We’re living in the age of the customer, where individual software users drive product decisions and market tactics. The trouble is, many software companies don’t know their customers—not really.
A huge reason for this is companies aren’t talking with their customers. A Hubspot analysis found 42% of companies don’t survey or ask for feedback. And 81% of companies don’t have a formal customer advocacy program. Even among companies who do talk with customers, data from User Interview suggests over 60% of stakeholders don’t know how to access customer research findings.
With findings like these, it’s no surprise teams struggle to put the customer first.
“Trying to stuff a product down the throat of an unsuspecting bystander is a good way to build the wrong thing.”
— Drew Houston, founder and CEO of Dropbox
Customer-driven is easy to say and hard to do. In truth, most teams run on hunches, directives from leaderships, and a good deal of Starbucks—not the customer. These teams reorganize every quarter and are tasked with new features that add theoretical value.
But data indicates these hunch-based features aren’t as valuable as teams think. When ProftWell asked 2,500 product leaders to assess the last 5,000 features they’d built, leaders put most features in the high-value and high-willingness to pay bucket. Yet when ProfitWell asked 1.2 million customers to assess those same features, customers put them in the opposite bucket—low value and low-willingness-to-pay. What ProfitWell considers “trash land.”
“Overall, if you want to deliver an AMAZING customer experience, the SINGLE MOST IMPORTANT thing you can do is LEARN more about your customers so you can custom tailor that experience to them. It's not magic. It's not science. It is simply building a tighter relationship with your customer.”
Eric Carlson - Founder, 10XFactory
To be fair, many teams want to run on customer insights, not hunches. They’ve seen the success of product-led companies like Shopify, Twilio, and Atlassian. (Product-led companies like these have over 2x enterprise value, over 1.5x revenue, and over 9% higher revenue growth than other SaaS players.) SaaS teams know customer understanding is the key to growth. It’s figuring out how to use the key that’s so difficult.
When it comes to customer insights, there are two types of useful data: quantitative and qualitative.
Quantitative data provides important context on what customers are doing and how much. Quantitative data is fixed and measurable. It provides valuable information in contexts like A/B tests and market research.
Quantitative data is incredibly useful, but it has at least one major shortcoming: For all its statistical significance, it cannot help you understand why customers are taking certain actions. For that insight, you need qualitative data. Qualitative data illuminates the why behind the what.
This why information is incredibly important. It allows teams to move from reactive to proactive—from what customers are doing this week to what they’ll adopt in the future. Amy O’Callaghan is a product manager at Snagajob who has practiced customer discovery for over a year. She’s experienced the reactive-to-proactive shift and explains, “I feel confident pushing for things that we’ve prioritized because we know they will bring value—we aren’t taking nearly as many risks with development time as we used to, and the dev team appreciates that.”
To many product managers (not to mention marketers), that level of confidence sounds magical. But obtaining this confidence isn’t magic or even a well-hidden secret. In fact, many teams could start gathering information from their customers today using the simplest of tools: the survey.
Surveys aren’t a revelational idea. The earliest ones date back to at least the Roman era, and they’ve been widely used in the US since the 1930s. You probably have one sitting in your inbox right now.
For many teams wanting to collect more qualitative data from existing customers, they’re a great starting point. Chances are, your team already uses a tool (such as Hubspot, Typeform, or Intercom) with survey capabilities built-in. Surveys are also lower cost and lower effort than more robust customer research options like interviews and focus groups.
There are many types of surveys, but in this course, we want to walk you through creating one specific type: the qualitative survey. This type of survey targets a specific group of people with specific feedback questions. Oftentimes, those questions are open-ended. The goal of a qualitative survey is to generate insights that help solve business-critical problems.
Other qualitative survey benefits include:
Next up we'll walk you through how to craft a qualitative survey so you can get these benefits. By the end of this chapter, you’ll have a firm understanding of who to talk to, what questions you should ask them, and which research method you should use.
One enduring myth about surveys is this: they’re a set of semi-random questions sent out to a very random group of people. A good qualitative survey is neither of these things.
“A problem: even the best personas tend to be descriptive, but not predictive.”
