Conecxt
Create account

Resources Insight

One in five .in homepages links WhatsApp, one in 160 shows a platform: our 113,102-site scan

Most WhatsApp Business statistics for India come from Meta or from vendors counting their own customers. We wanted a figure we could check, so we read the August 2026 Common Crawl copy of every Indian homepage it held and counted the WhatsApp links, the platform code and the chat widgets. This is the method article for the whole series.

Conecxt team 12 min read

Share WhatsApp LinkedIn

The short version

  • In Common Crawl's August 2026 archive, 24,059 of 113,102 .in homepages (21.3%) carry a WhatsApp click-to-chat link. Only 718 (0.63%) carry code or a domain from a WhatsApp platform vendor.
  • That is about 33 homepages running WhatsApp by hand for every one with a platform fingerprint. Platform use is undercounted by this method, so 33 to 1 is an upper bound.
  • Of the links we could read, 98.2% point to an Indian number, 9.1% are malformed (ten digits, no 91) and 5.7% of pages list two or more numbers.
  • co.in homepages link WhatsApp most (26.6%) and ac.in least among the common suffixes (12.5%). Another 43,025 .com homepages carry an Indian number, taking the combined count to 67,084.
  • The scan reads homepages only, as crawled between 7 and 20 August 2026, so the true share of Indian businesses with a WhatsApp link is higher than 21.3%.

Most WhatsApp Business statistics for India come from Meta, or from vendors counting their own customers. We wanted a number we could check ourselves. So we took the August 2026 archive of Common Crawl, an open copy of the public web, and read the homepage of every .in domain in it: 113,102 of them. On each page we looked for three things: a WhatsApp click-to-chat link, the code or domain of a WhatsApp platform vendor, and a generic website chat widget.

Three numbers came out. About one in five .in homepages links WhatsApp. About one in 160 shows a platform. And for every homepage with a platform fingerprint, 33 run WhatsApp by hand, on a plain link to a phone number.

We make Conecxt, a WhatsApp platform, so we have an interest in that third number. Conecxt leaves no homepage fingerprint, so it is absent from every platform count in this series, like every other tool that works through the API alone.

The three numbers

All percentages are of the 113,102 .in homepages, after the cleaning step described in the method.

MeasureHomepagesShare
Any signal (link, platform or widget)26,50523.4%
WhatsApp click-to-chat link24,05921.3%
Platform fingerprint (any of 137 detectors)7180.63%
Generic website chat widget3,1282.77%
WhatsApp link and no platform fingerprint (“manual WhatsApp”)23,86021.1%
Chat widget only, no WhatsApp link1,9271.70%
WhatsApp link and a platform fingerprint on the same page199
WhatsApp link and a chat widget on the same page1,162

The cohorts don’t overlap: any platform fingerprint makes a page “platform”, a WhatsApp link with no fingerprint makes it “manual WhatsApp”, and a generic widget alone makes it “widget only”. Manual to platform is 23,860 to 718, which we round to 33 to 1.

What we scanned and how

Common Crawl, in plain words

Common Crawl is a US non-profit that crawls the public web and publishes the result as open data. The August 2026 archive, labelled CC-MAIN-2026-34, holds 2.14 billion pages captured between 7 and 20 August 2026; the figure is on the archive’s own index page, which we read on 15 October 2026. It is a very large sample of the web, not a register of Indian businesses.

What we selected

From the archive’s URL index we took every record that was a homepage (the URL path was just ”/”), came back with HTTP status 200, was HTML, and whose host ended in .in. One homepage per registrable domain, so a shop subdomain and its parent count once. That gave 113,102 .in domains. We did the same for .com, which gave 8,497,286 domains worldwide; that set gets its own article.

How we fetched

Each homepage was pulled from the archive by byte-range request: the HTML as Common Crawl saw it in August, not the live site. Of the .in set, 113,084 came back and 18 failed. The .com fetch ran on a fleet of cloud machines and targeted all 8,497,286.

