Before interviewing 90 synthetic decision-makers, we wrote down what we expected to find

All respondents in this study are synthetic. This is synthetic pre-research on B2B parcel carrier switching in Spain.

When a synthetic study "discovers" something, the first question should be: would anyone in the industry have said it anyway? If you cannot answer that, there is no way to tell whether the study adds anything or just repeats the obvious in more words.

So in this pre-research we did something simple: before running a single interview, we wrote down six hypotheses, and for each one, the result that would knock it down. They were written and dated on 4 and 5 October; fieldwork was on the 6th. Then we compared.

What we did

• Question: what makes a Spanish company switch parcel carrier, what holds it back, and how it reacts to a price proposal versus a service proposal.
• Who: 90 synthetic decision-makers who buy parcel transport in Spain, from e-commerce, industry, and pharma and healthcare.
• How: a 12-question conversational interview, all open questions. The two proposals (one on price, one on service) come at the end, after everything spontaneous, so they cannot contaminate it.

What we wrote down beforehand

• H1. What makes companies switch is a badly handled incident, not a price increase.
• H2. Those who evaluated switching and stayed are held back by the risk of switching, not by satisfaction.
• H3. Those who do not look for alternatives do so out of reasoned inertia, not lack of knowledge.
• H4. A price proposal is objected to on risk; a service proposal, on credibility.
• H5. In pharma and healthcare, the objection to the price proposal is harsher.
• H6. Before signing, what people ask for is proof, not an argument.

What came out, hypothesis by hypothesis

Quotes are translated from Spanish.

H4 holds halfway

The cheap proposal is objected to on risk; the expensive one, with "prove it":

"Promising is free. I want to see it."

"Let them show me their failures."

Why only halfway: the price proposal handled incidents through a web form with a 48-hour response, and it is that form that triggers the objection. We cannot separate how much comes from the price and how much from the form we put in it.

H5 falls, or at least not as we wrote it

Pharma and healthcare do not reject the price proposal more harshly than e-commerce or industry. What changes is the reason: in pharma it is the cold chain and the hospital. That is sector context, not a finding.

"[...] if I save 12% but I find out about the breakage from the hospital, I have not saved anything, I have bought a cheaper problem."

H6 holds

What they ask for before signing is a trial with conditions, almost never a discount:

"[...] a one-month trial. Twenty shipments, and they call me when something goes wrong."

H1, H2 and H3 we cannot evaluate yet

Each one is read in a different group: those who switched carrier this year, those who evaluated switching and stayed, and those who renew without looking. When we reviewed the interviews we found that part of the respondents in the last group say they did ask for quotes, the opposite of what defines their group. Until we review that, we do not compare groups, and we do not call these three confirmed or refuted.

What was not on the list

Two ideas we had not written down that come up in more than one interview. They are hypotheses for fieldwork, not facts.

The threshold for switching is a pattern of failures, not a saving.

"[...] I do not switch over a rate. I switch over a pattern."

"Two failures in a row. And the sales rep not picking up."

Some keep their own incident log, so they have numbers when they are asked:

"I log every incident myself. Date, customer, what happened."

What we separated and what we threw out

A synthetic study is also measured by what it does not publish.

• What they said unprompted versus what they said on hearing the proposal. The rejection of the form with 48 hours is very strong, but it appears when we read them our proposal. We count it as a reaction to the proposal, not as their own reason to switch.
• The model's verbal tics. The same day of the week appeared in more than half of the interviews, in questions that had nothing to do with it, and several different people used the same stock phrases. It was not a finding: it was a tic. None of the quotes in this article contain them.
• Ideas from a single person. A brilliant line from one respondent is a question for fieldwork, not a result.
• Brands. Respondents name the real carriers in their market. We do not name them here: the opinions of synthetic people should not be attributed to real companies.

What we would take to fieldwork with real people

• How many failures, and of what kind, would make you look for another carrier? Do you have it defined, or do you decide in the heat of the moment?
• Do you keep your own incident log? What do you use it for and who sees it?
• When you have defended a switch to management, what figures did they ask for, and what actually convinced you?
• What would have to happen in a trial with a new carrier for you to consider it passed?
• A timely delay warning: does it count as delivering or as failing?

Technical sheet

• 90 synthetic respondents (StrataSynth). Spain. 3 sectors (e-commerce, industry, pharma and healthcare) and 3 buying behaviours, 10 per cell.
• Joint reading of the 90 interviews; this article does not compare groups with each other.
• Conversational interview, 12 questions, in Spanish from Spain; proposals at the end. The two proposals were always read in the same order (price first). Order was not rotated, and that may affect the reaction to the second.
• 1,080 responses. Mean conversational quality score 0.85; no interview excluded.
• The hypotheses were written on 4 and 5 October 2026, before fieldwork. The dated document is available on request.
• Use of AI: respondent generation, interview, analysis and report.
• Fieldwork: 6 October 2026.

What this study does not tell you

The 90 respondents in this study are synthetic: there is no real person behind these answers. It is for preparing a study with real people: sharpening the guide, finding the objections worth exploring, and reaching fieldwork with better hypotheses.

It does not tell you how many companies would switch carrier, which proposal the market would choose, or how much each reason weighs in the population: it is not a sample. It does not predict behaviour or replace fieldwork. Each respondent's description says who they are, not what they will answer.