Why Pairwise Comparison Works
for Client Discovery

Why asking clients to choose between competing priorities can reveal more than asking them to rate everything highly.

Speed, price, certainty, flexibility — ask someone to rate each one out of ten, and you'll get eights and nines across the board. Not because they're being difficult. It's just how people answer rating questions. Everything feels important when you're allowed to say so.

Ask them to choose instead, and the picture changes.

Overwhelmed by options

Confident, clear choice

The problem isn't the questions — it's the format

This isn't a KYP idea. It's a known problem in survey research, and a known fix.

Rating scales suffer from what researchers call response style bias. Some people agree with whatever is put in front of them. Some never go near the top of the scale. Some cluster everything around the middle. From the numbers alone you cannot tell which of those you're looking at — a nine from one person and a six from another may describe exactly the same strength of preference. The result is a wall of high scores that tells you very little about what comes first.

Forced choice removes the problem at source. When someone has to pick one of two, there is no scale to use generously or sparingly. They have to decide. And the decision carries information a rating never does: which one they picked, when picking cost them the other.

A 100-year-old idea with a modern application

L. L. Thurstone set this out mathematically in 1927, in A Law of Comparative Judgment. His argument was simple: people usually cannot put a reliable number on how much they want something, but they can usually say which of two things they'd take. The underlying idea is that enough small comparisons can reveal a person's relative ordering of preferences. KYP applies that principle using direct pairwise choices and a simple win-count ranking.

Thomas Saaty made the sharper version of the point in a 2008 survey paper for the Royal Academy of Sciences of Spain. Some things have measurement scales — length, weight, time. Many of the things that drive decisions do not. There is no unit of urgency, no agreed scale for how much certainty is worth to a particular person on a particular day. Where no scale exists, Saaty argues, comparison isn't a second-best substitute for measurement. It is the only route to a meaningful number at all, and the value of anything is established by what it is compared against.

That is precisely the situation a mortgage adviser is in. "How much does certainty matter to you, out of ten?" asks a client to read off a scale that was never built. "Would you fix your rate, or take the cheaper variable?" asks them something they can actually answer.

The same principle underlies some of the most widely used commercial research in the world. A related approach, used widely in commercial preference research to understand which attributes matter most to customers, is best-worst scaling — forced comparisons, not ratings, for exactly this reason.

Best-worst scaling is a later descendant of Thurstone's method. KYP uses the original: direct pairwise comparison, two statements at a time, one question — which of these matters more to you?

One difference is worth stating plainly. Saaty's own method asks people to judge how much one thing dominates another, on a scale of one to nine, and derives priorities from the resulting matrix. KYP does not do that. It asks only which of two, and counts wins. The reasoning for comparing rather than rating is the same; the arithmetic is deliberately simpler, and produces an order rather than a set of weights.

How a KYP session is built

Each session covers several stages: Borrowing Priorities, Mortgage Process, and Family & Protection Priorities.

Each stage holds a small set of statements drawn from real practitioner language. Every statement is compared against every other one, so whilst the pairs are rated in isolation, it is all in context — it's ranked purely by what it beat. Nothing is rated, nothing is scored out of ten, and there is no neutral option to fall back on.

Six isn’t an arbitrary number. George Miller’s 1956 work on short-term memory found that people can reliably hold and distinguish around seven items at once, give or take two. Saaty applied that limit to comparison: beyond about seven elements, the differences between successive values become small enough that judgments lose sensitivity, so he recommends comparing only a few elements at a time and grouping larger sets. We use four statements per set as standard, and never more than six, keeping every comparison comfortably within that range while still giving enough choices to reveal meaningful differences in a client’s priorities.

The ranking is calculated from those selections alone — head-to-head wins, counted deterministically. No model estimates it, no algorithm weights it, and nothing is inferred from it. Where two statements finish level, the tie is settled by the direct comparison between them, and the report both flags it in the summary ranking and shows it explicitly in the full results table.

Every comparison you make appears in the report, with its pair and the choice made. You can check any ranking against the decisions that produced it, line by line.

The report also carries a short written summary. That prose is generated by an AI model from the finished ranking, and it is checked by a second, independent pass before it reaches you. The model writes the sentences; it does not produce the ranking, and it is instructed to state what was selected rather than what it might mean.

What this means for the report in your inbox

A KYP session isn't a survey. It's the same forced trade-off used in preference research, for the same reason: to get answers that reflect a decision rather than an intention to be helpful.

In practice, that means:

You walk in prepared. You know which priorities your client put ahead of which others, before the conversation starts.
Rapport builds faster. Opening with "I can see you chose payment certainty over the lowest rate several times" lands very differently to "so, what are you looking for?"
You see the trade-offs, not just the headline. A client who says "I want the cheapest mortgage that's also totally secure" has told you nothing. A client who chose 'certainty' over 'rate' more times in the session has told you something you can work with.
Fewer surprises mid-meeting. The priority that would have come out at minute twenty is on the page in front of you at minute two.

That's the difference between a client telling you what they think you want to hear, and a client showing you, one choice at a time, what they picked.

The full report — typically five pages, in the adviser's inbox within seconds of the client finishing.

See how pairwise comparison is used for mortgage advisers

What it doesn't tell you

Being clear about the limits is part of the method.

It gives you order, not distance. You learn that certainty came ahead of rate. You don't learn by how much. A narrow win and a landslide can produce the same ranking position — which is why the report lists every comparison and its outcome, not just the finishing order. A statement that won all three of its comparisons and one that won two sit in different places, and the table shows you which you're looking at.

It only ranks what was put in front of them. The statements in each stage are fixed. If something matters enormously to a client and it isn't among them, the exercise won't find it. The report tells you how they ordered these priorities, not that these are all of their priorities.

It's the day they took it. Preferences move — a chain falls through, a rate changes, a family conversation happens. The session records the choices made at that moment, not a fixed characteristic of the person.

It needs a willing participant. The exercise suits someone prepared to complete it, even reluctantly. Someone clicking through to get to the end is not producing a meaningful result, and the output shouldn't be read as one.

It isn't a fact find. KYP performs no affordability calculation, no suitability assessment, no creditworthiness assessment, and makes no recommendation of any kind. It records preferences expressed during a structured comparison exercise, and nothing else.

A conversation aid, not a substitute for one

KnowYourProspect records the choices a client makes during a structured comparison exercise, and reports them in the order they were made. What you do with that, and any recommendation that follows, is entirely yours.

A note on evidence

KnowYourProspect is a novel application of an established, standardised method for measuring preferences.

Why Pairwise Comparison Worksfor Client Discovery