56 Million Justdial Listings, 11.5 Million Udyam MSMEs: Mapping India's Discoverability Gap

Two counts, two different objects

Justdial describes itself as India's largest local search engine, and its own company overview states a database of approximately 56.1 million listings as of June 30, 2026, alongside 192.9 million quarterly unique users across web, mobile, app, and voice, and 160.5 million user-contributed reviews and ratings. A "listing" here is a business record inside Justdial's database. Most listings are not added, verified, or maintained by the business itself; a large share originate from Justdial's own sales and data-collection operations, and business owners have the option to list for free, with no requirement to claim, correct, or ever revisit the entry.

The Udyam Registration Portal measures something structurally different. It is a one-time, self-declared government filing, tied to a PAN and, increasingly, a GST number, that classifies an enterprise as micro, small, or medium under the 2020 investment-and-turnover criteria and confers eligibility for credit schemes, priority-sector lending, and government tenders. It is not a directory a customer would search. Since the classification took effect on July 1, 2020, cumulative Udyam registrations, across the Udyam Registration Portal (URP) and the newer Udyam Assist Platform (UAP), reached 7.83 crore, about 78.3 million, enterprises as of February 28, 2026, according to a written reply the Ministry of MSME gave in the Rajya Sabha.

Putting these two totals side by side is not an apples-to-apples comparison, and this piece does not treat it as one. One number counts appearances in a commercial directory, with no deduplication against a real legal entity and no requirement that the listing still be accurate. The other counts formal registrations, one per enterprise by design, but carries no information at all about whether that enterprise has any online or directory presence a customer, or an AI engine, would ever encounter. Both numbers get used loosely as shorthand for "how many businesses are out there." The gap between what each one actually measures is the more useful story.

The number in most people's heads is a 2022 number

The Udyam Registration Portal launched on July 1, 2020. It took 25 months to reach its first crore, 10 million, of cumulative registrations, a milestone the Ministry of MSME announced on August 2, 2022. By the end of FY 2022-23 in March 2023, the cumulative count had reached 1.64 crore, 16.4 million. For most of 2022 and into early 2023, in other words, the actual scale of "formally Udyam-registered MSMEs" sat in the low double-digit millions, bracketing the roughly 11.5 million figure that circulated widely in that period and that many people still associate with the program today.

That figure is now badly out of date, and the reason is a specific, documented policy change rather than organic growth. On January 11, 2023, the Ministry launched the Udyam Assist Platform (UAP), built specifically to formalize Informal Micro Enterprises, businesses without a GST number that could not complete a standard Udyam filing. Within its first 14 months, UAP alone added 1.50 crore, 15 million, registrations, a milestone the Ministry announced on March 4, 2024.

The combined trajectory since is steep. Cumulative registrations across URP and UAP stood at 4.12 crore (41.2 million) at the end of FY24, 6.19 crore (61.9 million) at the end of FY25, and 7.83 crore (78.3 million) by February 2026, per the same March 2026 Rajya Sabha reply that carries the full fiscal-year trend line. The current answer to "how many Indian MSMEs are formally registered" is not 11.5 million. It is roughly seven times that, and the overwhelming majority of the growth is attributable to one platform built for the exact population this piece is about: small, previously undocumented, informal micro-businesses.

What Justdial's own numbers say about its own listings

Justdial's April 2025 investor presentation, covering the quarter ended March 31, 2025, discloses a metric that speaks directly to how machine-legible its own database actually is. Total listings that quarter stood at 48.8 million, up 11.9% year on year. Of those, only 32.7 million, about two in three, carried geocode data, the structured latitude-and-longitude field a listing needs to place correctly on a map or surface in location-based search. Roughly one in three Justdial listings, by the company's own disclosure, was missing a basic piece of structured location data as of that filing.

The same presentation, and Justdial's June 2026 company overview, disclose active paid campaigns, businesses that pay to actively manage and promote their listing, separately from total listings: 613,290 paid campaigns against 48.8 million total listings in March 2025, and approximately 639,200 paid campaigns against 56.1 million total listings in June 2026. Dividing the two figures at each point puts actively managed, paying listings at roughly 1.1% to 1.3% of the total database, both times. That division is this article's own arithmetic on numbers Justdial discloses directly, not a separately sourced statistic.

