Skip to content
Tuesday, 8 September 2026 · LondonENع
Rayan Azhari.Sustainability · Energy · Carbon · Built EnvironmentOccasional detours into philosophy, religion or programming, wherever curiosity leads
Property, Buildings & Sustainable Real Estate

Article 14: Vertical Postcodes: How the UK Address System Handles Stacked Property

A series mining the PhD thesis on London and UK office buildings (Azhari, 2025). Key takeaway. The UK postcode is a flat two-dimensional locator carrying three-dimensional cities on its back, and UPRNs plus the AddressBase Premium hierarchy fill the gap that lets digital property services work vertically.

Rayan AzhariChartered Environmentalist, MISEP · 13 min read
Aerial view of Canary Wharf office towers beside the River Thames under a cloudy sky, the title card for Article 14 on vertical postcodes and how the UK address system handles stacked property.

A series mining the PhD thesis "London and UK Office Buildings: Investigating Energy Use and Landlord-Tenant Influences" (Azhari, 2025).

Key takeaway. The UK postcode is a two-dimensional locator carrying three-dimensional cities on its back. UPRNs and the AddressBase Premium hierarchy fill the gap, and they are why every digital service in British property (delivery, electoral registration, tax assessment, stock-level energy work) can function in vertical urban environments.

Figure

One postcode at the door, fifty UPRNs going up

A single Central London postcode unit resolves to a parent UPRN and dozens of child UPRNs through the AddressBase Premium parent-child hierarchy.

EC2A 1AA

one postcode unit

One street address

single postal entry point

AddressBase Premium

parent-child UPRN hierarchy

53 UPRNs

1 parent shell, 50 tenant children, 2 retail children

Illustrative. AddressBase Premium assigns each addressable unit its own UPRN, organised in a parent-child hierarchy: a parent UPRN for the building shell, child UPRNs for floor 9 to floor 22 office tenancies (50+ in total) and ground-floor retail.

Source: Author's analysis; Ordnance Survey AddressBase Premium / GeoPlace UPRN

The collision: one postcode, many addresses

Walk into the lobby of a typical Central London office tower. The building has one street address, one postal entry point and one Royal Mail postcode unit at the door. Inside, twenty-five floors hold somewhere between five and a hundred distinct tenancies. The tower runs forty or fifty different electricity meters, more gas meters, and a small fleet of EPCs lodged against different parts of the demise. The lift goes up. The data does not.

A UK postcode unit (the seven-character string like EC2A 1XX) typically maps to between fifteen and thirty addresses. In a residential area those addresses sit at horizontally distinct front doors on the same street. In a Central London commercial setting, the same postcode unit can sit above hundreds of legally and operationally distinct property units stacked vertically. The flat identifier cannot describe the stacked reality without help.

The help comes in the form of a separate identifier system layered on top of the postcode: the Unique Property Reference Number (UPRN) and the AddressBase Premium hierarchy that organises them. This article is about how that system works, why it matters, and what an analyst who cares about British property has to know.

Anatomy of a UK postcode

The UK postcode has been with us since the 1960s. Its job was to help the Royal Mail sort and route letters efficiently across a growing nation. It does that job well. It was not designed to identify property, and the problems start when it gets used as if it were.

A postcode has four nested levels. The postcode area is one or two letters (EC, SW, M, BT). The postcode district adds one or two digits to the area (EC1, SW1, M1, BT9). The postcode sector adds a space and another digit (EC1A 1, SW1A 1, M1 1). The postcode unit completes the string with two more letters (EC1A 1BB). A postcode unit is the smallest formal postcode element and typically covers fifteen to thirty addresses.

What postcodes get wrong, for the purposes of identifying property, is twofold. First, postcodes are not stable. The Royal Mail reorganises postcode coverage every few years to reflect changes in delivery volumes, the closure or opening of sorting offices and demographic shifts. A property that sat in EC2A 1AB in 2019 might sit in EC2A 1AD by 2025 without anything having moved. For day-to-day mail this is fine. For longitudinal property research it is poison.

Second, postcodes are flat. They label horizontal area. They cannot distinguish between the ground-floor café and the twenty-fifth-floor trading desk in the same tower. Both share the same postcode unit. Both usually share the same street address. Anything that depends on telling those two units apart needs a different identifier.

UPRNs: the spine of the British property graph

The Unique Property Reference Number is a twelve-digit integer assigned by GeoPlace, a public-sector joint venture between the Local Government Association and Ordnance Survey. Every land or property unit in Great Britain that can receive an address gets a UPRN. A piece of land, a building, a floor of a building, a flat, a shop unit, a public toilet, a substation: all candidate UPRN holders.

There are several things that make UPRNs work where postcodes do not. UPRNs are stable. Once assigned, a UPRN persists for the life of the property unit. When the property is demolished or merged, the UPRN is marked retired rather than reused.

