
Almost nobody arrives at this subject with a clean question. They arrive with three answers. One page said the fixture should relate to the width of the table. Another said it should relate to the dimensions of the room. A third said it should relate to the height of the ceiling and gave a rule for adding to the size as the ceiling rises. All three were presented with the same confidence, and none of them explained when it applies. The useful conclusion is not that two of them are wrong. It is that the question you should be asking is different: not “what is the rule”, but “which rule was written for a room like mine, and how do I check it.”
This article does not add a fourth method to the pile. There is no shortage of methods, and one more would simply make the confusion worse. What it does instead is take the conflict seriously: it sets out the four ways that sizing guidance can disagree, explains why the disagreements arise in the first place, gives you a four-step audit you can run before anything is ordered, and finishes with what to put in front of a supplier so that the conversation is about your room rather than about somebody’s published rule.
The audience for this is anyone who has to commit to a fixture in a real space: a designer producing a schedule, a homeowner who has decided to get this right once, an architect checking somebody else’s schedule, a project manager reconciling a drawing with a quotation, and the sales side of a lighting business who has to explain why the numbers a client brings in do not match what the room needs. None of those people needs another formula. All of them need to be able to tell which published rule they are looking at, and what it left out.
Two related articles are the natural companions. Our measured survey of fixture sizes recorded across real installed projects is the data side of this question, and the guide to sizing in rooms with high or double-height ceilings covers the case where the vertical dimension dominates. Both give you numbers; this article is about how to handle them, and it is written deliberately without any, because the point of the exercise is that the numbers are inputs that belong to your room rather than constants that belong to the internet.
- 01Why Three Sources Gave You Three Answers
- 02The Four Ways Fixture Sizing Advice Can Disagree
- 03Conflict One: The Basis Changed
- 04Conflict Two: The Increment Changed
- 05Conflict Three: The Count and Spacing Method Changed
- 06Conflict Four: The Unit and the Conversion Changed
- 07Where These Conflicts Come From
- 08A Four-Step Audit You Can Run Before You Order
- 09Signing It Off: Measurement, Structure and Local Requirements
- 10What to Send a Supplier Instead of a Formula
Why Three Sources Gave You Three Answers
Start with the observation that makes the rest of this article useful: almost every published sizing rule is internally coherent. Take the method on its own terms, follow it properly, and it produces an answer that is defensible for the room its author had in mind. The problem is that the author had a room in mind, and that room was not described in the article. What was published was the arithmetic; what was left out was the situation.
So the three answers you collected are not three attempts at one calculation, with two of them wrong. They are three different calculations, each of which answers a slightly different question. One of them answers “how large should the fixture be in relation to the table it lights”. Another answers “how large should the fixture be in relation to the volume of the room”. The third answers “how much larger should the fixture be than it would otherwise be, given the height of this ceiling”. All three questions are reasonable. They are simply not the same question, and no page told you which one it was answering.
This is why comparing the outputs of two rules is the wrong move, and why people who do it end up more confused than when they started. If rule A is a table-based method and rule B is a room-based method, the difference between their outputs is not a discrepancy to be resolved. It is the difference between two design intentions: a fixture that behaves as a piece of furniture over a table, and a fixture that behaves as an object placed in a volume. Both exist in real interiors, and the choice between them is a design decision that you are entitled to make deliberately.
The practical consequence is a change of method. Instead of collecting formulas and looking for the one that seems most authoritative, you do three things in order. You identify the basis each rule is using, so you know which question you are being answered. You check whether the conditions each rule assumes are present in your room. Then you choose the rule whose assumptions match the room, and you record the choice along with the measurements it will be applied to. The rest of this article is that procedure, taken one piece at a time.
The Four Ways Fixture Sizing Advice Can Disagree
In practice, the disagreements cluster into four kinds, and each kind has a different remedy. Recognising the kind is most of the work, because it tells you immediately whether the two rules you are comparing can be compared at all.
