Build Design Systems With Penpot Components
Penpot's new component system for building scalable design systems, emphasizing designer-developer collaboration.

Inside Intercom | Sinéad Cochrane
If you’re not talking to your customers and listening to their feedback, your product is destined to struggle.
But how you go about gathering that feedback is as important as the feedback itself. For product teams not in regular contact with their customers, surveys can seem like the right solution to this problem, but they’re often misunderstood and misused. Here at Intercom, we’ve found that sending highly targeted, open-ended questions is far more valuable and reveals more of what customers really think.
Just like Intercom’s designers use Intercom to come up with better design solutions, the research team uses it to target and talk to our customers. This allows us to segment our customers and their answers, so we can treat our open-ended questions as qualitative (reducing the number of answers needed) rather than quantitative (i.e a survey requiring hundreds of respondents).
Data collected from your customers is a valuable commodity.
However, if done wrong, you can end up with questions that don’t map to the information you need or answers from the wrong types of customers.
A highly targeted, open-ended research question is like a 60-second interview with your customer, where you only get to ask one question. Unlike multiple-choice survey questions, it allows you to collect an infinite number of possible answers and as much detail as a customer wants to provide, which ultimately means you’ll end up learning things you didn’t expect. Even if you’re not an Intercom customer, you can still send research messages like these, via any available channels where you can accurately target and talk to your customers.
In this post, I’ll outline the key questions to ask yourself before starting to write a research message to your customers.
The biggest mistake I see non-researchers do is jump straight into writing the message they’re going to send to customers. This is about as efficient as trying to decide where to go for dinner when you don’t know what type of food your girlfriend wants to eat that evening.
It’s essential to plan ahead in order to figure out what you’re really trying to answer. Ultimately, the most valuable and reliable insights will come from sending the right research message, at the right time, to the right person. But it takes planning to figure out what the “right” everything is.
Think about the types of answers you would like to get back, and how that would impact the decision you are trying to make. What new information will you need in order to make your decision? Turn this into your research objective, which will define what it is that you want to learn e.g. “Why do customers visit the settings page?”
It can be helpful to write these down as simple questions – such as Who? What? When? Where? Why? How? – or around two predetermined answers, e.g. “Do customers want feature X because of A or B?”. When you write down all of your objectives you might find you have more questions than you expected, and some might be better answered via different methods, or you might have to send multiple research messages. Don’t panic. Talk to your team, prioritize what you want and need to learn and plan out how many messages you’re going to send.
Having a clear idea of who you want to message ensures you aren’t collecting data from the wrong customers. It will also stop you from sending an irrelevant research message and annoying your customers.
When defining the audience for your message, talk through these questions with your team:
Make sure to check-in with other teams in your company (such as marketing or customer support), to see if there are any rules around messaging your customers. For example, Ruairí, our Product Education Manager, has outlined a set of outbound message guidelines that ensure we’re not impacting any of the onboarding message flows for our new customers.
Always craft your message carefully, and remember that you’re representing your entire company when you speak to your customers. Make sure to write a message that your customers would be happy to receive when they’re in the middle of using your product. Your message should be clear, concise and polite. Ideally, it should be broken into three parts:
Believe it or not, people don’t spend all day thinking about your product, so asking a question when they’re using your product is best. The memories of the experience will be fresh in their mind, and they’ll be more capable of forming an opinion, remembering why they work in a particular way, or recalling problems. However, if you don’t have the capability to do this, you can try to use memory prompts e.g. screenshots of which part of your product you’re referring to.
The right time to ask a research question is less about figuring out the optimal day or time, and more a combination of two things – the where and the when. Discussing questions such as “Where should they see the message?” and “What action should they have just taken?” will help you and your team figure this out.
Whilst asking a question immediately after someone has done something or visited somewhere in your product can help you capture more relevant insights, you also want to make sure to balance this with not interrupting them in the middle of doing something important.
Whilst it’s easy to assume that your customers won’t expect a reply, the simple act of thanking your customers for their answers will make them feel listened to. They’ve taken a moment out of their day to help you out, so acknowledge it. This also increases the chances they’ll be happy to provide feedback in the future.
If your usual method for reaching out to customers makes it difficult to respond, try to find ways of working around the barriers. If you’re using email, set up an email address for conducting research (e.g. [email protected] instead of [email protected]), so all of the customer answers are going to one place and can be managed by someone on your team. If you don’t have access to email and are using a survey tool to capture answers, make sure to include an email capture field with the message, “Please leave your email address if you would you like us to reach out to you directly to discuss your feedback.” Just make sure to reach out to any customers that do leave their email address.
Whatever your method of collecting answers, the speed at which you reply to your customers is very important. Promptly replying will give you opportunities to ask follow-up questions if something they have told you is really interesting, or if anything is unclear in their answer. Don’t shy away from a research message starting a conversation, as this experience alone is a great opportunity to build empathy with your customers.
Sending a message to your customers without documenting and sharing your findings is not research. As the answers come in, you’ll naturally start to develop a feeling for what they are collectively pointing to. However, without systematically exploring the answers for themes, you’re still going on intuition, with just a splash of customer feedback thrown in.
At this point, you should be referring back to the plan you had at the beginning, reviewing the answers against your research objectives and then grouping the answers by theme. When we wanted to understand “Why are some customers sending a high volume of test emails?” some of the themes we saw were: Visual previews of the email design, proofreading the content, gathering external feedback, and checking that the links work. Some of these themes were unexpected, as we hadn’t realised our email message composer wasn’t sufficient for some of these activities, which showed the value of us sending the message in the first place.
Data collected from your customers is a valuable commodity; it should be accessible and reusable later. Collating the answers somewhere like Google Sheets will allow you to easily group and theme the answers and share them with the rest of your team. As a rule-of-thumb, we collect 20 answers for each question, and then stop messaging, in order to review if any themes have emerged. This proves a fast and effective way of gathering customer insights.
Once you’ve established themes, it’s essential to document your research and, more importantly, don’t keep it to yourself. Documentation is important because it enables you or someone else to return to the research at a later date and understand what everything means without having to read through every response, thus reducing the risk of insights being forgotten.
Your summary of research should answer the questions you and your team had at the beginning and be used to help your team make decisions. This could be as simple as a Google Doc or as complex as a presentation or poster. The key is to make the research visible and enable your team to have a discussion around it. The research findings should also be shared outside of your team, to prevent customer insights and product knowledge from becoming siloed.
Once you get into a rhythm of always having a plan, you’ll find it gets easier and faster each time to conduct this type of research. Not only will it give you reliable, actionable insights within a short space of time, being in regular contact with your customers will bring you closer to them and help you understand their problems better.
We’ve put together a cheat sheet outlining a simple three-step process to help you send the right research message at the right time to the right person. It also includes some tips on how to write your message. Try it out next time you want to get feedback from your customers, and if you have any suggestions or questions I’d love to hear them in the comments.
The post 7 questions to ask before messaging your customers for research appeared first on Inside Intercom.
AI-driven updates, curated by humans and hand-edited for the Prototypr community