In real startups you work at speed. Features ship every two weeks. You cannot do multi-month focus group studies before every decision. The question is how to get actionable insights quickly while keeping up with the engineering team's need for direction.
Why focus groups and surveys often disappoint
Focus groups have a few consistent failure modes: a few vocal participants dominate and suppress more valuable dissenting views; the group dynamic creates artificial consensus; and the differences between individual participants — which are often the most interesting data — get averaged out.
Online surveys have a different problem: the "choose an answer" model bakes in the researcher's assumptions. You define the options; users choose from them. The interesting texture of what users actually think rarely fits cleanly into preset responses. And you can never ask the follow-up question that the answer makes you want to ask.
Neither method is useless — surveys are valuable for tracking attitudes at scale, and focus groups can give a useful top-of-mind view of brand perception. But for product direction and user insight, something else usually works better.
Two methods that work at startup pace
Extended one-on-one open-format interviews: Long-form conversations, ideally face-to-face, with open-ended questions that invite detailed answers. Questions like: "Describe your needs in your own words. How else would you solve this problem? What would you value in a solution? What do you like and dislike about our current approach?"
The goal is to learn the unexpected — to pan for the occasional golden insight that no survey could have anticipated. Take real notes (not summaries). Share them internally. Three of these interviews a week for a month gives you more customer knowledge than most people at your company have accumulated in years.
Narrow, goal-directed single-issue research: You have identified a specific problem — say, users are dropping out of the sign-up flow at step 3. Engineering has proposed several solutions. Which (if any) are the right ones?
Pull a list of relevant customers, send personal emails, ask for 15 minutes by phone or screen share. Ask general questions first to confirm you are talking to your actual audience. Then probe the specific issue: "What were you trying to accomplish when you got to this step? How did you expect it to work?"
This is not a focus group. It is a targeted investigation. One issue, five to ten participants, 15 minutes each. You can run this in a day.
Practical tips for running user sessions
A few practices that consistently improve the quality of user research:
- Invite a mix: current customers, potential customers, advocates, and a few people who do not know your product at all. Each group gives you different signal.
- Early morning works best. People are more energetic and less distracted.
- Use rough wireframes or high-level specs in the session. Do not wait for polished mockups — early sessions on rough materials generate richer feedback.
- Track time spent on each screen or section (normalizing for content complexity). This reveals where users are confused or interested — more reliably than asking them.
- Debrief with your co-facilitator immediately after each session. Do not wait — immediate synthesis is far more accurate than notes revisited a week later.
- Run at least two sessions before drawing conclusions. Averages from a single session are noise.
B2B buyer research: the separate problem
In B2B and enterprise products, users and buyers are different people. The person using your product every day is not the person writing the check. If you only talk to users, you are solving for one half of the equation.
The single best method: Sit in a room with buyers and ask targeted questions. Your goal is to understand their problems, their goals, and how they are solving those problems today — whether with a competitor's product, a spreadsheet, or a manual process.
The access problem: Buyers in enterprise contexts are senior and protected. Your attempts to meet them will initially look like sales pitches. They will not respond to gift card incentives. You have to travel to them (expensive). Sales will want to tag along (which invalidates the research — it turns into a sales call).
What actually works:
- Source a list of potential candidates via your CRM, LinkedIn, and network. Prioritize second- and third-degree connections — an introduction dramatically increases response rates.
- Contact the list in order: phone, then email, then social media. Expect few responses. That is normal.
- When you reach someone, lead with a concise statement of the problem you believe they have: "As a VP of Sales, I think you experience pain with X because of Y. We are solving that problem and want your help." If you are right about the problem, this gets attention. If you get repeated rejections at this stage, you may have identified the wrong problem.
- Keep the session focused on discovery, not selling. Understand how they solve the problem today. What does their workaround look like? How much time does it cost? Stay on track — the time goes fast.
Synthesizing what you learn
Once you have conducted research, the PM's job is synthesis — distilling a large, messy set of observations into clear, shareable product direction.
Two tools that consistently work:
2×2 matrix: Choose two dimensions that define the competitive landscape and map your product against competitors. Jobs used this at the iPhone launch to show exactly where every competitor's phone sat — and why none of them were addressing the quadrant Apple was entering. The insight is in choosing the right axes.
Feature gap analysis: A comparison table showing which value propositions your product and competitors deliver, with ratings for each. Surfaces where your product is strong and where gaps exist that you can exploit. Start with this when you are deciding what to build next, and revisit it quarterly as the competitive landscape shifts.
Both tools force you to take large amounts of research and compress them to a single visual.