UPRNs are granular. Each addressable unit gets its own. A twenty-five-storey office tower with fifty tenancies generates fifty-plus UPRNs (one per addressable unit) plus a parent UPRN for the building shell. The flat-postcode problem is dissolved.

UPRNs are open. In July 2020 the UK government released the UPRN as open data under the Open Government Licence. Before that release, working with UPRNs required a commercial licence with Ordnance Survey, which significantly limited academic and civil-society use. The open release is one of the quieter and more consequential data-policy decisions of the past decade.

UPRNs have a cousin. The Topographic Identifier (TOID) is the equivalent identifier for OS Master Map topographic features (buildings, parcels, roof outlines). Where UPRN identifies an addressable unit, TOID identifies a cartographic feature. The two link in AddressBase Premium.

AddressBase Premium and the Basic Land and Property Unit

UPRNs do not sit in a flat list. They are organised in a hierarchy maintained by Ordnance Survey in a product called AddressBase Premium. Three concepts matter.

The Basic Land and Property Unit (BLPU) is the atomic property concept. One BLPU is one addressable unit (one UPRN). It has coordinates, a parent UPRN where applicable, a logical status (proposed, approved, current, retired) and an organisational link to the local authority that maintains the underlying Local Land and Property Gazetteer (LLPG).

The Land and Property Identifier (LPI) is a name or address attached to a BLPU. One BLPU can have multiple LPIs: the postal address, an internal building name, a historical address that has been superseded. The LPIs are alternative human-readable labels for the same machine-readable UPRN.

Parent and child UPRNs are the relational glue. A multi-let office tower has a parent UPRN for the building shell and child UPRNs for each let unit. A multi-flat residential block has a parent UPRN for the block and child UPRNs for each flat. The parent-child relationship is what lets a query of the form "everything in this tower" return all the addressable units without needing to know each one in advance.

Lifecycle states matter for longitudinal work. A UPRN with logical status "retired" used to be live and has since been demolished, merged or otherwise removed. Analysts joining datasets across years have to handle these explicitly. Treating them as missing data will distort time series.

The vertical problem, solved

A worked example. A twenty-five-storey Central London office tower, completed in 2005, with a single street address, one postcode unit, fifty tenant let units across floors three to twenty-five, retail at ground and first, and a sub-basement plant room. In the postcode system it is one identifier. In AddressBase Premium it is fifty-three UPRNs (one parent, fifty tenant children, two retail children) all sharing the same street address but each carrying its own coordinates, lifecycle state and parent reference.

Drop in a mixed-use building of the same era: ground-floor retail, floors two to fifteen office, floors sixteen to twenty-five residential, one street address, one postcode. The number of UPRNs roughly doubles because the residential floors subdivide into flats. The same hierarchical machinery resolves the question of what is what.

Compare to other countries. United States street addressing is famously variable. Most jurisdictions do not have an equivalent stable identifier. Address strings are the primary key, with all the inconsistency that implies. France Base Adresse Nationale gives every address a stable identifier and an Open Database Licence, closer in spirit to UPRN, but does not have the same depth of hierarchy. Germany Lagebezeichnung uses a different system based on the cadastre. The UK system is unusually well equipped for digital uses, in part because of decades of investment in AddressBase Premium and in part because of the 2020 open release.

Standards and the politics of addressing

The UPRN system did not appear by accident. It sits on a stack of standards and a slow politics of inter-agency cooperation.

BS 7666 is the British Standard for spatial data on residential and non-residential property. It sets the rules for LLPG, BLPU and LPI and is the technical reference for everything described above. The standard is in its fourth revision and runs to several hundred pages across its parts.

INSPIRE compliance, the EU directive that pushed member states to publish standardised spatial datasets, is one reason AddressBase Premium exists in its current form. Even after Brexit, the UK has retained the INSPIRE-derived datasets, including the Land Registry INSPIRE Index Polygons (covered in the next article in this series).

GeoPlace is the central custodian: a joint venture between the Local Government Association and Ordnance Survey that allocates UPRNs, maintains the central GeoPlace dataset, and coordinates with the 348 local authorities in England, Wales and Scotland that maintain their LLPGs.

Northern Ireland sits on a parallel system called Pointer, maintained by Land and Property Services. Pointer assigns its own UPRN-equivalent identifier and has been gradually aligning with the GB approach. For UK-wide analytical work, the Pointer to LLPG difference is a real but manageable nuisance.

Why this matters for energy work

The 3DStock model that sits behind every empirical article in this series uses UPRN as its primary join key. Without UPRN, the model could not match BEIS gas and electricity meter data to EPCs, to VOA non-domestic rating list entries, to OS topographic footprints. With UPRN, those joins become tractable.

The seven-data-source flow described in Article 4 is, in essence, a series of UPRN joins. AddressBase Premium provides the spine. Each data source attaches additional attributes to UPRNs that already exist in the address graph. The work of constructing 3DStock is the work of cleaning those joins, handling lifecycle state changes, reconciling parent-child relationships and dealing with the edge cases where one of the source datasets has been slow to update.

