What tools do you use to survey your customers and how do you extract qualitative vs quantitative insights from it?

This probably won’t come as a huge surprise, but our most used tool for surveys at Airtable is...Airtable. A lot of our customer surveys are done using Airtable Forms, which help us store that feedback directly in Airtable to report on and sync to other important bases, so everyone in the org has access to valuable customer feedback. I won’t make this just an Airtable product pitch, so I'll share that I’ve also used SurveyMonkey and GetFeedback and both have worked well for my survey needs.

Extracting qualitative and quantitative insights is really about creating a good survey. Often, we want to hear as much data and feedback from customers as possible, and there’s a temptation to structure your survey with a lot of open ended, free response questions to get all those good nuggets of intel. The thing you have to be careful about is that it’s extremely difficult to synthesize that qualitative info when you’re getting hundreds or thousands of responses to your survey. Here are a few ways I've gone about creating an impactful survey: 

• Marrying quantitative and qualitative questions: Give respondents a multiple choice or true or false question, but then give them the option to add more detail or information if they’d like. That way, you get the clean quantitative data, but also leave room for any qualitative insights a respondent feels like providing. 
• Keep qualitative feedback snappy: Reading paragraphs of text from survey respondents makes it hard to pull out key learnings. I like to keep qualitative insights in a survey short and sweet. For example, if I’m trying to understand how a customer explains/positions our product, I’d ask the question “How would you explain Airtable to a friend or colleague who doesn’t use the product in one sentence?” rather than asking “What do you think our product is for and what value do you get out of it?”. One will give you meaningful, short answers and the other might get you long-winded paragraphs. 
• Make sure you’re not skewing responses in your question wording: This is a big one that's taken me a long time to learn. Many of the surveys I sent early in my career were poorly worded and incredibly skewed. A survey should be objective, not guide a user to the “right” answer. Make sure your questions are clearly worded, concise, and free of any bias or language that might influence the respondent’s answer. PMMS are good at many things, but we're not experts in survey design so I'd highly recommend leaning on a partner team, like user experience research, to check your work and give suggestions here. 

I like Typeform (super simple for users, works for my use case right now) for surveys where I have pointed questions on a research topic, and also work with other teams to run evergreen G2 campaigns and NPS surveys which I think give you an ongoing thread of general feedback.

Extracting value depends a lot on what you're working with. For example, a 1-question free text survey that 10 users complete can help you get a TON of qualitative insights but not as much quantitative you can pull from there due to the small sample size.

I like to think about that up front! If you have a small audience and are trying to understand broad topics, focus on the qualitative information you can extract. If you have a large audience and a very specific topic (e.g. pick which messaging resonates the most with you), focus on what the group in aggregate is telling you.

In terms of extracting value, go into it curious and with no expectations. Letting the survey data tell you its own story will help you actually pull the right insights instead of trying to force it into a story that you were hoping works. That works against what user research is meant to do—it's meant to get you out of your head and into the users!

There are a few ways to collect customer insights. First, you can look at your own data to understand how your customers are using the product and what they are doing on your site. Tools like Google Analytics, MixPanel, Crazy Egg, and Segment are great for journey mapping customer behavior at scale.

Next, you can do qualitative and quantitative studies to understand why your customers are using your product and who they are - this is important for segmentation. You can run surveys in-house using tools like SurveyMonkey or Qualtrics or you can outsource studies to an agency. I recommend getting a few quotes if you are going that route. You can also quickly collect feedback via an in-product survey using Intercom or a short poll embedded in your email using something like Movable Ink.

Your customer support team is a treasure trove of insights. Be sure to connect with them often and set up an easy automated system for phone/chat agents to log common issues, questions, and feature requests.

Finally, nothing beats just getting a customer on the phone - send a quick email invite and ask for a 15-min call for customers to share their experience with you. These conversations are usually very enlightening.

At Atlassian, we use many methods for understanding customers both qualitatively and quantitatively. 

The most standardized, larger-scale tool we use across all of our cloud products is our Happiness Tracking Survey known as HaTS (developed by Google). Our research teams sends out weekly emails to employees who subscribe that give the overall customer satisfaction score and short clips of customer feedback such as what customers find frustrating about our products or what they like best. This is a helpful way to keep customer feedback top of mind. 

For more in-depth research on a particular audience, product, or feature, we use Qualtrics to send wider surveys to specific audiences. 

On my new products team at Atlassian, we use the Product-Market-Fit score which asks users 'how would you feel if you could no longer use the product?". We usually aim for early products to at least reach 40% of their earlier adopters to say they would be very disappointed. You can read more about the survey here. 

We also use a variety of analytics tools to measure our funnel and in-app engagement such as Segment, Amplitude, Redash, Tableau, etc. 

