loyal to one source only, so there is opportunity.
Clear byproduct opportunity. 5.5K people are looking for refills! That hints for a strong satisfaction with the product category, as half of them want to repurchase in some way.
Furthermore, it hints that you may already start thinking of a byproduct, the refill — and even opportunize monthly recurrent revenue.
Luxury or health? Antibacterial emphasis is the highest ranking attribute searched! Should it look more like perfume or have that clinical purity aesthetics?
Substitutes? One substitute may be homemade products. Positioning as handcrafted may alleviate this, as long as the price holds. Curiously, the bar soap doesn't look very threatening when in immediate comparison.
You get the point. Can you see the power of Search Behavior in Service Design?
EXPLORATION & FEEDBACK
BE THE COGNITIVE SCIENTIST
This is a really fun part of the process, and the way you approach it depends on the stage of your product development or the nature of your business.
First, recruit "agents"
You need to recruit people and give them a crash course on how to use the tools you will give them. A base instruction is: simply report when something affects your mood.
Users should keep a diary of the experience and report emotional changes and what triggered them.
How many, and what profile?
It is a well established definition that 6 persons are enough to uncover most usability problems on a service. Apply the same rule for real-life businesses!
As for the profile, do not get caught up in segments, here: map your service based on humans, not on specific niches. You want to uncover your essential customer journey, the one performed by most customers. Thus, look for the goal of ground exploration.
THE SPLIT
4/5
BIOFEEDBACK IN MOMENTS
PHYSICAL SERVICES
Mapping a physical is an incredible process. The main goal is to understand how physically and emotionally your customer experiences the service or product.
HOW IT WORKS
In here, you will recruit users to do some "mystery shopping" — that is, they have a task at hand to complete — and report how they felt along the way.
This can last an hour shopping or a week waiting for something to arrive from the mail.
You may use gadgets to collect physiological data: FitBit for heartrate and steps, MoodMeter for reporting emotions and a GoPro to record everything, if that is possible.
After the process is over, you will conduct an interview where you discuss how they feel about the service.
Report emotions: Mood Meter
I use a brilliant app that costs $0.99 and was developed by the Department of Psychology of Yale University. Users are able to plot how they feel very easily, in very nuanced ways. It also timestamps the usage so you know exactly when each thing happened.
Measure physiology: Fitbit
Retail is a physical experience. I like to know a day in life of the user, but also a day in data. I want to know their heart rate when they reach the store from their car (Is the parking lot 2km away? Are there big staircases?), at what point they get thirsty, at what point they want to go to the restroom and this type of physiological data. Creepy? No, scientific!
REVIEW
Once the agents perform the task, you will sit down with them separately and discuss how it went
it's all part of a mosaic of data that, when put together, give a comprehensive view of the experience.
Yale's Mood Meter App Recruit users and equip them with the app, asking them to report emotions as they experience the service.
EXPLORATION & FEEDBACK IN A CONTROLLED ENVIRONMENT (RECRUITED AGENTS)
Methodology
Self-reported emotions
Geotracking, biotracking
Nature of data
Qualitative data, quantitatively
Moderately accurate
Review session and assessment
Goal
Recognize the service and map it according to emotional reactions to the service.
Markers
Level of Ability — Can they perform the given task?
Level of Confidence — Are they confident or confused?
Level of Sentiment — How do they feel about the service?
Peak
Moment of Truth
Insight, Pride, Connection, Elevation
Optimal number of subjects
Around six, as established by usability research.
Duration of tracking
From 1h (a visit to a shop) up to a week (placing an ecommerce order, waiting for delivery, unboxing).
Resources
A week planning and recruiting
A week collecting data
3-5 days collating and reporting data
Suggested software / hardware
Mood Meter App
GoPro camera
Fitbit (or other biotracker)
Podometer app (if not using Fitbit)
GSR sensor & software
4/5
BIOFEEDBACK
DIGITAL SERVICES
If you are working with digital services you are able to collect biofeedback from users. It doesn't matter if you have a live service or an interactive prototype mockup.
Don't get impressed with the biofeedback term. You may go to eye tracking and other gimmicks, but you don't have to.
To me, the most illuminating thing you can get is a good heatmap from HotJar and proper recorded sessions — also available, for free, in HotJar. Other tools are Google Analytics (or any Business Intelligence suite). If you want to be really fancy, go for Galvanic Skin Reaction devices.
Heatmaps
What percentage of people are scrolling your pages till the end? Can you make a business out of the few ones who get to your call to action?
