Carcin publishes the research behind its claims. Each figure used anywhere on this site is a cited third-party statistic, a measured product fact, or it does not run. These are the packs that hold the citations.
One published pack. Every statistic in it names its publisher, its year, and links to the source document.
Small business owners have been sold to using invented numbers for a very long time. The market that Carcin sells into is the most statistic-battered market there is, and a made-up figure is the cheapest thing in it.
Carcin operates under a claims policy adopted in June 2026. It has one operative rule: every number that appears on a customer-facing surface is either a cited third-party statistic, a measured product fact, or absent. There is no fourth category. No estimates dressed as findings, no rounded-up figures that lost their source somewhere between a slide and a landing page, no simulated activity counters, no invented testimonials.
A policy like that is unverifiable if the citations live in an internal document. So they do not. The research section is where they are published, in full, with the links intact and the weak sources labelled as weak. If a statistic appears on the pricing page or the homepage, the pack it came from is here and you can check the original yourself.
Three rules govern every figure in every pack. They are deliberately restrictive, because the failure mode they prevent is the one nobody catches later.
Every finding links to a document anyone can open. A statistic quoted secondhand from an article that cites another article gets chased back to the original publisher or dropped. Most of what gets dropped is dropped here.
Publication and year stay attached to the figure wherever it appears, on the research page and off it. A statistic without its provenance is not evidence. It is a persuasion device, and it is exactly how bad numbers propagate.
Research packs cite their publishers. Product facts, such as price, are labelled separately. For client outcomes, explore the measured studies on the results page.
It gets labelled and kept, not quietly deleted. Some figures in these packs come from aggregated industry roundups rather than a single surveyed sample with published methodology. Those are marked as aggregated on the page so a reader can weight them accordingly. Removing an inconvenient source without saying so is a worse practice than publishing it with a caveat, because the reader cannot audit an absence.
The research is not a separate exhibit. It is the supply chain for the claims made everywhere else. The 57% figure on the homepage, the finding that customer acquisition is the top operational problem for US small employers, comes from the Federal Reserve Small Business Credit Survey and is cited in pack 01. The marketing confidence figure used alongside it comes from Constant Contact. The security framing on the comparison pages comes from the Verizon Data Breach Investigations Report. In every case the pack is the origin, and the page carries the publisher and year next to the number.
Product facts work differently and are labelled differently. The free workspace includes starter credits; paid plans have fixed monthly prices starting at $19.99 and included credits. That is a product fact, not a research finding. The channel library count describes channels Carcin has mapped, not campaigns launched for a customer. Neither claim is a survey result.
Client case studies document a specific engagement, its scope, and its measured results. Each published number should have a named source.
The packs are written to be quoted by people who have to defend the quote. That shapes the format.
Each pack is an index, not a primary source. Follow the link, confirm the figure in the original document, and cite the publisher. If you want to credit the compilation as well, name the pack and its compiled date.
Every line carries the year its survey was fielded or reported, not the year of the newest revision. Several of these are recurring instruments, so a fresher wave may already be published.
Each pack publishes its markdown source next to the page. The SMB pain pack is at /research/smb-pain.md, plain text, no scraping required.
Packs are published when there is a body of citable third-party work worth assembling, not on a schedule. Nothing below carries a date, because a research calendar is a promise about output rather than about quality.
Bring your business context into a private workspace and review what to investigate first.
Start freeFree workspace with starter credits · Paid plans from $19.99/mo · No card to start