We looked for three kinds of signal in the HTML.

  1. A WhatsApp click-to-chat link: wa.me/, api.whatsapp.com/send, web.whatsapp.com/send or whatsapp://send (and wa.link/ on .com).
  2. A platform fingerprint: the registrable domain of one of the 119 companies in our June 2026 copy of Meta’s Business Partner directory, or of 15 WhatsApp tools common in India that were not in that copy (Interakt, Zoko, Gallabox, DoubleTick, LimeChat, MSG91, Verloop, BusinessOnBot, Kaleyra, Trengo, Rasayel, Wappbiz, Whatso, Bikayi and Zixflow), plus three widget markers (the AiSensy widget, the Wati Shopify app and the Gallabox widget). That is 137 detectors. Every pattern is anchored to the vendor’s domain, never a brand word, so “Swati” never matches Wati.
  3. A generic website chat widget: Intercom, Freshchat, Tawk.to, Crisp, Tidio, Zendesk, Chatwoot, Drift, LiveChat or HubSpot chat.

How we read the numbers

WhatsApp’s help page on click to chat says the link takes the full phone number in international format, with zeros, brackets and dashes left out: https://wa.me/91XXXXXXXXXX for an Indian number. We kept every number in that full form. A link that carried only the ten national digits, with no 91, was recorded as bare. Such a link is malformed and will not open the chat the business meant. The WhatsApp Business app generates a correct short link for every account, so a bare link was most likely typed by hand.

How we cleaned

We removed homepages whose title reads as a parked or placeholder page (“for sale”, “coming soon”, a default server page), pages whose only numbers were demo numbers (a run of ten zeros, the keypad digits in order, ten nines and the like), and any number shared by 20 or more domains, which is what a domain-parking service’s number looks like. That removed 692 .in homepages and 6,327 .com ones. All tables here are after cleaning; the raw “any signal” figure was 27,197 (24.0%).

How many businesses in India use WhatsApp on their website?

On .in homepages, 21.3%. Add the .com homepages that carry an Indian number, covered below, and the total is 67,084 (24,059 plus 43,025). Both are counts of homepages, not businesses, and both are floors: a business whose WhatsApp link sits on the contact page and nowhere else counts as “no link” here.

The link comes in four forms; a page can use more than one.

Link formatWhatsApp-linking .in homepagesShare of 24,059
wa.me/13,21554.9%
api.whatsapp.com/send10,24342.6%
web.whatsapp.com/send2,1068.8%
whatsapp://send4491.9%

web.whatsapp.com/send forces the desktop client on a phone, and whatsapp://send fails wherever WhatsApp isn’t installed; the link mistakes article goes through both.

Of the 24,059 WhatsApp-linking homepages, 21,759 (90.4%) carried a number we could read from the link. The rest were text-only, catalogue or QR links.

Among the readable ones, 19,661 (90.4%) carried a country code we could parse, 1,988 (9.1%) were bare ten-digit numbers with no 91, and 110 (0.5%) had a prefix we could not recognise. So roughly one WhatsApp link in eleven does not work as the business intended.

The first number on the page points to India on 21,376 pages (98.2%). After that come US and Canada numbers (68), Indonesia (38), the UAE (34), Pakistan (16) and the UK (15). Of the 22,797 Indian numbers we extracted, 22,604 (99.2%) have the ten-digit mobile shape, starting 6 to 9: business WhatsApp in India runs on mobile SIMs.

Most pages list one number: 20,520 (94.3%). Two appear on 1,068 pages, three on 105, four on 24 and five or more on 42. Two or more, together, is 1,239 pages (5.7%), the subject of two WhatsApp numbers on one website.

Before cleaning, about 280 homepages (1.2% of WhatsApp-linking pages) linked a number nobody owns: ten zeros on 123 pages, the keypad sequence on 55, ten nines on 42, a nine and nine zeros on 17. One US number sat on 316 homepages, all “domain for sale” pages. All of that went in cleaning.

