Why Indian addresses use landmarks instead of house numbers

"Opposite the temple, behind the water tank" is not laziness. It is the best fix people found for streets with no names and plots with no numbers.

A narrow lane in Ballimaran, Old Delhi, lined with old buildings, shop signs and overhead wires
Photo: Shadabgdg, CC BY-SA 4.0, via Wikimedia Commons

Indian addresses lean on landmarks because most Indian streets were never formally named and numbered in sequence, so "near the temple" is often the most precise shared reference available. Researchers at MIT and Delhivery estimated that about 80% of Indian addresses are written relative to a landmark, only about 30% follow a structured street-and-number format, and the average PIN code covers 179 square kilometres. Poor addressing costs India an estimated US$10-14 billion a year; India Post's DIGIPIN, launched in 2025, gives every 4 m by 4 m square its own 10-character code to fix the gap.

Key takeaways
  • About 80% of Indian addresses are written relative to a landmark; only about 30% follow a street-and-number format.
  • The average PIN code covers 179 sq km, and 20–40% of written PIN codes are wrong.
  • Poor addressing costs India an estimated US$10–14 billion a year, about 0.5% of GDP.
  • Planned cities such as Chandigarh show Indians use numbers when the streets are numbered.
  • DIGIPIN (2025) splits India into about 4 m squares: 36° ÷ 410 ≈ 3.8 m.

Ask for directions in most Indian towns and you will hear a chain of landmarks, not a number: second left after the petrol pump, the blue gate opposite the school. Outsiders see chaos.

In fact the landmark system is an efficient adaptation to how most Indian settlements grew: organically, plot by plot, faster than any municipality could name streets. This post explains where that comes from, what it costs, and what the new digital address layer changes.

~80%Indian addresses written relative to a landmark
179 sq kmaverage area of one PIN code
$10-14 bnestimated yearly cost of poor addressing
4 m × 4 msize of one DIGIPIN grid cell

Why do Indian addresses use landmarks?

Because for most places there is nothing more precise to use. A structured address needs three things: a named street, numbers assigned in order along it, and a register that keeps both up to date. Most Indian towns have at best one of the three.

A 2018 paper by Kabir Rustogi, Santanu Bhattacharya, Margaret Church and Ramesh Raskar, written from MIT Media Lab and logistics firm Delhivery's experience, put numbers on it. About 80% of Indian addresses are written with respect to a landmark, which typically sits 50 to 1,500 metres from the actual door.

Only about 30% of addresses are in a clean "26 Gandhi Road" style format; around 70% of sites have no street name at all (Rustogi et al., 2018). A typical address in the same town reads more like "B56 Niman Nagar, Near Green Park", with a colony name, a plot number and a reference point.

Several forces keep it that way:

  • Organic growth. Villages absorbed into cities keep their lanes, which were never surveyed as streets.
  • Unnumbered and subdivided plots. Plot numbers come from layouts, and when plots are split or built in stages, numbers stop being sequential or are duplicated across "blocks" and "phases".
  • Renamed streets. Official names change; people keep using the old one, or the landmark.
  • Many languages and scripts. The same place is transliterated many ways into English, so text matching fails.
  • Landmarks are stable socially. Everyone local knows the temple, the water tank or the old cinema, even if nobody knows the street's official name.
The story

The Pune-based non-profit Shelter Associates, which works in informal settlements, has recorded what a missing address costs in daily life. One resident, Bhagyashree, runs an electrical store. With no address of her own, she had to have her supplies delivered to the nearest landmark.

From 2018 the organisation began giving slum homes Plus Codes, short codes computed from a home's latitude and longitude, and by October 2023 it had coded more than 87,000 homes. Bhagyashree's goods now come straight to her home. Another resident, Rajashree, who runs a small shop in front of her house, used her code one monsoon to tell the authorities exactly where a manhole was overflowing with sewage, so they could fix it before it flooded her shop. The same address, Shelter Associates' director Smita Kale wrote, opens the door to bank accounts, ID cards, and electricity and gas connections that ask for a permanent address.

