Original research

Nothing published here yet. This page says what would be.

Ariadne intends to publish original research on how service businesses actually behave between an enquiry and a customer. Many claims in this category are vendor-produced or presented without enough methodology to evaluate them.

No report exists today. Rather than leave an empty section or fill it with numbers borrowed from a marketing blog, this page describes the first planned study and the standard it will be held to.

Planned studyNot started. No data collected, nothing published.

Toronto Service Business Digital Benchmark

Across Toronto service businesses, how long does a web enquiry actually wait for a first human response, and how many are never contacted a second time?

Intended method

  • A defined sample frame: publicly listed service businesses in specified categories within the GTA
  • An identical, disclosed enquiry submitted to each, from an address that identifies itself as research
  • Time to first human response measured to the minute
  • Second-contact attempts recorded over a fixed observation window
  • Public technical measures collected: Core Web Vitals, mobile usability, structured data presence

What would be published

  • The full methodology, including how the sample was drawn and its limitations
  • The raw aggregate dataset, downloadable
  • Every business anonymised — the finding is about the category, not about naming anyone
  • What the data does not support, stated as plainly as what it does

The standard any report here will meet

Research that supports a commercial argument has an obvious incentive problem, and the only reasonable response is to make the work checkable.

  • Method published before findings. Sample frame, collection period, and analysis approach stated so the result can be evaluated rather than trusted.
  • Data released. The aggregate dataset downloadable, so anyone can check the arithmetic or reach a different conclusion from it.
  • Limitations stated up front. Sample size, selection bias, and what the data cannot support, in the report rather than in a footnote.
  • Participants anonymised. The finding is about a category. Naming individual businesses would be both unkind and worse research.
  • No claim beyond the evidence. If a finding is directional rather than significant, it gets described that way.

If a report ever appears here that does not meet all five, it should be treated with exactly the suspicion that standard exists to prevent.

Why this matters commercially

Most of what gets asserted about response times, conversion rates, and follow-up in this industry traces back to a vendor blog citing another vendor blog. That is why this site avoids benchmark statistics almost entirely and instead tells you to measure your own numbers. Publishing real data would be a better answer than avoiding the question, which is the reason to do it properly.

In the meantime

The insights section contains explanatory writing rather than research: how the seven layers work, how to diagnose which checkpoint is failing, and how growth engineering differs from the alternatives. Nothing there depends on a statistic that cannot be sourced.

If you research this space and want to collaborate, or you would like to be told when the first study is published, email support@ariadne.fyi.

Read the insights

Your practical starting point

Your own numbers beat any benchmark.

A growth system audit measures what your business actually does — real response times, real capture rates, real attribution — rather than comparing you to an industry average nobody can source.

Request a growth system audit