The idea behind personas is a good one. They’re supposed to humanize software users and equip teams to make better decisions. However, the problem with most personas is they’re largely made up. It’s easy to pull a variety of data from sources like Google Analytics to make a semi-educated guess most of your customers are “Jill,” a late-20s, iPhone-wielding, trend-setting millennial who lives in the midwest.
While these visualizations may look great, they bring little predictive power to the table. Will Jill adopt the latest feature? Maybe—she’s trendy, after all—but we can’t really know. Because while personas may tell you what customers look like, they don’t tell you why customers buy or what is going on in their lives when they do. Personas fail to tell you whether Jill even needs that feature to begin with.
To truly understand customers’ behavior, you need to look at them from another angle.
When Clayton Christensen popularized the Jobs to be Done (JTBD) framework, he gave teams a powerful new way to view their customers.
While there’s plenty of nuance to the JTBD framework (we cover that in another resource), the gist of it is this: Customers don’t buy your product. They hire it to do a specific job.
A job, in this context, is a specific type of progress your customer wants to make. To find that job, you have to understand what your customers’ lives look like. You also need to know what pains and frustrations they experience, alternative solutions they considered, and so on. This is the kind of information that helps teams make smart moves around marketing, acquisition, retention, and churn.
One way Typeform leverages customer understanding is through their quarterly “customer voice” report. The customer experience (CX) team generates this report by pulling data from support tickets, churn surveys, sales calls, and other touchpoints. In one instance, the CX team figured out the majority of Typeform’s churn is due to one thing: Customers don’t know what to do after they create their first form. This insight helped the Typeform team combat churn with the “What’s your next Typeform?” campaign. This campaign appears to customers who have completed a form, and it delivers inspirational content around creating more forms.
“Through our churn survey, we found that a lot of our churn isn’t actually due to customers being unhappy, but rather from people successfully completing a project and not knowing what to do next. As a result, beyond the typical ‘feature-based’ content one would expect in a help center, our education team also creates content that is ‘job-to-be-done-based’ in order to inspire customers to do more with Typeform than they had initially intended.”
– David Apple, VP of Customer Success and Sales at Typeform (former)
Looking at the customer journey as a whole is important. But in this course, we want to hone in on what you can learn from the journey stages that intersect your product.
From website visitors to recent cancellations, here are the lifecycle points that are particularly useful for building a JTBD understanding.
This is a customer who’s browsing your website or landing page, but hasn’t committed to your product. These customers can help you understand where your positioning does (and doesn’t) resonate and what information about your product isn’t clear.
This lifecycle point is especially useful for marketers. Insights from website visitors can help marketers:
This is a customer who recently signed up for your product or service. They can help you understand what the customer is:
Maggie Crowley, Director of Product Management at Drift, asks new customers questions like, “What are your big goals for this year?” and “What kind of conversion rates are you trying to hit?” Because if you know what your customers are trying to achieve, you’re one step closer to making them superheroes with your product.
This is a customer who recently upgraded their relationship with you. These customers can help you answer questions like why did they upgrade now? And, what changed and pushed them to upgrade? What you want to identify here is a buying trigger. Triggers tell you how to move customers through the upgrade process more efficiently.
New Feature Activation
This is a customer who recently adopted a new feature. These customers can help you answer feature-specific questions such as:
Failed Feature Adoption
This is a customer who started using a newly released feature but then abandoned it. These customers can help you identify what is difficult about a feature, where it fails to meet expectations, and what pieces of the puzzle you may still be missing.
Feature Milestone Usage
This is a customer who has reached a specific “power user” level. They’ve hit some metric that marks them exceptionally successful with your product. This metric is usually quantitative, such as active days per month, number of actions completed, shares on social media, or total hours in the software.
Because power users often have very different usage from normal users, they can help you:
This is a customer who has recently left your product. They can help you understand:
Over time, surveying new cancellations will help you identify common churn patterns and prevent these, just like Typeform in the case study earlier in this lesson.
Some segments are very easy to identify. New signups, recent upgrades, and new cancellations should be apparent in every product or payment dashboard.
Other segments, such as new feature activation, failed feature adoption, or feature milestone usage will be easiest to see in third-party software. Tools like Mixpanel, Heap, and Amplitude all surface this information. Check your specific tool’s research page for additional guidance.