Google blog: How Plus Codes helped create new home addresses in India (Smita Kale, Shelter Associates, October 2023)
Wide plaza and modernist buildings in Sector 17, Chandigarh
Sector 17, Chandigarh. In a planned grid city, a sector number and house number get you to the door; most Indian towns never had that plan.Photo: Owais Khursheed, CC BY-SA 4.0, via Wikimedia Commons

Are planned cities like Chandigarh any different?

Yes, and they prove the point. Chandigarh, planned from 1949 onwards and built under Le Corbusier's team from 1951, was laid out as numbered rectangular sectors served by a seven-tier road hierarchy, from highways down to pedestrian paths. An address there, a house number in a numbered sector, resolves almost to the door without any landmark.

The city was inaugurated in 1953, its first developed sector was Sector 22, and it had about 10.55 lakh residents in the 2011 Census. Similar logic applies to planned colonies elsewhere: Delhi's DDA blocks, Navi Mumbai's nodes and sectors, many state housing board layouts. Where planning came first, Indians use numbers naturally. The landmark habit is a response to unplanned growth, not a cultural dislike of numbers.

Type of placeTypical addressHow precisely it locates you
Planned grid city (Chandigarh)House no., Sector no., cityUsually to the door
Planned colony in an older cityPlot no., block, phase, colonyUsually to the street; numbering may jump
Old city core or urban villageHouse no. (if any), mohalla, "near" landmarkTo a lane, then ask
Rural villageName, father's name, village, post officeTo the village; the postman knows the rest
Any of the above with DIGIPIN10-character codeTo a 4 m × 4 m square

What do landmark addresses actually cost?

A lot, mostly in time. The MIT and Delhivery researchers estimated that poor addressing costs India US$10-14 billion a year, about 0.5% of GDP, counting only logistics and financial services.

Where the money goes

The mechanism is simple. Without a reliable geocode, logistics firms sort parcels by PIN code, the only number in most addresses. But they found:

  • about 20-40% of written PIN codes are wrong;
  • the average PIN code covers 179 sq km, with roughly 1.35 lakh households and more than a lakh businesses and institutions;
  • the last mile, from local depot to doorstep, made up about 30-35% of delivery cost in India, compared with 10-12% in Western countries with structured addresses, despite much cheaper labour.

Drivers lose time calling customers and asking shopkeepers. When Delhivery moved from PIN-based to locality-based sorting, last-mile productivity rose 40-60%.

One address, three levels of precision

The precision problem is easy to see in one example from the paper. An address of the form "College Tilla, PO Agartala College, Agartala 799004" resolved on Google Maps to an area of about 76 sq km. Adding "near College Tilla Lake" narrowed it to about 3 sq km, and the last stretch still took a driver 15 to 45 minutes of asking around.

How much area an Indian address narrows down to
See the data as a table
Area (sq km)
Average PIN code179
Agartala address without landmark76
Same address with a landmark3
Average locality1.5

Source: Bhattacharya et al. (2018) and Rustogi et al. (2018), arXiv · For comparison, one DIGIPIN cell is about 0.000016 sq km (4 m × 4 m).

What is the right Addressing scheme for India?Kabir Rustogi, Santanu Bhattacharya, Margaret Church, Ramesh Raskar · arXiv (MIT Media Lab) · 2018

Drawing on Delhivery's delivery data and a review of systems such as eLoc, Zippr, Plus Codes and what3words, the authors describe India's addressing problem and compare fixes. Key numbers: about 80% of addresses are written relative to a landmark lying 50-1,500 m from the address; only about 30% are structured; the average locality covers about 1.5 sq km and the average PIN code 179 sq km; 20-30% of written PIN codes are wrong; the average gap between a written landmark and the actual doorstep is around 400 m; and the top 200 cities alone hold about 10 million points of interest. They argue for open, grid-based machine codes. The limit: the figures come largely from one logistics company's data, and the paper is a short position piece rather than a peer-reviewed study.