The first kind is a difference of basis. The rule measures the fixture against something: the table below it, the floor area of the room, or the height of the ceiling. Change the basis and you change the answer, without either rule being faulty. The second kind is a difference of increment: both rules accept the same basis and then add to the fixture as some condition increases, but they add different amounts or add at different moments. This is the kind that most often appears as a small, annoying discrepancy, because two rules of this kind look almost identical and produce answers that differ by a margin you cannot explain.
The third kind is a difference of arithmetic structure rather than of measurement: how a number of fixtures is divided across a length, and where the gaps are placed. Rules of this kind can agree on the total extent and still disagree on the arrangement, which matters because the arrangement is what a person actually sees. The fourth kind is a difference of unit and conversion, which sounds trivial and is responsible for a surprising share of the confusion, because the two systems of measurement in common use do not convert cleanly and because a formula that was rounded after conversion is no longer the formula it claims to be.
Two further distinctions are worth holding while you read the next four sections. The first is between a rule that produces a size and a rule that produces a limit: some published methods are answering “approximately how large”, and others are answering “no larger than” or “no smaller than”, and a maximum used as a target produces a fixture at the edge of what the room can carry. The second is between a rule that assumes the fixture is the sole light in the space and a rule that assumes it is one layer among several, because the second kind is quietly answering a question about visual weight rather than about task illumination.
Conflict One: The Basis Changed
Take the basis first, because it explains more of the confusion than any other single cause. A method based on the table treats the fixture as belonging to the table: the relationship that matters is between the size of the object above and the size of the surface below, and the room enters the calculation only as a constraint. A method based on the room treats the fixture as belonging to the space: what matters is whether the object holds its own in the volume, and the table is a piece of furniture that happens to be underneath it. A method based on ceiling height treats the fixture as a function of vertical space, which is the basis that produces the rule people quote most often in tall rooms.
Each basis produces a fixture with a different character in the room, and the difference is visible. A table-based answer tends to produce something that reads as a considered centrepiece over the table, generous enough to relate to the surface but not attempting to fill the volume. A room-based answer tends to produce something larger, because rooms are bigger than tables and the fixture is being asked to register at the scale of the walls. A ceiling-height answer produces something that either grows with the vertical space or does not, and in a tall room the growth is what keeps the fixture from looking stranded.
Because the three bases are answering three different design questions, the correct move is not to average them but to decide which one you are actually specifying for. If the fixture is the principal light over a dining table and the room has other layers doing the ambient work, the table is your basis. If the fixture is the single significant element in a tall or open volume and will be seen from across the space, the room is a better basis, and you should expect the answer to be larger. If the room has an unusually high ceiling and the fixture will be viewed from directly underneath as well as from a distance, height becomes a modifier on top of whichever basis you chose, rather than a basis of its own.
One more distinction inside this first kind of conflict is worth naming, because it changes the arithmetic: the difference between the footprint of a surface and its longest edge. Two methods can both be described as table-based and still disagree, because one scales the fixture to the width of the table and the other scales it to the diagonal reached across the corners. Neither is dishonest, and the second is not an attempt to inflate the fixture: in a long, narrow room the diagonal is often the more useful measure, because it is closer to what the eye reads from the doorway.
Conflict Two: The Increment Changed
The second kind of conflict appears when two rules agree on the basis and disagree on how the fixture should grow. This is the most frustrating kind, because the two rules can look like the same rule with a different number in it, and the numbers are exactly what is not transferable. The growth may be attached to the height of the ceiling, to the size of the room, or to the number of floors the fixture is seen from, and each attachment produces a different progression.
In tall spaces, the common device is an addition for each additional unit of height above a threshold: a fixture that would suit the room at the base height grows a little for every step above it. Two published rules of this type can differ in the threshold, in the size of the step, or in whether the step is applied to the width of the fixture or to its drop. The third of those is the one that catches people out, because a rule that grows the drop produces a taller, narrower object, and a rule that grows the width produces a wider, shallower one, and in a double-height room those two objects behave completely differently in the sight line from the balcony.