On the qualitative side, we run customer interviews, user testing sessions, focus groups, customer advisory boards, etc. 

I mentioned this in a previous answer, but I believe the best way to structure research is to develop hypotheses you are trying to prove or disprove before crafting your survey or research questions. You want to be clear about what you're trying to understand. Typically, you can start with customer interviews, a qualitative approach, develop your "hunch" and hypothesis, and then use wider surveys to validate the hypothesis.

It’s no surprise since at Intercom we offer an in-product surveys tool, so I’m naturally a big fan of in-product surveys. Even before Intercom, I had always seen a 3-4x response rate to in-product surveys than email or any other channel. I think surveys in-product are generally more valuable for quantitative insight and tactics like interviews, a customer advisory board and focus groups are more helpful for those rich emotional or granular feedback. How does your customer feel about your product and which problems are most urgent to them are best revealed in a live conversation.

The way to extract insights that are meaningful is to really make sure you write your survey questions or qual discussion guide questions with as little bias as possible. Good research is designed to minimize bias. I always have a friend or researcher review my discussion guides if I’m writing them myself to make sure I’m not leading the participant and stripping out bias.

Here’s a few watch outs that are common:

• Avoid asking yes/no questions - you want to actively encourage discussion by posing open-ended questions and understand the “why”
• Confirmation bias - This is a tendency to find what you’re looking for and ignore any contradictory information that doesn’t support your hypothesis (the cherry picking I mentioned earlier)
• Selection bias - it’s easy to get stuck in an echo chamber of your existing customers who love you. Be really thoughtful about who the target audience is for research and if you’re interviewing existing customers, make sure they’re not actively engaged in other betas or giving input on other product areas so as to not overwhelm them and remember you’re not building a product for a single customer.

Give participants permission to be critical - Encourage participants to be comfortable sharing a critical perspective. If you’re showing them a new product concept, let them know it’s okay if they say their willingness to pay is zero or if they’d have no use for the product at all - in fact this is helpful information for you!

Great question — it's all too easy for self-serve marketers to get disconnected from customers, especially if you're used to gathering this feedback through customer-facing orgs.

As a leader at SurveyMonkey, I'm going to be a bit biased here, but I'll provide some general guidance on choosing a feedback solution, and you can guess what product I'm recommending:

• Research teams are great — we have an awesome one — but lots of orgs don't have dedicated product/market research functions, or those that do have extensive backlogs. Sometimes you need to be able to get insights quickly or set something up yourself, so choose a survey solution mere mortals can use. It's no good to you if you need a PhD to work it.

• When you're writing a survey, think about the types of insights you're looking for: while it's important not to introduce bias in your questions, I always think about what my ideal results look like and how I plan to leverage them internally or externally, and that helps quite a bit with determining what questions to ask.

• Survey templates, pre-written questions, and even generative AI can provide really solid starting points that will result in actionable feedback.

• Make sure your feedback solution supports industry-standard methodologies for getting the insights you want to uncover: If you want to know whether customers love you, make sure you can measure NPS. If you're looking to uncover customer and market preferences, MaxDiff can be a useful tool. You don't want feedback scattered across a dozen different platforms, so pick one that's proficient at everything you need.

• Surveys are a touchpoint for your brand and should reflect it when they're sent to customers, so pick one that can leverage your brand guidelines across the survey itself, email invitations, etc. Certain free tools really fall short here and are not cut out for customer-facing duty.

• Ask both quantitative and qualitative questions. The quant will tell you what is happening. Qual will tell you why it is happening.

• With a great feedback platform, you won't feel like you have to extract insights from your surveys — a great solution will give the insights to you so you can focus on what to do about them. Sentiment Analysis is a great example of technology helping here: in the old days you would have to manually sift through open-ended feedback and categorize it yourself before you could make sense of it. Now machine learning can do that for you and sort responses into Positive, Neutral, or Negative sentiment, so it's easy to see where you stand and dig into "why" with a couple of clicks. I'm incredibly excited about the future of AI when it comes to helping people uncover interesting segments and insights from survey responses.

I hope that's helpful. With a solid feedback platform and a plan to capture feedback on a regular cadence, a self-serve business leader can easily keep well informed when it comes to what customers are thinking and feeling.

Utilize product data for quantitative insight. For example, how many people clicked through the in-app product tour? 

Utilize surveys and interviews for quantitative insight. Extracting qualitative insight from interviews is a special skill to hone - you've got to think critically and analyze patterns in real time to make the most of it. Surveys are great for snapshots, but they do have a low response rate so come prepared with a reward! 

Big fan of Pendo for self-serve analytics. I've used Google Forms, SurveyMonkey, TypeForm, etc in the past for surveys. Customer interviews need to be recorded - you can use zoom recordings or Chorus/Gong as well.