Start with the decision, not the data
A freelancer notices that fewer prospective clients are accepting proposals. A small team hears that its onboarding process feels confusing. In either case, collecting more information may help—but only if the team knows what it is trying to decide. Before opening a spreadsheet or sending a survey, write a specific question: “Which part of onboarding should we improve first?” or “Are prospects dropping off after seeing the price or before receiving a proposal?”
Quantitative data consists of values you can count or measure, such as the number of inquiries, proposal acceptance rates, time spent on a task, or ratings on a fixed scale. Qualitative data captures descriptions, experiences, and context, such as customer comments, interview answers, support messages, and observations. Neither type is automatically better. The useful choice depends on whether you need to measure a pattern, understand an experience, or do both.
Use quantitative data to see patterns and changes
Numbers are useful when you need to compare periods, services, or steps in a process. A design studio might track how many inquiries become consultations, how many consultations lead to proposals, and how many proposals become projects. Those counts can reveal where to look more closely. Tracking the same measures consistently can also help the studio see whether a process change appears to coincide with a change in outcomes.
Define each measure before collecting it. Does an “inquiry” include a casual social media message? When does a proposal count as accepted? Consistent definitions make comparisons more meaningful. For a small business, a short list of measures tied to real decisions is often more manageable than a dashboard full of figures no one uses.
Be careful with small samples. If two out of four prospects accept a proposal one month and three out of four accept the next, the percentage rises sharply, but the difference is only one client. Record the underlying counts alongside percentages, and avoid treating a brief fluctuation as proof that a change worked. Numbers may identify a question worth exploring without explaining its cause.
Use qualitative data to understand reasons and needs
Qualitative information is especially helpful when you do not yet know which answer choices to offer or why people behave as they do. An open-ended question such as “What was difficult about getting started?” may uncover unclear instructions, uncertainty about timelines, or terminology customers do not recognize. Watching someone try to complete a task can reveal friction they might not think to mention.
Keep the conversation focused but leave room for unexpected answers. Ask about a specific experience rather than requesting a general opinion of your business. For example, “What did you expect to happen after you submitted the form?” is more actionable than “Did you like the website?” Avoid questions that suggest the answer you hope to hear.
A few vivid comments can guide improvements, but they do not establish how common a problem is. A customer who offers detailed feedback may have needs that differ from those of quieter customers. Look for recurring themes, note disagreements, and keep individual examples separate from claims about your whole customer base.
Combine both types for a clearer picture
Suppose a consultant finds that many prospects book an introductory call but fewer request a proposal afterward. The booking and follow-up counts show where the process changes. Notes from calls or brief follow-up questions might suggest possible reasons: the service scope is unclear, the timeline does not fit, or the prospect was only gathering information. The comments help the consultant decide what to adjust; later counts can help check whether the adjusted process is performing differently.
The sequence can also work in reverse. Customer conversations might repeatedly mention slow replies, prompting a team to measure response times and compare them with project outcomes. Start with whichever kind of evidence matches what you do not know. If you know something is happening but not why, seek context. If you have heard a possible problem but do not know its scale, consider a relevant count or measure.
Do not force a precise numerical conclusion from a handful of interviews, or assume a numerical change has only one explanation. Changes in client mix, workload, or how data was recorded could also matter. Treat the two types of data as complementary checks rather than competing verdicts.
Keep collection proportionate and act on what you learn
For a small team, a simple routine is often enough: choose one decision, select a few relevant measures, gather a manageable set of customer observations, and review them together. Write down what the evidence suggests, what remains uncertain, and what small change you will test. If you revise an onboarding email, for instance, you might monitor related support questions and ask new clients whether the next step was clear.
Collect only information you have a reason to use. Be respectful when requesting feedback, avoid sharing identifiable customer comments unnecessarily, and store notes where the appropriate team members can access them. The goal is not to build the largest dataset. It is to make a better-informed decision, see what happens next, and revise your understanding when the evidence calls for it.
