As founders or operators, we’re often faced with challenging questions:

• What problem should I focus on?
• What features should I build first? 
• What go-to-market motion is best with my target customer?

Fortunately, there’s a great approach to informing answers to these types of questions - talk to customers! 

As a startup studio, we’re always talking to customers. We’ve run our voice of customer ("VoC") process - targeting customers, running interviews, and synthesizing findings - over 400 times, resulting in 40 venture-backed spin-outs over the last nine years. We’ll be honest - getting the "VoC engine" going takes work, but it is a highly rewarded effort. It is what propels good ideas forward and kills bad ones quickly, mitigates the “hammer looking for a nail” problem, and encourages a tight fit between a concept’s ideal customer profile ("ICP"), value proposition, key features, and business model (collectively, "CVFB" - see related blog post explaining this process). At PSL specifically, it drives our decisioning to advance from our “validation” phase, into our “creation” phase, at which point we begin to invest significantly more resources. Note: At PSL, we run a five-phase company creation process: Ideation, Validation, Creation, Spin-Out, Scale Up.

What follows are some best practices that we use for targeting, running and synthesizing VoC interviews.

Targeting VoCs

Targeting VoCs ideally starts with a hypothesis for the ICP. Who specifically does the business seek to "super-serve"? We find that getting sufficiently narrow in this exercise leads to better outcomes. This is because with an ICP-focused effort, you can more quickly build conviction around that ICP. For example, if we’re thinking about an AI-powered compliance solution, we might have a hunch that financial services companies would be strong candidates for early adopter customers. As such, we might initially narrow our search to speak only with the economic buyers at banks, fintechs, and investment managers. From there, we can quickly get enough signal to determine if we correctly identified our ICP, need to narrow further (e.g., double-click on banks because they showed the most interest within financial services), move on to a new vertical ICP (e.g., healthcare), or kill the idea entirely. Caveat: Sometimes, you may not have a strong hypothesis for the ICP, or even if you do, that ICP may be really difficult to get in touch with. In those instances, having “generalist” conversations with those who are knowledgeable about or well-connected within a given market, regardless of ICP potential, may make the most sense!

Once we’ve thought through our targeting, our next goal is to find those ICPs and reach out to them. Personal relationships and warm leads from our network have the highest connect rates, but cold outreach can be worthwhile and is often necessary. On either front, LinkedIn Sales Navigator or Apollo.io are helpful tools to aid in this process. We’ve also engaged services such as Tarka.ai to provide cost-effective leverage in cold outreach. In addition to these and several other external GTM tools, we’ve built several of our own. Picco is a PSL spinout that identifies leads, scores them, and generates personalized messaging for easy outreach. PSL Spark is an internal tool, which among other things, significantly improves our ability to find the right people in our collective network. It works by bringing together PSL's LinkedIn connections (70K+ people!), enriching the profiles with publicly available information, and using LLM-powered search functionality. With this tool, we can easily search for people who match our ICP and identify who internally is the source of the connection, making it significantly easier and faster to get in front of prospective customers.

For consumer ICPs, Prolific is another tool that we find helpful. For a few hundred dollars, we can run a quick, targeted survey of several hundred people. We always add a question to the end of our survey to identify those who’d be open to a follow-up conversation.  

Often in parallel with finding our ICPs, we’re also crafting the right messaging in preparation for outreach. This is a critical piece of the process to build a strong VOC interview pipeline and often requires an iterative approach. As alluded to earlier, AI helps with this too, but typically still requires human intervention. Messaging at the earliest stages is often focused on better understanding the problem space as opposed to selling a product. We find that in any case, succinct and personalized messaging works best. It’s even better when we can offer some form of immediate value to our potential interviewee (e.g., "I’d be happy to share our perspective on X", or "I’d happily add you to the top of our beta waitlist"). 

Running VoCs

With our VoC interviews starting to hit the calendar (we use cal.com to help with this), it’s important that we have a plan for how to best facilitate them. As you never know how an interview will go until you’re in it, we find that maintaining a healthy balance of deliberateness and flexibility works best. One way to help on the deliberateness front is to create an interview guide. 

We like to be hypothesis led when creating our interview guides and we think about this activity in terms of "sprints". For each sprint (which tends to be weekly), we develop three or four hypotheses that we want to test in our calls (typically 30 minutes long). For each hypothesis, we develop a short set of questions that if answered in their entirety will give enough signal to prove or disprove our hypothesis. For example, we might have something that looks like the following:

Hypothesis: Not having a “single source of truth” significantly contributes to the problem of software developers finding information at work.

• Talk to me about how information is memorialized, stored, and shared at your company. Where does it live? How do you access it?
• Do you experience issues at work related to information redundancy?  
• How do you or people at your organization make sure information is up to date? Is it easy for you to know if the information you’re looking at is stale? 
• Do you ever run into issues with version control? How often? Please describe.
• To what extent do the issues we just discussed impact your ability to search for and find the right information? 
• What else, if anything, negatively impacts your ability to find information at work more than what we’ve already discussed?