What that ratio implies is straightforward: the remaining 98-plus percent of Justdial's listings are not under active commercial management. Some were added by a business itself years ago and never revisited. Some were added by Justdial's own field and data operations without an owner ever claiming or correcting the entry. Structurally, these are exactly the kind of records that drift out of date, accumulate near-duplicates under slightly different spellings of a business name or address, or continue to describe a business that has closed, moved, or rebranded since the listing was created.

Where the duplication and the disappearance actually happen

The premise this cluster is built around, that comparing Justdial's listing count against the Udyam registration count would show which businesses are duplicated many times over versus which never made it online at all, deserves a direct answer rather than an invented one: no public dataset performs that comparison today. Justdial's listing export carries no Udyam registration number field. The Udyam registry, built for lenders and scheme administrators, carries no directory-presence field. The two systems were built for different regulators, run on different identifiers, and have never been published in a form that lets one be matched against the other.

What the surrounding, verifiable evidence does support, stated as inference rather than measurement, runs in two directions at once. First, the structural conditions for duplication are well documented: a single small business can plausibly accumulate a Justdial listing, a near-duplicate Justdial entry from a separate data-collection pass, a Google Business Profile, an IndiaMART or Sulekha listing, and a Udyam certificate, several records describing one real shopfront, none cross-referencing another, because none of the systems that created them share an identifier. Second, a real remainder still sits outside every formal count. YourStory's review of Ministry of MSME data reported that Udyam registrations reached 6.2 crore in March 2025 against an official estimate of roughly 7.34 crore total MSMEs operating in the country, and that micro enterprises, the segment least likely to run a website or manage a directory listing, account for over 98% of everyone who has registered. Both directions can be true simultaneously: a searchable minority duplicated across every free directory in reach, and an unsearchable majority still outside the formal count and the directory layer alike.

A method exists for turning this from a structural argument into a measured one, and it is worth stating plainly rather than skipping past: a joined study would need to match records on a shared identifier, GSTIN, PAN, or a normalized name-address-phone triple, the same fields local-SEO deduplication work already uses, across a Udyam data extract and a directory's listing export. Neither dataset currently publishes a field the other side can match against. That gap, not a missing willingness to run the analysis, is why this remains an open measurement question rather than a reported finding.

Why the count matters less than the record

Regardless of the exact overlap, the more consequential question for a small business is not which directory or registry it appears in, but whether the underlying record is structured, consistent, and corroborated enough for a generative engine to use it. The 2024 peer-reviewed paper that introduced Generative Engine Optimization tested which content properties change whether a source gets cited inside an AI-made answer, and found that adding cited statistics, quotations, and authoritative sourcing measurably raised a source's visibility in the systems tested. The lever that earns a citation is corroboration, not directory count.

The stakes for getting this wrong have changed. A 2025 Pew Research Center browsing-panel study found that users clicked a traditional search result in about 8% of searches when an AI summary was present, against 15% without one, and clicked a link inside the summary itself only about 1% of the time. Ranking, or simply being listed, no longer reliably earns the visit it used to. Being named inside the answer increasingly does, and that depends on which version of a business's record the underlying system encountered and trusted.

A business duplicated across five free, unmanaged directory listings, each carrying a slightly different address, a disconnected phone number, or a category that no longer matches what it sells, and registered correctly exactly once under Udyam, has produced one clean government record and several conflicting public ones. The conflicting versions are also the more numerous and the more crawlable ones; the Udyam registry was built for lenders and scheme administrators, not for a language model synthesizing a local recommendation. That asymmetry, not the raw listing count, is the actual discoverability gap.

How to read these numbers

Two conclusions hold, and both resist the temptation to round the numbers into a cleaner story than the sources support. The first is that "56 million Justdial listings" and "11.5 million Udyam MSMEs" were never measuring the same population, and the second number is now roughly seven years and one deliberate policy intervention out of date; the current, PIB-sourced figure is over 78 million. The second is that neither total, current or historical, tells a business owner or a researcher whether a specific business is duplicated, stale, missing, or actually discoverable to the systems that increasingly decide what gets recommended. That measurement does not exist yet in public form, and this piece has tried to say so plainly rather than manufacture a percentage to fill the gap.

What the evidence supports is a structural read: India has built two enormous, separately maintained, non-interoperable records of small-business existence, a commercial directory optimized for advertising revenue and a government registry optimized for credit and compliance, and neither was designed with a generative answer engine as its reader. Closing that gap for a specific business starts with the same discipline as any other machine-legibility question, measuring where that one business actually stands across the surfaces that matter, rather than trusting either headline count to answer for it.