POMI has had its day

Why construction cost information needs to move beyond free text and paper-era spreadsheets

POMI is often defended because it is familiar. Clients recognise it, consultants have old templates built around it and quantity surveyors know how to prepare a bill using it. Familiarity, however, is not the same as suitability. The Principles of Measurement (International) were first issued by RICS in 1979 and archived in March 2023. This does not prevent parties from using POMI where a contract expressly adopts it, but it does confirm that it is no longer a current RICS publication.

POMI was useful for its time. It provided broad international measurement principles where no suitable local method existed, or where project teams from different countries needed a common starting point. It was a practical fallback rather than a detailed system for storing construction cost information. Even the foreword added in 2003 noted that many countries had developed their own methods of measurement. Of the 32 countries that replied to an RICS survey, 14 already had a national method.

The world for which POMI was written has changed. The larger question is whether the traditional bill of quantities, and the spreadsheet layout that usually supports it, is also starting to show its age.

The bill is still treated as a document

The usual bill layout has changed very little: item number, description, unit, quantity, rate and amount. It is familiar to quantity surveyors, easy to print and simple to issue as a tender document, but it is far less useful as a database.

Most of the meaning is placed in one free-text description or left elsewhere in the document. An item may depend on the bill heading, pricing preamble, specification, drawing reference or an earlier description using phrases such as “ditto”, “as before” or “complete”. A quantity surveyor can read up and down the page to find that context, but a computer generally receives one row in a spreadsheet without any clear record of where the missing information sits.

Consider four descriptions: “200mm concrete block wall”, “Walling in 200 thk HCB”, “200mm CMU including mortar” and “Internal blockwork, 200mm thick, complete”. These may describe the same work, but they may not. One may include mortar, ties, reinforcement, cutting and scaffolding, while another may depend on a general preamble. “HCB” may mean hollow concrete block to one surveyor, while another office may use a different abbreviation. The word “complete” can hide a remarkable amount of uncertainty.

A good bill description is a pricing instruction, not merely a sentence about the work. It should clearly state the action, material, size, type, location, finish, standard and relevant inclusions. The difficulty is that the traditional bill tries to place all those fields into one line of prose.

Flexibility produces inconsistent information

POMI does not stop a skilled surveyor from preparing a clear bill, but a reliable company process should not depend on every surveyor making the same personal choices. POMI permits wide choice over bill format, classification and the content of descriptions. It also allows descriptions to refer to drawings, specifications and other documents rather than repeating the information within the item itself.

That freedom helped POMI gain international use, but it also allows different surveyors, teams and offices to describe the same work in very different ways. This becomes expensive when quantities later need to be grouped, compared or used for procurement.

On one programme involving more than 80 buildings, a mixed team prepared bills covering different uses, disciplines and building sizes. When the quantities had to be brought together for procurement packaging, inconsistencies in the descriptions made reliable grouping almost impossible. More than 100 hours were then spent rewriting descriptions, even though the measurement had already been completed. The additional effort was required because the output could not be used properly as data.

Different surveyors use different spelling, syntax, abbreviations, word order and units. One starts with the material, another with the location. One includes performance requirements in the description, while another assumes they can be found in the specification. “m2”, “m²” and “sqm” may all appear in the same bill, as may “nr”, “no.” and “each”, or “reinforced concrete”, “RC” and “R.C.” These differences seem minor when reading one bill, but across thousands of items and many projects they prevent direct grouping, comparison and analysis.

Then comes the Excel archaeology

The problem does not end with the descriptions. Before the information can be analysed, someone usually has to spend hours making the spreadsheet look consistent. Columns need to be realigned, merged cells removed, hidden rows found, units corrected, blank lines deleted and formulas traced. Then comes the familiar moment when Excel asks whether you would like to update links to an external file.

The source file was created in 2003.

The spreadsheet being used today was first prepared in 2001, copied for another project in 2007, altered again in 2014 and carried faithfully into 2026. Nobody knows who created the original link and nobody wants to remove it in case the whole bill collapses.