A second form of increment conflict concerns how the rule stops. Some published methods add to the fixture without limit as the ceiling rises; others cap the addition at a point and let the remaining height be dealt with by suspension length rather than by fixture size. The capped versions exist because a fixture that keeps growing eventually stops looking like a fitting and starts looking like a structure. Knowing which type you are holding matters: if you apply an uncapped rule in a very tall space, you will produce a very large object, and you may conclude that the rule is wrong when in fact it is a rule that was written for ordinary domestic ceiling heights and was not intended for the volume you have.
The remedy for increment conflicts is to test the rule at the extremes of your own space before adopting it. Apply it at the minimum and maximum height the fixture will be seen from, and look at the two answers. If the progression produces a fixture that is plausible at one end and absurd at the other, you have found the assumption the rule was written under, and you can then decide whether that assumption applies to you. This is a five-minute exercise, and it is far cheaper than discovering the answer after installation.

Conflict Three: The Count and Spacing Method Changed
The third kind of conflict concerns arrangements rather than sizes, and it is the one that most often produces a result that satisfies the formula and still looks wrong. Three structures are in common use. The first divides the length evenly, giving every interval the same size. The second treats the gap between adjacent fixtures as the variable to be controlled, and sizes that gap in relation to the diameter of the fixtures themselves. The third works from the centres of the fixtures, placing each centre at a fixed position and letting the edges fall where they fall.
Run the three structures across the same length and the same number of fixtures and you get three different arrangements with the same total extent. The even division spreads the fixtures across the whole run and leaves no obvious focal point; it suits a passage or a long counter where rhythm matters more than emphasis. The gap-controlled version holds the visual interval constant, which suits a series of identical objects where the eye is reading the space between them as much as the objects. The centreline version groups the fixtures toward the middle of the run and leaves more space at the ends, which is a different design statement: it makes the arrangement read as a cluster rather than as a line.
Conflicts in this family also arise from what the rules assume is being spanned. One method assumes the run is bounded by two ends and places fixtures inside those bounds; another assumes the arrangement should be centred on the feature below it, whether that is a table, a counter or a doorway. In a symmetrical room the difference is small. In an asymmetrical one it is the whole question, because centring on the feature and centring on the room put the arrangement in different places, and one of them will look deliberate from most viewing angles while the other looks like a mistake.
There is also the question of what happens at the ends. A rule that holds the gap constant will produce a different relationship to the wall at each end of the run as the run changes length, long before the change is large enough to be noticed in the fixture count. This is why an arrangement can look right in a drawing and slightly off on site: the drawing was made to the nominal length, and the built length is different by an amount that the arrangement absorbs at its ends rather than in the middle. If the wall or the counter matters visually, decide in advance whether the arrangement is anchored at the ends, at the centre, or at the fixtures that light the seats.
The remedy for this family is a plan, drawn to the built dimensions rather than to the nominal ones, with the arrangement laid out and the leftover allowance distributed where someone has consciously decided it should go. When two spacing methods disagree, the disagreement is not a numerical problem to be settled by preference for one page over another; it is an instruction to decide what the arrangement is for. Once that decision is made, the arithmetic follows it rather than the other way round.
Conflict Four: The Unit and the Conversion Changed
The fourth kind of conflict is the least glamorous and produces more of the confusion than it deserves. Two measurement systems are in circulation around this subject, and they do not convert into one another cleanly. A rule published in one system and quoted in the other has been rounded twice: once when the author simplified their own figures, and again when the reader converted them. By the time your search returns it, the rule may no longer be the rule that was written.
The visible symptom is a discrepancy that is too small to be a different design intention and too large to be arithmetic error. Two sources give you the same structure, the same basis and the same reasoning, and their answers differ by a margin that no amount of reading resolves. When that happens, the likely explanation is conversion, and the way to test it is to work the rule in the unit it was published in rather than in the unit you prefer, and to compare the structure of the rule rather than its output.