Read the paper (free)
Economic Impact of Discoverability of Localities and Addresses in IndiaSantanu Bhattacharya, Sai Sri Sathya, Kabir Rustogi, Ramesh Raskar · arXiv · 2018

The authors estimate the cost of poor addressing in India's logistics and financial services sectors, using delivery-operations data and industry figures. They find India's last-mile cost is around a third of total delivery cost against 10-12% in the West, that locality-based sorting raised Delhivery's last-mile productivity by 40-60%, and that poor addressing costs India US$10-14 billion a year. The limit: the paper's own figures vary slightly between sections (30% vs 35% last-mile share, 20-30% vs 30-40% wrong PIN codes across the two papers), and the national estimate is extrapolated from a few industries.

Read the paper (free)
Front of the India Post branch office at Sasatgre in West Garo Hills, Meghalaya, with its PIN code on the signboard
The branch post office at Sasatgre, West Garo Hills, PIN 794002. The PIN tells the system which delivery office handles a letter; the last few hundred metres still depend on local knowledge.Photo: This Photo was taken by Timothy A. Gonsalves. Feel free to use my photos, but please mention me as the author. I would, CC BY-SA 4.0, via Wikimedia Commons

Why doesn't the PIN code solve the problem?

Because a PIN code was designed to route mail to a delivery post office, not to find a door. India Post introduced the six-digit PIN in 1972:

  • the first digit is the region;
  • the next two narrow it to a sorting district;
  • the last three identify the delivery office.

Everything after that relies on the postman's memory.

PIN code area: average 179 sq km landmark sits 50–1,500 m from the door temple (landmark) about 400 m on average door, inside one DIGIPIN cell (about 4 m) Figures: Rustogi et al. 2018 (landmark gap, PIN area); India Post (DIGIPIN)
Why the rider still phones you. The PIN gets a parcel to the right depot, the landmark to the right neighbourhood, and the last few hundred metres are guesswork unless the address carries a precise point.

That design worked when the postman walked the same beat for years. It struggles with app-based delivery riders who change daily, ambulances that must reach an unfamiliar lane at night, and lenders checking that a borrower's address is real. I explain how the PIN system was built in how India's PIN code system works.

How does DIGIPIN give every place an address?

DIGIPIN, India Post's Digital Postal Index Number, divides the country into a grid of squares about 4 m by 4 m and gives each square a 10-character code derived from its latitude and longitude. The basics:

  • Built by the Department of Posts with IIT Hyderabad and ISRO's National Remote Sensing Centre;
  • released as open source under the Apache 2.0 licence;
  • rolled out in May 2025.

The maths is tidy. The grid covers longitudes 63.5°E to 99.5°E and latitudes 2.5°N to 38.5°N, a 36° by 36° box that includes India's maritime zone. Each level splits a square into 4 × 4 = 16 smaller squares, labelled with 16 symbols, and there are 10 levels. So the side of the final cell is:

\[ \frac{36^\circ}{4^{10}} = \frac{36^\circ}{1{,}048{,}576} \approx 3.43 \times 10^{-5\,\circ} \approx 3.8\ \text{m} \]

That is where the "about 4 m" figure comes from (one degree of latitude is about 111 km). Because the code is just encoded coordinates, India Post says it stores no personal data, and anyone can compute it offline. It does not replace your postal address; it adds a precise point to it. My guide to finding your PIN code and DIGIPIN shows how to get yours, and how to write an Indian address covers putting it on a form.

Level 1: 36° box, 16 cells Level 2: each cell split into 16 again ... 10 levels: 36° ÷ 4^10 ≈ 3.8 m per side One symbol per level → a 10-character code, shown as 3-4-3
How DIGIPIN works: every level divides a square into 16, and each of the 10 characters picks one of them. Ten picks take you from a 36-degree box to a square about 4 m across.

