Methodology
This page is the contract between us and everyone who reads this site. If we ever break a rule written here, the site is not worth reading. Please hold us to it.
What we are, and what we are not
VA Index is an independent index of virtual assistant companies. We do not supply virtual assistants, place staff, broker deals, take a commission on any engagement, or accept payment to recommend a provider. We are not affiliated with any company listed here, and we are not an auditor or a certification body.
Where the data comes from
Every field is taken from a page the company itself publishes, or from an independent review platform. We do not use private data, and we do not contact companies to ask for figures they have not published — what a company chooses to publish is itself part of what we are recording.
When we record a figure, we store the sentence it came from, the URL, and the date. A quote that cannot be matched against the archived page is discarded, not published. That check is mechanical: the sentence either appears in the source or it does not.
How we collect pages
We think a directory that publishes other companies’ data owes them a plain account of how it was gathered. So, precisely:
| We identify ourselves | Our crawler sends VAIndexBot/1.0 (+https://vaindex.com/methodology; [email protected]). We do not disguise it as a browser. |
| We obey robots.txt | If a domain’s robots.txt cannot be read, we do not crawl it at all. Disallowed paths are skipped and recorded as skipped. |
| We go slowly | At most one request every four seconds per domain, at most a dozen pages per site, re-checked monthly. We are not a load on anyone’s server. |
| We never work around a barrier | Login walls, paywalls, bot checks, 403s — we stop and record the field as not disclosed. We do not rotate addresses, solve challenges, or use logged-in sessions. |
| We only read public business pages | Pricing, FAQ, about, compliance. Nothing behind an account, nothing from a private source. |
| We do not collect data about individuals | No assistant names, photos, CVs, personal emails or phone numbers — ever. This index is about companies. Individual assistants never agreed to be listed anywhere, and we have no business putting them in a searchable database. |
| We quote, we don’t republish | Quotes are capped at roughly a sentence. We do not reproduce third-party review text, and we do not rewrite a company’s own description as if it were ours. |
If you would rather we did not crawl your site at all, say so at [email protected] and we will stop — no conditions attached.
What each badge means
| ✓ Verified from source | Quoted verbatim from the company’s own page and matched against our archived copy of it. Click the ⓘ to read the sentence and open the source. |
| ◇ Third-party data | Taken from an independent source rather than the company itself. Not currently in use — review-platform ratings are on hold while we work out a route that respects those platforms’ terms. We would rather show nothing than scrape something we shouldn’t. |
| ❝ Company-stated | A claim the company makes that we cannot independently check — typically HIPAA, BAAs, background checks, certifications. We report the claim. We do not endorse it. Verify these yourself before signing anything. |
| — Not publicly disclosed | We looked and found nothing. The pages we checked are listed. This is not a gap in our data; it is a fact about the company. |
On compliance claims
How ordering works — and what money cannot buy
Default ordering on every listing page is alphabetical by company name. Nothing else.
When paid listings open, a paid tier will affect two things only: whether a listing is
visually highlighted, and whether the company’s outbound link is do-follow.
It will never affect:
- the default ordering of any listing page
- the content, wording, or verification status of any field
- whether a company appears in the index at all
- whether unflattering data — undisclosed pricing, long lock-ins — is shown
Paid listings are labelled as such wherever they appear. If you ever find a case where this is not true, tell us, and we will fix it publicly.
How often data is re-checked
Every record is re-checked monthly. Fields that have not changed have their “last verified” date updated. Fields that have changed are re-extracted and reviewed before publication. Each company page shows both the last check and the next scheduled one.
We also keep a small hand-verified reference set and re-run it whenever the pipeline changes. If field-level accuracy against that set drops below our threshold, the change does not ship.
Corrections, claims and removal
- Something is wrong. Email [email protected] with the page and the field. We respond within 24 hours.
- You work at a listed company. You can claim your record and correct it yourself, free, permanently. Corrections are sanity-checked, not gatekept.
- You want the entry gone. Any company may have its record removed on request, with no conditions and no questions asked.
Known limitations
We can only verify that a company said something — not that what it said is true. Where information sits behind a contact form, a PDF or a sales call, we cannot record it, and we mark it as not disclosed rather than guessing. Some sites block automated access; those records will be sparser than others, and that is a limitation of ours, not a judgement about the company.