GrowKittOriginal research, the verifiable kind
Every study here is built from live checks on real websites — never estimates, never "industry averages" copied from someone else's PDF. We publish the raw data with every study so you can verify any number yourself.
Why we publish research at all
GrowKitt is a marketing tool, and marketing-tool blogs are full of recycled statistics nobody can trace. We decided to do the opposite: run the checks ourselves, against live Indian websites, and publish every raw result — site by site, check by check. If a number on this page is wrong, you can catch us, because the dataset is downloadable and every check is a plain HTTP request you can repeat from your own laptop. That standard — checkable or it doesn't get published — applies to everything in this section.
The research also powers our guides: when a blog post claims Indian businesses are missing FAQ schema, that claim links back to the data here instead of to someone's opinion.
Current study: AI-readiness of Indian D2C websites (pilot)
In October 2026 we live-checked 25 real Indian D2C websites for llms.txt, AI-crawler access in robots.txt, structured data and on-page basics. The headline finding surprised us: every single site "has" an llms.txt file — and almost none of them put it there. Here is the one-table summary:
| Check | Result | What it means |
|---|---|---|
| Has an llms.txt file | 25 / 25 (100%) | All Shopify boilerplate — none hand-written |
| Blocks GPTBot (OpenAI) | 1 / 25 (4%) | Indian D2C is wide open to AI crawlers |
| Blocks PerplexityBot | 0 / 25 (0%) | No blocking at all |
| Has any JSON-LD schema | 23 / 25 (92%) | Platforms add the basics automatically |
| Has FAQPage schema | 3 / 25 (12%) | The markup that gets businesses QUOTED by AI |
| Title tag ≤ 60 characters | 9 / 25 (36%) | 64% overflow what Google displays |
The AI-Readiness of Indian D2C Websites
Full findings, per-site results table, complete methodology and the downloadable raw dataset (CC-BY).
Read the study + download the raw data →
Blog version · 7 minThe short, shareable write-up
The same findings as a plain-words blog post — what the numbers mean for your business and the exact fixes, with FAQs.
Read the blog post →
How our methodology works
Every check is a real HTTP request against the live site — fetching /llms.txt, parsing robots.txt for AI-crawler rules, reading the homepage HTML for JSON-LD blocks, title length and meta tags. No scraping services, no third-party databases, no sampling tricks: the study page documents the exact request for every column, and the raw JSON lists every site's result for every check. Limitations are stated plainly on the study page too — a 25-site pilot is a pilot, not a census.
What's next
The full two-cohort study
Next: ~200 sites in two cohorts — online-first D2C brands vs traditional service SMBs (clinics, CAs, manufacturers) — with a documented sampling method. The readiness gap between the two is the question. Follow the RSS feed or the blog to catch it.
Where does your site stand?
The free scan runs these exact checks (and more) on any website in about a minute — no signup. The GrowKitt app runs them on your whole site, monthly.