Often with each new sprint, our guide will change to address new hypotheses and questions. Depending on the importance of what it is we’re testing, having five to 10 or more VoC responses to our questions is necessary. As such, it can help to stick to the guide, but it is also reasonable to deviate from it if you feel the conversation going in a different, more valuable direction.

One line of questioning we like to get at early in our process relates to Willingness to Pay (”WTP”). This is due to its critical nature within our idea vetting process - customers need to want a product AND be willing to pay for it in order for us to have high conviction. A great way to get at this answer is to have our interviewee assign relative value to the features we’re thinking about building. For example, if we’re interviewing software developers, we might ask something like “Relative to the application you currently get the most value from at work, where would you place NewCo on the same value scale?”. This helps everyone involved better understand the scale. Later, we can determine the cost of the example applications ourselves to help triangulate a dollar value WTP for NewCo. This relative value method also helps mitigate potential issues related to what we call the payee dissociation problem, whereby an interviewee may be a user but not a buyer of a given product.

As we host our VOC interviews, note taking is an important aspect that we’ve almost entirely farmed out to AI. Whenever possible, we like to conduct interviews over Zoom and use Read.ai* to transcribe and summarize the calls. We find this incredibly valuable as there tends to be new learnings each time that we go back to the recording. We’re also able to share summarized meeting notes immediately with other stakeholders to eliminate time lag in our process and later leverage the full transcripts to extract customer quotes for our pitch decks. 

Lastly, any time we speak with someone, we always ask if they know someone else they can introduce us to. The success of this exercise is telling in and of itself (i.e., more introductions = good signal), and perhaps the best way to get the VoC machine going!

Synthesizing VoCs

As we alluded to at the onset of this post, the primary goal of our VoC efforts is to extract sufficient signal to make better informed decisions early in our concept validation process. It’s important to recognize that in this process, we accept that we’re trading off precision for greater speed. As such, when we look back on our notes and synthesize our findings, we should keep that point in mind. 

A simple framework that we use to assess our VoC process is “Volume, Velocity, and Value”, referring to the number of interviewees that we’re able to attract, the speed at which those interviewees engage with us, and their ultimate willingness to pay for what we’re proposing. 

For example, very early in the process we start to get signal around how easy or difficult it is to reach our ICP and how interested they are in speaking with us. After a few weeks of trying different targeting and outreach tactics, if we can’t find our ICP, or if we aren’t able to get enough of them to speak with us, that is the first sign that we likely need to adjust course. How will we ever sell to that ICP if it’s so difficult to reach them in the first place?  

When we are able to sufficiently find and engage interviewees, we take note of the speed at which we’re able to do so. When interviewees (re: potential customers) are prioritizing conversations with us and making them happen quickly, that is a good sign. 

When we analyze the actual responses to our interview questions, we often look at responses to questions as percentages. Did we have a high percentage of VoCs express WTP? If 14 out of 15 said they would pay $X or more a month, that’s probably enough to have conviction at the $X level. What if it’s 14 out of 30? That could be enough too. Or not. Context matters a lot as we seek to draw conclusions. This is clearly not an empirical approach, but it can be helpful to move quickly in the earlier stages of concept validation.

In practice for us, VoC synthesis takes place at least every week. Sometimes it’s obvious that the business we’re testing is a “kill” (e.g., when we can’t find the ICP, there’s no WTP, we learn about many competitors with significant advantages), but often it requires at least a few weeks of different lines of questioning with variations of ICPs to get there. A strong signal that we have validated a business to move on to our “creation” phase is when our ICP, Value Proposition, Feature Set, and Business Model are in sync with one another and are remaining stable after 20+ VoC conversations. 

Conclusion

The VoC process - from targeting to hosting to synthesizing - is the backbone of our startup exploration process. We see it as the necessary mechanism that aligns ideas with their ICP, value proposition, key features, and business model. While the process requires time, commitment, and adaptability, it is a valuable and rewarding practice. It gives us the confidence to kill or validate a business idea based on real customer input, enabling us to avoid the "hammer looking for a nail" situation. As we continue to venture into new territories and test out new business hypotheses, VoC interviews remain our compass, guiding us towards the most promising opportunities and away from potential distractions. 

*Read.ai is a PSL Ventures portfolio company

‍

About Us

Pioneer Square Labs (PSL) is a Seattle-based startup studio and venture capital fund. We partner with exceptional founders to build the next generation of world-changing companies, combining innovative ideas, expert guidance, and investment capital. PSL operates through two primary arms: PSL Studio, which focuses on creating new startups from scratch, and PSL Ventures, which invests in early-stage companies. Our mission is to drive innovation and growth by providing the necessary resources and support to turn big ideas into successful, impactful businesses. If you have a groundbreaking vision, connect with us hello@psl.com, and let’s build something extraordinary.