Why post-sale surveys and not only non-customers?

Because information from customers who have already bought best reflects who actually buys from you: what motivated them, what they hesitated about, who decided. If you only ask those who do not buy, you may be refining the profile of someone who will not close. Non-customer surveys are useful for contrast, but prioritise customers and recent lost opportunities.

What does «over 60 % or 80 % correlation» mean?

That a trait, context or profile repeats in more than 60 % (or 80 %) of your closures or your best opportunities. It is not a strict mathematical rule; it is a practical criterion: if most of who buys from you shares a pattern, that pattern is useful for prioritising and for messaging. Below 60 %, the pattern may be noise.

Why not ask about my idea or product directly?

Because people often say yes out of politeness or speculation about the future. You do not get reliable information. It is better to ask about past experiences and real needs: «how do you solve it today?», «what have you tried before?». That way you discover the context and the real «job», not a false validation.

How often should I repeat the Research–Patterns–Survey–Contrast process?

When the market, the offer or the type of customer you target changes. There is no fixed frequency; what matters is not leaving the profile frozen. Reviewing with real data (closures, losses, surveys) at least once or twice a year is usually a good habit.

Process

10 March 2026

Defining the customer well —whether with target audience, buyer persona, Jobs to be Done or ICP— requires data and a method to avoid falling into stereotypes or the «typical salesperson». This article describes a four-phase process: (1) Research: put all available information in one place, with sources, surveys and tools; (2) Patterns: look for patterns based on data, with useful correlation thresholds; (3) Design survey: check whether hypotheses have a real basis, asking to learn and not to validate; (4) Contrast: compare survey results with theories. It is a practical framework for the pre-demo phase and for aligning research and sales.

1. Research: put all information in one place

The first step is to centralise all the information you have. If data is scattered across sheets, emails and people's heads, hypotheses about the customer are built on intuition or clichés. Grouping everything in one space —a document, a repository or a shared tool— lets you see what you really know and what is missing.

a. Information sources

Include external databases (company lists, sector, size), internal databases (CRM, customer history, won and lost opportunities), the sales team (what they see day to day, what objections they hear, which profiles close) and communities (forums, networks, events where your audience is). Combining internal and external sources reduces the bias of «we only look at who already buys from us» and adds market context.

b. Surveys

Ideally, post-sale questionnaires to get information from the real customer and not just from those who do not buy. Those who have already bought can explain what motivated them, what they hesitated about, who was involved in the decision and what problem they solved. Surveys of non-customers are useful for contrast, but if you only ask those who do not buy from you, you may be refining the profile of someone who will never buy. Prioritise surveys of customers and recent lost opportunities.

c. Tools and data

Incorporate data from GA4 (web behaviour, traffic sources, conversions), Google Trends (interest in topics or terms related to your offer) and social networks (where the audience is, what content they consume, what questions they ask). These tools do not replace interviews or team judgement, but they provide a quantitative base so you do not rely only on anecdotes.

2. Patterns: based on data, not stereotypes

With the information centralised, the next step is to look for patterns. The important thing is to base this on data and not fall into stereotypes, clichés or the «typical salesperson». If a profile repeats across your customers or your won opportunities, quantify it: what percentage of closures share that trait or situation?

A practical rule of thumb: over 60 % correlation with a trait or context is considered good; over 80 % is very solid. You do not need 100 %; what matters is that the pattern is real and actionable for prioritising accounts and messages. If you do not have enough data, document the hypothesis and design the next phase —the survey— to test it.

3. Design survey: ask to learn

Now we want to understand whether our hypotheses have a real basis. The temptation is to ask to validate what we already believe; the goal should be to ask to learn. For that, follow a few simple rules.

Do not ask about your idea specifically: ask about their experiences and needs. If you ask «would you use a product that does X?», most will say yes out of politeness or speculation. Better: «how do you solve [the problem X addresses] today?» or «what have you tried before?».
Ask about past experiences: the future does not exist in the interviewee's head; any answer about what they «would» do or «would» buy is speculation. Focus questions on what they have already done, decided or tried.
Do not stay on the surface: if they talk about the problem, ask again. «Why?», «what happened then?», «what would you have needed?» lead to the real cause and context, not the commonplace.

These rules align with Jobs to be Done practice and with qualitative research in customer discovery: the goal is to understand the «job» and context, not to get a «yes» to your product. Organisations such as the Nielsen Norman Group document good user research practices that can serve as a reference.

4. Contrast: surveys versus theories

The last phase is to contrast: compare the results of your survey set with your theories and hypotheses. Do the patterns you identified in the Research phase hold up? What surprises appear? Which segments or pains gain weight and which do not?

With that contrast you can update the customer profile (whether target audience, buyer persona, job or ICP), refine messages and qualification criteria, and better prepare the pre-demo phase. The process does not end in a single cycle: it is worth repeating Research, Patterns, Survey and Contrast when the market or the offer changes.

Fitting the process with client profiling and ICP

This process fits directly with ideal customer definition and client profiling. Research and Patterns feed the ICP (who to look for, what pain, where to find them); a well-designed survey and contrast feed the operational customer profile (how they decide, what objections they have, what they have tried). If you want to go deeper on ICP or client profiling, in our pre-demo resources you have the framework; here you have the method to fill it with real data.

Next steps

If you want to apply this process to defining your ideal customer or designing your pre-demo research, we can review your case in a no-obligation call. At Miranda's Consulting we work with sales teams on pre-demo preparation, qualification and conversion improvement.

Frequently asked questions

Why post-sale surveys and not only non-customers?: Because information from customers who have already bought best reflects who actually buys from you: what motivated them, what they hesitated about, who decided. If you only ask those who do not buy, you may be refining the profile of someone who will not close. Non-customer surveys are useful for contrast, but prioritise customers and recent lost opportunities.
What does «over 60 % or 80 % correlation» mean?: That a trait, context or profile repeats in more than 60 % (or 80 %) of your closures or your best opportunities. It is not a strict mathematical rule; it is a practical criterion: if most of who buys from you shares a pattern, that pattern is useful for prioritising and for messaging. Below 60 %, the pattern may be noise.
Why not ask about my idea or product directly?: Because people often say yes out of politeness or speculation about the future. You do not get reliable information. It is better to ask about past experiences and real needs: «how do you solve it today?», «what have you tried before?». That way you discover the context and the real «job», not a false validation.
How often should I repeat the Research–Patterns–Survey–Contrast process?: When the market, the offer or the type of customer you target changes. There is no fixed frequency; what matters is not leaving the profile frozen. Reviewing with real data (closures, losses, surveys) at least once or twice a year is usually a good habit.