I have started to think of this as bill-sheet debt. It is the gradual build-up of copied formatting, merged cells, hidden formulas, broken links, unused tabs, old macros and conditional formatting rules that no longer serve any clear purpose. Every project adds another layer until the workbook contains hundreds of thousands of formatted but empty cells. It can take 15 minutes to open, several minutes to save and long enough to recalculate that the user begins to wonder whether the computer has frozen.

This is apparently the price of preserving a cell border selected by someone during the early years of the Iraq war.

The result is not merely an irritating spreadsheet. It is a fragile commercial record. Files become difficult to review, difficult to move between systems and risky to change. A small alteration may break formulas several tabs away, while links may point to private drives, former employees’ folders or files that no longer exist. The traditional bill format has encouraged this because the spreadsheet is treated as both the working database and the final presentation document. It is trying to do two jobs and often performs neither particularly well.

Poor descriptions do not become useful data through volume

Construction firms now hold more historic cost information than ever, yet the value of that information is often overstated. A database may contain millions of measured items, but it is of limited use if the same work cannot be identified across projects.

Before rates can be compared, someone must establish whether the items cover the same scope, use the same unit and relate to similar project conditions. Suppose three projects contain an item called “200mm blockwork”. On the first project, the rate includes blocks, mortar and ties. On the second, it includes reinforcement and scaffolding. On the third, it includes plaster as well. An average calculated from those rates may look precise, but it says very little.

The same issue affects procurement. A client may have enough combined demand across several buildings to obtain better prices for doors, tiles, sanitary fittings or mechanical equipment, but that buying power is lost where similar items cannot be found and grouped without a manual review. Poor descriptions also create gaps and overlaps between packages. Work may appear in two packages because the wording differs, or be missed because each package assumes another includes it. The saving made by producing a quick bill can be lost many times over through tender queries, qualifications, variations and final account disputes.

POMI is also creating a training problem

The continued use of POMI is not only producing poor information. It is also allowing weak measurement practice to continue. Its broad rules make it too easy for inexperienced quantity surveyors to rely on old bills, copied descriptions and inherited office templates. A previous item can be changed slightly and reused without the surveyor first considering the construction method, measurement rules, package boundaries or pricing risks.

The result may look like a bill of quantities, but the reasoning behind the take-off can be weak. A proper take-off requires more than extracting dimensions from drawings or clicking objects in a model. The quantity surveyor must understand what is being built, how it will be constructed, what must be measured separately, what is included in the rate and how the item will be used by the tenderer and client.

In too many cases, surveyors are producing quantities without being able to explain why the work has been measured in a certain way. More concerning still, some struggle to follow the written rules when more detailed methods such as NRM 2 or CESMM4 are used. Items are placed in the wrong class, work that should be measured separately is grouped together, deemed inclusions are repeated as separate items, units do not match the stated class and descriptions omit the features required by the method.

The problem is not that NRM 2 or CESMM4 lack rules. The rules exist. The problem is that many surveyors have not been properly taught how to read, apply and check them.

The profession risks creating people who know how to operate CostX, measure from a PDF or extract quantities from a model, but cannot explain the basis behind the output. Software can make a good surveyor faster, but it can also allow poor practice to be repeated on a much larger scale. A person can click hundreds of objects, generate thousands of lines and produce a document that appears complete, without the take-off properly following the selected method of measurement.

The real test should not be whether the software produced a quantity. The surveyor should be able to explain why the item was measured, why that unit was selected, which rule applies, what is included in the rate, what must be measured separately and which information must appear in the description. If those questions cannot be answered, the take-off is not complete, regardless of how neat the workbook looks.

Measurement knowledge is being lost

This training problem is becoming more serious as experienced quantity surveyors retire, leave the profession or move into senior management. Much of their knowledge was gained through hand measurement, close review of drawings, site experience and detailed checking by senior staff. Junior surveyors were once expected to prepare take-off sheets, apply the method of measurement and have their work reviewed line by line. That process was not always fast, but it taught people how construction work was put together and why measurement rules existed.