Common Crawl also records each page’s language. Of the WhatsApp-linking homepages, 23,483 (97.6%) are in English; Hindi follows at 236, then Marathi (66), Malayalam (18), Tamil (17), Bengali (16), Gujarati (16), Kannada (15) and Telugu (11). The Indian web that links WhatsApp is written in English, even where the chats that follow won’t be.

An invented example of how the bare-number mistake survives: Ritika Malhotra runs a nine-person chartered accountancy practice in Chandigarh. Her web designer pasted the button in 2023 as wa.me/ plus her ten digits. Nobody noticed for two years, because the number is printed under the button and clients typed it in. The link was decoration.

SuffixHomepages fetchedWith a WhatsApp linkRateWith a platform fingerprintRate
.in (second level)87,92918,75421.3%5140.58%
co.in15,8154,20226.6%810.51%
ac.in2,85635812.5%70.25%
org.in2,08424011.5%60.29%
edu.in1,96033317.0%110.56%
gov.in924444.8%646.9%
nic.in65530.5%335.0%
net.in5469216.8%10.18%

co.in, the suffix companies pick on purpose, links WhatsApp at the highest rate, just over one in four. Colleges and schools on ac.in and edu.in link it less, and almost never with a platform; the education article covers those 691 homepages. The platform rates on gov.in and nic.in look high, but those fingerprints are chatbot and CRM vendors on government portals, not WhatsApp inboxes.

The .com extension: 43,025 more Indian homepages

Indian businesses don’t only register .in. In the .com scan, 43,025 homepages link a +91 number, more than any other country’s. Of those, 42,643 show no platform fingerprint and 382 (0.9%) do. Two or more numbers are commoner there, on 3,096 pages (7.2%), and 42,212 (98.1%) are in English.

Put the two sets together and India has 67,084 homepages with a WhatsApp link and 1,100 with a platform fingerprint (718 plus 382). We did not scan .co, .org, .net, .shop or .store, so even this is not the whole picture.

The widget layer

A generic chat widget sits on 3,128 .in homepages (2.77%). Tawk.to is on 2,079 of them (1.84% of all homepages), then HubSpot chat on 386 (0.34%), Tidio 296 (0.26%), Zendesk 130 (0.11%), Crisp 118 (0.10%), LiveChat 71, Freshchat 48, Chatwoot 30, Intercom 26 and Drift 2. These are counts of the vendor’s code on the page.

Across the .in set, 1,927 homepages have a widget and no WhatsApp link; 1,162 have both. The chat widget article looks at who runs which, and what each can do that the other can’t.

What it means: a manual-WhatsApp market

WhatsApp Business API adoption, as far as a homepage can show it, is 0.63%. Read as a market map, that is the smallest thing on the page. The vendors in Meta’s partner directory, plus the Indian tools we added, show up on 718 .in homepages between them. Meanwhile 23,860 homepages link WhatsApp straight to a phone, nearly all mobile numbers, which in practice means someone’s handset. The contest is not platform against platform; it is the plain link, and the phone behind it, against everything else.

Where that phone is the owner’s, the link works until enquiries outgrow one person’s day. Where it is a staff member’s, the chat history leaves with them. Adding a second number to the page, as 5.7% have, is the cheapest answer to either limit, and it doesn’t solve the history problem. On the WhatsApp Business Platform the replies themselves are charged from 1 October 2026 once a number passes 1,000 service messages in a month, at ₹0.115 per delivered message to India before GST (Meta’s pricing page, updated 30 September 2026). That is small against the cost of a lost thread. If any term is new, the glossary has it.