There is a second, subtler conversion problem in this family: the difference between a linear measure and a proportional one. A rule that says the fixture should be a particular fraction of the room works in any unit, because fractions are dimensionless. A rule that says the fixture should be a particular length of addition for every unit of ceiling height does not, because it carries a unit with it. When a proportional rule and a linear rule are presented side by side, they look similar and behave differently, and mixing them in one specification is a reliable way to produce an answer nobody can defend.
Finally, note that rounding matters more than the unit itself. Rules are published as memorable figures, and a figure that is memorable is by definition rounded. Two variations of the same method, one rounded up and one rounded down, will bracket your calculation without either of them being the version the original author used. If you are reconciling a range of published rules rather than a single one, the honest approach is to treat each as a range rather than a point, and to let the spread inform the decision rather than trying to collapse it to one figure.
Where These Conflicts Come From
Understanding the four conflict families tells you how two rules can differ. It does not explain why the same subject is covered by so many incompatible rules, and that explanation is worth having, because it tells you how much confidence any published figure deserves. Four production causes account for most of what is out there, and none of them is a conspiracy. They are what happens when helpful information is produced at scale, for an audience that is not described, about a subject that has a real technical core.
The first cause is templated production. Content about home improvement is produced in volume, and volume rewards material that can be adapted to many subjects with small changes. A sizing article is an easy candidate for that treatment: the structure is reusable, the arithmetic is short, and the subject appears to be self-contained. The result is a large body of advice that is competent in form and unanchored in application. When you read a published rule, one of the useful questions to ask is whether the page shows any sign of having been written for a particular room, or whether it is a well-formed answer to a question nobody asked in those terms.
The second cause is repeated conversion. As set out above, a rule that has been carried between two measurement systems and rounded at each step is no longer the rule that was written. Conversion damage accumulates quietly, because each individual step looks reasonable and the result still produces a plausible figure. A rule that has travelled a long way is not necessarily wrong; it is simply no longer precise, and it should be treated as an indication rather than as a specification.
The third cause is the default that escaped its context. A method developed for one room type — a dining room in a house with ordinary ceiling heights — is presented as general advice about fixtures, and is then applied to a double-height foyer, a restaurant, a hotel lobby or an atrium. The method was not designed for those spaces, and it does not become suitable merely because it is the first result that appears. The fourth cause is the missing scope statement. Almost no published rule states the range of conditions it was written for, which is precisely the information a reader needs, and it is absent because stating it would make the advice less useful to a general audience and less quotable.
Put together, these four causes explain something that is otherwise puzzling: the rules are not converging on an answer as the subject gets more attention. They are multiplying, because production at scale rewards new angles rather than reconciliation. The practical implication is that the burden of establishing scope falls on you, which is the subject of the rest of this article. It also means that the most valuable thing you can bring to a supplier conversation is not a formula but a description of your room, because the room is the part of the problem that no published rule has.
A Four-Step Audit You Can Run Before You Order
The audit below is deliberately short enough to run in one sitting, and it produces a defensible answer without requiring anyone to decide which website was right. It assumes you have the room, or at least a measured plan of it, and it produces a range rather than a figure, because a range is what the evidence supports.
Step one: fix the basis. Decide what the fixture is being sized against, and write the decision down in one sentence. If the fixture is the centrepiece over a table and the room has other layers of light, the basis is the table. If it is the single significant object in a volume and will be seen from across the space, the basis is the room. If the ceiling is unusually high, height is applied as a modifier on top of the chosen basis rather than as a competing basis. Writing the sentence matters because it is the part that later readers of your schedule will disagree with, and it is much easier to defend a stated intention than an unexplained figure.