Mouse cursors
Also a heatmap, what is the mouse cursor behavior? This sounds funny, but it's an incredible indicator for what people pay attention to. Yes, people do read using their mouse. So you definitely don't need eye tracking gadgets (they will mostly track the eyeballs of a few users in a highly controlled an awkward environment, as opposed to free tracking of the mouse).
Screen recordings
You may spend hours watching the behavior of users online, like a legit scientist would do with their subjects. There is just so much that can be observed. Do users seem confused? Where do they get detained? Where are they spending time? At which point do they abandon the website?
Analytics
Analytics are not entirely user-friendly tools, but if you have a knack for it, dive in. If not, ask your analyst to give you some insights on how the site is used lately, the profile of users (location, gender, age), the pages they most visit, the time they spend on the site, the bounce rate (that is, how many people arrive at the site and leave immediately). Those won't explain everything, but will definitely be very valuable hints for what you are looking after — that is, more confidence in making decisions, raising hypotheses and challenging assumptions.
Galvanic Skin Reaction (GSR)
Once in a while I hear about this, and I go along, because they are fun tests. Galvanic Skin Reaction gadgets are attached to the user's hands and measure the electric current on the hands. If the user gets an increase in palms sweat, the current increases. It should be an indication of arousal, for the good (excitement) or bad (confusion).
GSR sounds all very illuminating, but the truth is simpler: why do you need to measure these somewhat imprecise nervousness levels if you can, quite simply, just measure if the design you are offering sells more than the previous? So forget GSR and focus on A/B tests.
HotJar of a Wikipedia page Heated areas show clicks or mouse movement, indicating clearly the areas and functions of most interest.
EXPLORATION & FEEDBACK IN LIVE DIGITAL SERVICES (DATA TRACKING)
Methodology
Tracking user sessions
Observation
Nature of data
Quantitative data
Goal
Observe users interacting with the service on their own.
Markers
Level of Ability — Can they perform the given task?
Level of Confidence — Are they confident or confused?
Minimum number of subjects
50-100, minimum
Duration of tracking
A few minutes, or as long as a typical session takes.
Resources
A week planning and recruiting
A week conducting data
3-5 days collating and reporting data
Suggested software
HotJar
SEM Rush
Google Trends
Google Analytics
PRO TIP
CONTROLLED ENVIRONMENT OR OUT IN THE OPEN?
As you may see from the takeaways, what you get varies from each method. Having a low number of respondents tracked quantitatively can give you "qualitative data, quantitatively", which may inspire you for great service design along the journey. It's greatly precise method. You measure closely, but you may be measuring irrelevant things if they don't ultimately represent many customers.
Quantitative studies, however, validate enormously if your design is working. It's a highly accurate method — you have less precision, but accurately indicates what is relevant.
Try to combine both according to your budget, urgency and need.
THE REVIEW
It's time to collate the data we gathered and turn it into actionable insights.
REVIEW WITH AGENTS
It's time to collate the data we gathered and turn it into actionable insights.
If your research had recruited agents (either in physical services or digital prototypes), you will discuss the following markers:
Markers
You will ask, from 1 to 10, how they felt about three main markers: Ability, Confidence, Sentiment.
Ability
Could they perform the task?
Confidence
Were they confidence on how to perform the task, or confused?
Sentiment
How did they feel about the service in the end?
Qualitative Markers
These are not quantified, but described by users.
Moment of Truth
When did they feel they actually "got what they were looking for"?
Peak
What would they say was the peak of the experience?
Special moments
Did they encounter any moment of insight, pride, connection or elevation?
Emotion variations
Along the journey, how moods change. A quadrant of emotions may help:
Good high energy
Excited, Cheerful
Good low energy
Satisfied, calm
Bad high energy
Angry, Nervous
Bad low energy
Bored, frustrated
REVIEW WITH TOOLS
It's time to collate the data we gathered and turn it into actionable insights.
If your research was based on live services tracked with tools like HotJar and Analytics, you will investigate the following markers, and assess them:
Markers
You will estimate from 1 to 10 how they felt about three main markers: Ability and Confidence, and identify these markers in the main screens of the service (homepage, product page, contact form, for example).
Ability
Could users perform the task?
Confidence
Did users seem confused and wandering at points, or confident and to the point?
Building the Journey Map
You have now a lot of documented insights on how customers experience your service. You need to distribute those insights over a map, chronologically, into different layers.
The big differences between this methodology and other methodologies:
-
Insights come from the real world, through the experience of customers, not from the designer's point of view.
-
We will break the journey into moments, not into objective touchpoints.