Operational rating policy is the other place where UPRN matters. NABERS-UK, the operational rating scheme covered in Article 9, has to decide whether a rating belongs to a whole-building UPRN (the parent), to a base-building UPRN (a notional subset of the parent), or to individual tenancy UPRNs (the children). The UK addressing system gives policy-makers the technical option to do any of those. The decision is then a regulatory one, not a data one.

The vertical problem reappears with force whenever a policy intervention has to operate on a building-level abstraction rather than the address-level reality. The 1,000 square metre threshold for the proposed operational rating scheme (see Article 1) sounds simple at the building level. At the UPRN level it requires a rule for how to aggregate child UPRNs up to the parent, and a rule for what to do when the aggregate sits below 1,000 square metres but one of the children would otherwise be exempt. These are addressing-system questions disguised as policy questions.

Practical advice for analysts

Five rules of thumb for working with UK property data.

First, use UPRN as the primary join key. Never use postcode strings or address strings except as last-resort fallbacks. Postcode and address change. UPRN does not.

Second, decide explicitly which level of the UPRN hierarchy you are joining at. Building, let unit, flat, plant room. Document the choice. Different policy questions sit at different levels.

Third, handle lifecycle states. A retired UPRN is not the same as a missing UPRN. Treat retirement events as data rather than as noise. They often mark the events you actually want to study (demolition, merger, change of use).

Fourth, cross-check against the VOA non-domestic rating list for commercial premises and against Council Tax records for residential. These are the live sources of property activity in the UK and they update faster than AddressBase Premium in some local authorities.

Fifth, build a UPRN-level data dictionary for any non-trivial project. The variations in how local authorities populate the LLPG are real, and the only way to keep your analysis defensible is to know which authorities use which conventions for which property types.

The next article in the series stays in the methodology mini-arc but moves from addresses to buildings. If UPRN tells you what counts as one address, four other public datasets tell you what counts as one building, and they do not always agree.

Limitations

The article relies on the documented behaviour of UPRN, BLPU and AddressBase Premium. It does not have access to internal GeoPlace operational data on update cadences across local authorities. The Northern Ireland Pointer dataset is acknowledged but not treated in depth. AddressBase Premium remains a licensed product for most uses (beyond the open UPRN release). Practice varies across local authorities and the article does not enumerate the variation. Some technical nuance (TOID, OSGB grid linkage, sub-BLPU concepts) is referenced but not exhaustively treated.

References

About this series

This article is part of a fifteen-piece series adapting the 2025 PhD thesis "London and UK Office Buildings: Investigating Energy Use and Landlord-Tenant Influences" (Azhari, 2025) for a mixed academic and industry readership. The empirical findings draw on the 3DStock model of 6,038 office Self-Contained Units in Greater London with metered energy data for 2017, supplied by BEIS under a data-sharing agreement, alongside the Better Buildings Partnership Real Estate Environmental Benchmark. The qualitative findings draw on semi-structured interviews with seven major UK property organisations, conducted during the 2021 lockdown. Interviewees and their organisations are anonymised by role and organisation type. Please cite the original thesis for academic use.

Author. Rayan Azhari completed his PhD at the UCL Bartlett School of Environment, Energy and Resources in 2025, supervised by Paul Ruyssevelt and Kathryn Janda. The research was supported by the EPSRC Centre for Doctoral Training in Energy Demand (LoLo) and UK Research and Innovation through the Centre for Research into Energy Demand Solutions.

Other articles in the series. Article 1 The 30/85/89 Problem; Article 2 Why EPCs Do Not Tell You How Much Energy a Building Uses; Article 3 Eighteen Per Cent; Article 4 Mapping the Stock; Article 5 Height, Age and the Fuel Question; Article 6 The Split-Incentive Problem; Article 7 Green Leases and Service Charges; Article 8 From 38 to 73 Per Cent Energy Savings; Article 9 NABERS for Britain; Article 10 Time to Retire ECG-19; Article 11 Can London Speak for England and Wales; Article 12 The Hybrid-Work Footprint; Article 13 Why I Used Linear Regression Over Random Forest; Article 14 Vertical Postcodes; Article 15 What Is a Building?

Further reading

Office energy, part 14 of 15

Related posts

Article 12: The Hybrid-Work Footprint: What COVID-19 Did to Office Energy

A series mining the PhD thesis on London and UK office buildings (Azhari, 2025). Key takeaway. Hybrid work did not deliver the linear energy saving the headlines implied: some demand shrinks when occupancy drops, base-building load holds steady, and fresh-air ventilation actually grows.

· 8 min

Essays in your inbox

New writing on Syria, sustainability and finance, a few times a month.

Unsubscribe anytime. Read by 4,200+ professionals.