Step two: fix the applicable conditions. List the conditions each candidate rule assumes, and check them against your space. The working list is short: ceiling height, the height of the surface beneath the fixture, the presence of a partition or a change of level, the principal viewing positions, and the number of layers of light in the room. A rule that assumes a surface at ordinary height is not applicable over a counter that is taller, and a rule that assumes a solid ceiling is not applicable to a recessed void. Most published rules fail at this step rather than at the arithmetic, and failing here is a success: you have eliminated a wrong answer cheaply.
Step three: carry the candidates as a range. Apply the rules that survived step two, and keep the answers as a range rather than collapsing them to one figure. Read the range for what it says: a narrow range means the methods agree and the choice is mostly aesthetic; a wide range means the methods are answering different questions, and you should revisit step one rather than average them. Averaging two answers produced by two different intentions produces a figure that satisfies neither intention, which is one of the ways a fixture ends up looking neither generous nor restrained.
Step four: sign off against the site. Take the shortlisted figure back to the room and test it against the things that are not in any formula: the position of the fixing point, the structure that the fixing point is attached to, the sight lines from the door and from the seats, and the clearance between the fixture and anything it will be seen against. Then record the figure as provisional until it has been confirmed by measurement on site and by a structural check of the fixing arrangement, and confirmed against the requirements that apply where the building is. An audit that stops before this step has produced a good drawing and an unverified installation.

Signing It Off: Measurement, Structure and Local Requirements
The audit produces a size. What turns that size into something you can order, install and defend is a small amount of documentation, and the documentation is where this subject is usually abandoned. Three items belong in it, and none of them is a formula.
The first is measurement. Sizes used in the audit should come from the room as built, taken with a tape and recorded, rather than from a drawing that may predate a change or represent a nominal dimension. The difference between nominal and built dimensions is small in most rooms and decisive in some: a change of level, a recessed ceiling, a later addition or a wall that is not parallel will all move the figure that the formula was applied to, and they will move it in the direction that matters for a fixture hung in the middle of the space. Record every dimension with the date it was taken and the person who took it.
The second is structure. The size of a fixture and the ability of the ceiling to carry it are two separate questions, and the second is answered by a structural check of the fixing arrangement rather than by the fixture supplier alone. What has to be established is where the load is going: whether the fixing point lands on a structural element, whether a spreader or a reinforced section is needed, and what the condition of the supporting construction is. This is the part of the process that people are most tempted to skip, and it is the part with the least tolerance for a guess.
The third is local requirements. Sizing and installation in a building can be affected by requirements that apply where the building stands — in some cases to the fixing and the ceiling construction, in others to the electrical installation, and in some to the room’s use. Those requirements are not universal and they are not published as interior design advice, which is exactly why they are missing from the pages that gave you the formulas. They have to be confirmed locally, by whoever is responsible for compliance on the project, before the schedule is frozen.
One sentence is worth carrying out of this section and into every schedule, because it is the honest summary of the whole subject: all dimension values must be based on our own on-site measurement and the supplier’s structural calculation, and confirmed against the requirements that apply to the building. Once that sentence is in the document, the formulas stop being authorities and become what they actually are — starting points, useful for narrowing the field, and not the thing that decides.
The same discipline applies to the fixtures that are not the centrepiece. Sconces, wall lights and secondary fittings have their own published rules, and they suffer from the same conflicts; our guide to heights and sizes for wall-mounted fittings deals with that family separately, since the conflicts there are vertical rather than horizontal. Where a fixture sits in a multi-storey volume, the questions change again, and the guide to fixtures in stairwells and atria sets out how the viewing positions multiply.
What to Send a Supplier Instead of a Formula
The practical payoff of all this is a different first message to the people who will make the fixture. A formula sent to a supplier invites a quotation for a fixture of that size, produced without reference to the room, and it puts the supplier in a position where the sensible reply is to build what was asked. A description of the room invites a schedule, and it gives the supplier the information needed to tell you when the size in the formula will not work.