Once you know which segments you can survey, your next step is deciding which segment and what to ask.
“Ask them what they care about. Ask them how they are measured in their role...you should know how your customers are evaluated in their performance reviews. You should know what they care about, what they wake up thinking about. Start there.”
— Maggie Crowley, Director of Product Management at Drift
If you put garbage into your survey, you’re going to get garbage out of your survey. To gather insights instead of trash, be strategic. Start with a specific problem and choose the right people and questions to answer that problem.
“Your objective directly impacts every aspect of the research, from the scope of the study down to the questions you ask. Spend the time upfront to define what you want out of it.”
— Jesse Caesar
You can’t afford to spend time and resources gathering insights for the sake of insights. You need insights to solve specific problems.
To find those insights, start with a painful problem you want to solve and to what end. For example:
Stating your problem and why solving it matters does three important things:
If you manage people or read much psychology, you know humans aren’t totally honest. They also don’t know what they want and are terrible about predicting their future behavior. So, how can you get reliable qualitative data from customers?
You ask smart questions.
“Rather than asking: ‘Why did you buy our product?’ ask ‘What was happening in your life that led you to search for this solution?’ Instead of asking: ‘What's the one feature you love about [product],’ I ask: ‘If our company were to close tomorrow, what would be the one thing you’d miss the most?’ These types of surveys have helped me double and triple my clients.”
Talia Wolf - Founder and Chief Optimizer at GetUplift
This, of course, is easier said than done. Especially when you consider the many kinds of questions you can ask in surveys. Question types include open-ended, close-ended, nominal, and rating scales. Each of these is useful in their own way.
To save you reading time and to make survey creation easier, we’ve put together a collection of survey playbooks to help you get started. A playbook starts with a type of problem you’d like to solve (e.g. churn) and ends with the exact questions you can add to your survey.
Check out your options in the Playbooks gallery below. You can filter by problem you’re solving, lifecycle stage of the customer, or area of the company. We’ve covered who to survey and what questions to ask them. Now, one of your last decisions is where to launch your survey.
Choosing where you will survey customers is one of your last major hurdles. Common distribution options for SaaS companies include:
There are a few things you need to consider when you weigh these channels:
What do you have access to?
If the customer hasn’t provided their phone number, for example, that avenue isn’t an option. Or if you can’t easily segment your email list, a targeted email survey may be more headache than it’s worth.
How contextual does the survey need to be?
If you’re hoping to understand why a customer is adopting a particular feature, surveying them while they’re interacting with the feature is ideal. Likewise, if you want to understand why a customer upgraded, asking them on a thank you page right after they upgrade, or in-app after they adjust their subscription plan, is ideal.
Which customer segment are you targeting?
If you’re surveying churned customers, they’re no longer in the app. This rules out in-app surveys. If you’re surveying website visitors, your only option is on the website.
What do your customers prefer?
Some customer types may have particularly full inboxes, prefer text messages, or have an aversion to text messaging. As always, meet your customers on their terms, and keep their entire context in mind.
How much does response rate matter?
In-app surveys are easier for customers to complete and have higher response rates than email surveys. However, email surveys usually gather richer qualitative data because the respondent intentionally sets aside time to respond.
One other thing teams should note is the traditional in-app pop-ups and email surveys aren’t your only option. Particularly with the in-app approach, there are many creative ways to prompt customers.
Hiten Shah, for example, came up with a particularly clever way to answer new feature questions for his most recent business, FYI. “If we’re going to add a new feature,” Shah says, “we test it by adding a button for it first before we’ve built out any of the functionality. When people click the button, we ask them questions about their motivations.” This not only tells Shah how many people are interested in a feature (a quantitative measure), it tells him why people want it (a qualitative measure).
In short, don’t be afraid to break the mold when you engage with your customers!
“You need to come into qualitative market research with an absolutely open mind. If you’re fixed on a certain outcome, you’ll selectively read the output in your favor. The point of research is to be humbled by it — and inspired to do better.”
– Jesse Caesar
After you get your survey up and running, you’ll need a way to gather responses, organize them, and look for patterns. A few tips here:
LearnWhy analysis and tagging make it easy to organize feedback
Visuals are a fast and easy way to track patterns in your data