What is DHRUVA, and will it replace landmarks?

DHRUVA, the Digital Hub for Reference and Unique Virtual Address, is the policy framework India Post released on 30 May 2025 to build an "Address as a Service" layer on top of DIGIPIN. The idea is a consent-based address label, similar to a UPI handle, that you can share with a delivery app or a bank instead of typing out a long address.

In December 2025 the Department of Posts invited public comments on draft amendments to the Post Office Act, 2023 to give DHRUVA legal backing. When I checked, I found no source confirming that the amendments had been enacted. The direction is clear, though: India is not going to rename and renumber millions of lanes; it is going to put a precise, machine-readable point under the address people already use, much as UPI put a simple handle on top of bank account numbers.

A code tells a machine where to go; a landmark tells a person what to look for when they arrive.Why the Indian address of the next decade will carry both

Will landmarks disappear? I doubt it. A code tells a machine where to go; a landmark tells a person what to look for when they arrive. The best Indian address of the next decade will probably carry both: the DIGIPIN for the app, and "blue gate opposite the temple" for the rider standing in the lane. More everyday-India questions like this sit in Kyun, the Why India world.

Wooden-handled cancellation seals on a counter at the India Post office in Marathahalli, Bengaluru
Cancellation seals at the Marathahalli post office in Bengaluru. The postal network that once relied on the postman's memory now publishes an open digital grid.Photo: Subhashish Panigrahi, CC BY-SA 4.0, via Wikimedia Commons

Questions people ask

Why do Indians give directions using landmarks?

Because most streets are unnamed or unsigned and house numbers are not sequential, so a well-known landmark is the most reliable shared reference. Researchers estimate about 80% of Indian addresses are written relative to a landmark.

What efforts are being made to standardise Indian addresses?

India Post's DIGIPIN (2025) gives every 4 m square a 10-character code, and the DHRUVA framework proposes consent-based digital address labels, with draft amendments to the Post Office Act published for comment in December 2025.

Does DIGIPIN replace my PIN code or postal address?

No. India Post says your postal address stays the same. DIGIPIN adds a precise location code, and it stores no personal data because it is computed from latitude and longitude.

Why do delivery riders still call me even with a full address?

Because most addresses resolve only to a locality or a landmark, not a door. The MIT and Delhivery papers found the last mile costs about a third of total delivery cost in India, largely from time spent finding addresses.

Sources

  1. Rustogi et al.: What is the right Addressing scheme for India? (arXiv) (2018) (source for the 80% landmark, 30% structured, 179 sq km and 400 m figures)
  2. Bhattacharya et al.: Economic Impact of Discoverability of Localities and Addresses in India (arXiv) (2018) (source for the $10-14 billion cost and last-mile figures)
  3. India Post: DIGIPIN (2025-2026)
  4. India Post: DIGIPIN open-source repository (GitHub)
  5. The Week: What is DIGIPIN (2025)
  6. The Week (PTI): Postal Dept releases policy document on Digital Address DPI, DHRUVA (2025)
  7. Insights on India: DHRUVA draft amendments (2025)
  8. Wikipedia: National Level Addressing Grid (DIGIPIN)
  9. Wikipedia: Chandigarh (planning history, 2011 population)
  10. Kale, S., Google blog: How Plus Codes helped create new home addresses in India (2023; source for the Shelter Associates story and the 87,000 homes)
addresseslandmarksPIN codeDIGIPINDHRUVAIndia Postlast-mile deliveryChandigarh
Sapta

Sapta (Saptarshi Sarkar) is a mechanical engineer turned marketer in Bengaluru and a solo touring rider. He writes the seven Saptaverse blogs. Work with him or read more of Whyblog.

Keep reading in Whyblog

From the other blogs