In many consultancies, that knowledge is no longer being passed down in the same way. Senior staff are under pressure to manage clients, fees, programmes and commercial risks, while junior staff are expected to produce quantities quickly. Software is then treated as a replacement for training rather than a tool to be used after training.

Old bills become the unofficial teaching material. A junior surveyor copies what was done on the last project, the next surveyor copies that bill again and any error in the original measurement, wording or classification is passed from one project to the next. After several years, nobody remembers why the item was written that way. It has become office practice through repetition rather than because it is correct.

POMI gives poorly managed organisations too much room to avoid this issue. Its simplicity can be mistaken for ease of use, when in practice it places more weight on the judgement of the person preparing the bill. A more detailed method such as NRM 2 or CESMM4 does not remove the need for judgement, but it provides a clearer structure within which that judgement can be taught, reviewed and tested.

Consultancies should not allow measurement knowledge to disappear into old templates and unexplained office habits. Senior quantity surveyors should be expected to teach the reasoning behind the take-off, review how items have been formed and explain why work has been separated or grouped. AI can assist with checking descriptions and finding missing information, but it cannot replace basic professional education.

We already have better structures available

No single standard answers every information need. Measurement methods, cost plans, classification systems, specifications and international reporting standards each serve different purposes.

For building works, NRM 2 provides detailed rules for measurement and description, while NRM 1 provides the elemental structure used for order-of-cost estimates and cost planning. It allows building costs to be reviewed by elements such as the substructure, frame, external walls, internal finishes and services. For civil engineering works, CESMM4 provides detailed classes and rules for roads, drainage, earthworks, utilities, structures and other civil or trade-based work.

ICMS 3 provides a high-level international structure for reporting costs and carbon, allowing projects to be compared across sectors, regions and countries. Uniclass can classify complexes, entities, activities, spaces, systems, products and other forms of construction information throughout the life of an asset. In North America, OmniClass, MasterFormat and UniFormat provide related methods of arranging construction information, while CSI standards are widely used for specifications, work results and elemental analysis. NBS provides specification and product-information systems that use structured classification, particularly Uniclass.

Construction data dictionaries can go further by defining common names, properties, units and permitted values. Instead of allowing every surveyor to invent a new way of describing a fire door, a data dictionary can state which fields must be completed and which values may be used.

These systems should not be treated as competing answers. A measured item can carry several codes at the same time, including a detailed NRM 2 or CESMM4 measurement code, an NRM 1 element, an ICMS 3 cost group, a Uniclass or OmniClass classification, a specification reference, a procurement package code and a location or asset reference.

Crosswalking between these structures provides several views of the same information. The detailed measurement view supports tender pricing and valuation, the elemental view supports cost planning and building benchmarking, the civil engineering class view supports analysis of roads and utilities, the system and product view supports design and specification, the package view supports procurement and the ICMS view supports international comparison.

The same quantity does not need to be measured several times. It needs to be stored once, with enough structure to allow it to be viewed in several ways. A client could then compare structural frames across countries, review façade systems by building type, group civil engineering quantities by CESMM4 class and report the same projects under ICMS 3. That is far more useful than a folder containing hundreds of unrelated bills.

The bill should be an output, not the database

The industry should reconsider whether the familiar bill layout should remain the main source of cost information. A modern measured item should not consist only of a description, unit and quantity. It should be stored as a record containing a permanent item identifier, work result, system, product, material, size, location, finish, performance requirements, inclusions, exclusions, unit, quantity, drawing source, specification reference, building, zone, package, measurement code, classification code, cost-plan element and international reporting code.

The bill can still be produced for tendering and contract purposes, but it would become one report generated from the information rather than the place where all the information is stored. The same source could also produce a procurement summary, elemental cost plan, ICMS report, carbon schedule, package breakdown and benchmarking file.

This does not require the removal of bills of quantities. A well-prepared bill remains a useful pricing and contract document. The change concerns how the information is created and retained. The bill should become a human-readable view of structured cost information, rather than the place where all the information is trapped.

Where AI can help

AI is well suited to identifying patterns in inconsistent descriptions. It can recognise that “HCB”, “hollow block” and “hollow concrete masonry unit” may refer to the same material, standardise units, correct spelling, separate dimensions from text and suggest likely classification codes.