Every caveat

  • Common Crawl is a very large sample of the public web, not a census of Indian businesses. A site has to be reachable and crawlable to be in it; large, well-linked sites are over-represented and tiny or new sites under-represented.
  • Homepage only. A business that links WhatsApp from its contact page but not its homepage counts as “no link”. The real share of businesses with a WhatsApp link is higher than our figure.
  • A platform fingerprint means the vendor’s code or domain is on the homepage. A business that runs a provider’s inbox behind a plain wa.me link leaves no trace, so platform use is undercounted and the 33 to 1 ratio is an upper bound. Conversely, some vendors on Meta’s partner list sell many things (Salesforce, HubSpot, Twilio, Zendesk): a fingerprint there means “runs that vendor’s code”, not “uses the WhatsApp API through it”.
  • Industry labels, where the series uses them, come from words in the homepage title (“dental clinic”, “coaching”, “real estate”), so they are approximate. We write “homepages whose title mentions”, never “clinics”.
  • The .com set is worldwide. A .com site is attributed to a country by the country code of the WhatsApp number on it, not by where the business is registered.
  • One business can own several domains. Every count here is of homepages, not companies.
  • The figures describe August 2026. Any site can have changed since.

What we publish separately

Each takes one slice of the scan and goes deeper than this article can:

To cite the study: Conecxt, “WhatsApp on Indian websites, 2026”, a scan of Common Crawl CC-MAIN-2026-34, analysis of 15 October 2026. For the per-suffix or per-industry tables as a file, or to query a count, write to us through the contact page.

Common questions

How many businesses in India use WhatsApp on their website?

In our scan of Common Crawl's August 2026 archive, 24,059 of 113,102 .in homepages (21.3%) carried a WhatsApp click-to-chat link, and a further 43,025 .com homepages carried an Indian WhatsApp number. The scan reads homepages only, so a business that links WhatsApp from its contact page is not counted. The true share is higher.

How many Indian businesses use a WhatsApp Business API platform?

Few leave a trace on their homepage. Of 113,102 .in homepages, 718 (0.63%) carried a platform vendor's code or domain, against 23,860 that linked WhatsApp with no platform fingerprint: about 33 to 1. A business running a platform inbox behind a plain wa.me link leaves no fingerprint, so the ratio is an upper bound.

What is Common Crawl and why use it for WhatsApp statistics?

Common Crawl is a US non-profit that crawls the public web and publishes the pages as open data. Its August 2026 archive holds 2.14 billion pages captured between 7 and 20 August 2026. We read the archived HTML of every .in and .com homepage in it rather than visiting live sites, which makes the count repeatable by anyone.

What is the most common WhatsApp link format on Indian websites?

wa.me, on 13,215 of the 24,059 WhatsApp-linking .in homepages (54.9%), followed by api.whatsapp.com/send on 10,243 (42.6%). Another 2,106 pages use web.whatsapp.com/send and 449 use whatsapp://send; a page can use more than one. 1,988 links carry a bare ten-digit number with no 91, which WhatsApp's link format does not accept.

Which Indian domain suffix links WhatsApp most?

co.in, where 4,202 of 15,815 homepages (26.6%) link WhatsApp, ahead of plain .in at 21.3%, edu.in at 17.0%, net.in at 16.8%, ac.in at 12.5% and org.in at 11.5%. gov.in sits at 4.8%. The platform fingerprints on gov.in and nic.in portals are chatbot and CRM vendors, not WhatsApp inboxes.

Sources

  1. Common Crawl, August 2026 crawl archive (CC-MAIN-2026-34) · Crawled 7–20 Aug 2026; page read 15 Oct 2026
  2. Common Crawl · Read 15 Oct 2026
  3. How to use click to chat (WhatsApp Help Centre) · Common Crawl capture of 6 Sep 2026
  4. How to create short links (WhatsApp Help Centre) · Common Crawl capture of 4 Sep 2026
  5. Pricing on the WhatsApp Business Platform (Meta for Developers) · Updated 30 Sep 2026
Share WhatsApp LinkedIn

Get the next one

One email when something new is published. Nothing else.

Unsubscribe from any email. See our privacy policy.

Run this on your own number

Create an account, connect your WhatsApp number, and invite your team.