Send four things. Send the measured plan of the space, showing the room outline, the position of the table or counter if there is one, and the fixing position if it is already established. Send the vertical measurements that matter: the ceiling height, the height of the surface beneath the fixture, and the height from which the fixture will be seen from the principal viewing positions. Send the structure — what the ceiling is made of and where the structural elements run, or a note saying it is not yet established. And send the intent in one sentence, in the form you wrote at step one of the audit: what the fixture is being sized against, and whether the room has other layers of light.
Then ask for three things back. Ask for a schedule that states the fixture dimensions and the suspension arrangement as a proposal rather than as a confirmation. Ask which dimensions the supplier would check before manufacturing, and what information they would need to close those checks. And ask what the fixture will weigh and how that load reaches the structure, because that answer is needed regardless of the size, and asking for it early is far cheaper than discovering it late. A supplier who answers those three questions in writing is a supplier you can work with on a project where the dimensions matter.
For custom work, the same package is the input to a brief, and our guide to writing a brief for custom lighting sets out how to structure it so that the answers come back comparable between suppliers. Where the project involves a run of public areas rather than a single room, the discipline of reading a specification the way an engineer does is the appropriate next step, because at that scale the fixture size stops being the variable and the schedule becomes the instrument.
None of this removes the need for judgement, and that is the point. Size, in a room that actually exists, is not a constant to be looked up. It is an input taken from the space, a decision about what the fixture is for, and a verification against the building it hangs in. Once those three things are in place, the confusion that brought you here stops being a problem: you can read any published rule, see what it is answering, and take from it only the part that applies to you.
Frequently Asked Questions
Why do different sites give different chandelier sizes for the same room?
Because they are answering different questions. One rule sizes the fixture against the table below it, another against the volume of the room, and a third against the height of the ceiling. Each produces a defensible answer to its own question, and the pages rarely state which question they are answering. Comparing the outputs of two rules is therefore not a way to find an error; it is a way to find out that two different design intentions are in circulation.
Should I average the sizes that different rules produce?
Averaging is best avoided. A figure that sits between a table-based answer and a room-based answer satisfies neither intention, and it is harder to explain to anyone who reviews the schedule later. The more useful approach is to decide what the fixture is for, apply the rule that matches that decision, and record the other answers as a range so that the sensitivity of the choice is visible.
Does a bigger ceiling mean a bigger fixture?
Usually, but not without a limit. In tall rooms the fixture has to register at a greater distance, which is the reason methods add to the size as the vertical space increases. Published methods differ in how much they add and in whether the addition stops at some point; the ones that stop exist because a fixture that keeps growing stops behaving like a fitting. Test any method you adopt at the lowest and highest ceiling in your space before you commit to it.
What if the fixture only has to look right from one place?
Then say so, and let it shape the decision. A fixture viewed from one seat, one door or one counter has a smaller set of relationships to satisfy than one seen from a balcony and from underneath. Naming the principal viewing position is a legitimate input to the audit, and it often resolves a conflict between two published rules, because the two rules were written for two different viewing conditions.
Do I need a structural check for an ordinary fixture?
You need to know what the fixing is attached to, whatever the fixture weighs. For a light fixture in a sound ceiling that is often a matter of confirming what is there; for anything heavy, for anything hung in a void, and for anything suspended above people, it becomes a calculation. The size of the fixture and the ability of the ceiling to carry it are separate questions, and the audit in this article settles only the first of them.
Choosing the Right Size for a Real Room?
Send us the measured plan, the ceiling height and the height of the surface below the fixture, the construction of the ceiling, and one sentence on what the fixture is for. We will return a proposed size range with the reasoning stated, the suspension arrangement that suits the available height, and the list of checks that need to be closed on site before the fixture is manufactured.
Our measured survey of fixtures recorded across installed projects shows what these decisions look like once built, and the guide to fixtures in tall and double-height rooms covers the case where vertical space drives the choice.