It can also split an old description into fields such as work result, material, thickness, location, unit, mortar inclusion, strength, fire rating and reinforcement. Where the original description does not contain enough information, the output should state “not stated” or “unknown” rather than filling the gap with an assumption.

AI can compare descriptions across offices, identify likely duplicates, flag missing information and suggest mappings between POMI, NRM 2, CESMM4, NRM 1, ICMS 3, Uniclass and other systems. It can also check whether the unit appears to match the selected class, whether required description features are missing and whether a line has been placed under an unsuitable heading.

It cannot replace knowledge of measurement rules. It may flag that an item appears inconsistent with CESMM4 or NRM 2, but a competent quantity surveyor must decide how the rule applies to the design and contract.

The best use of AI is at the time of measurement. A surveyor entering an item into CostX or another take-off system could receive an immediate warning that a required field is missing, an unapproved abbreviation has been used, the selected unit does not match the class or the description does not meet the stated measurement rule. AI can help repair historic records, but it should not become an excuse to continue producing poor records or avoid training staff.

Consultancies need to step up

When a client requests POMI because it was used on earlier projects, the consultant should not simply accept the instruction without advice. The consultant should explain that POMI may produce a contractually usable bill, but it will not by itself produce a consistent portfolio cost dataset. The client should also be told about the later cost of cleaning descriptions, mapping items, correcting spreadsheets and resolving gaps between packages.

Pushing back does not have to mean refusing to use POMI. Where it remains a client requirement, it can continue as the contractual measurement layer while the consultant also provides a controlled item library, structured data schedule, classification codes and mapping into an agreed reporting structure.

A client working across several countries should be able to compare similar projects without first paying a team to translate every bill. Local measurement rules may differ, but the reporting fields, classification rules and data dictionary can remain consistent. That is real added value from a consultant because it gives the client cost information that can be used across projects, offices and countries, rather than a collection of isolated tender documents.

This will not happen simply because better standards and software exist. Consultancies must educate their staff properly. Quantity surveyors need formal training in measurement principles, not only software instructions. They must be taught how to apply NRM 2, CESMM4 and other relevant methods, including the rules on classification, units, deemed inclusions, item coverage and description requirements. A training session showing someone which CostX button to press is not measurement training.

Junior staff should prepare take-offs that are reviewed by experienced quantity surveyors, and those reviews should test the measurement basis rather than merely check arithmetic and spelling. The profession must restore the link between measurement, construction knowledge and commercial judgement.

Quantity surveyors must also understand that item descriptions are no longer written only for tenderers. They may later be read by procurement teams, cost planners, data analysts, asset managers and AI systems. Staff should therefore be trained in approved description syntax, classification systems, data dictionaries and the use of structured fields. Lead quantity surveyors should be responsible for enforcing those rules, rather than treating them as an optional formatting preference.

Consultancies must also educate their clients. A client may request POMI because it appears simple, familiar or cheaper, but the consultant should explain what the client loses by making that choice and what could be gained through a consistent international approach. That discussion should take place before the bills are prepared, not after hundreds of hours have been spent cleaning them.

Consultants are paid for judgement, not merely for following an old template. Where a client requests a method that will produce poor information, the consultant should say so, explain the risks and propose a better route.

The larger question

POMI has outlived the problem it was written to solve. It provided broad international measurement rules at a time when few other common options were available. Today, the industry has detailed methods for building and civil engineering work, elemental cost structures, international reporting standards, specification systems, classification codes and construction data dictionaries.

Yet much of this information is still compressed into free-text descriptions and passed around in spreadsheets designed before some of the people using them were born.

The question is no longer only whether POMI should be replaced. It is whether the traditional bill of quantities should remain the main way construction cost information is stored at all.

The future bill may look familiar when printed, but behind it should sit structured, coded and machine-readable cost information that supports pricing, procurement, cost planning, international benchmarking and future asset decisions.

Clients should expect that, and consultancies should be teaching their staff how to provide it, checking that they can apply the relevant measurement rules and advising clients why it matters.