Your Spec vs Ours: Building a Gap Table That Works

Most specification comparisons are run as a single question: does the material meet our specification. The question is answerable from the certificate, and the answer is useful exactly once. It says nothing about which differences between the two documents matter, how often they will bite, or what to do when they do.

A gap table answers those instead. It is not a more elaborate version of the same check; it is a different document, built once and used for years, and it changes the conversation from whether a batch passes to where the two specifications actually diverge.

Two documents with two purposes

The specification we release against exists to define commercial material and to make release decisions reproducible. It states limits for the tests that are performed and it is written so that a batch either conforms or does not.

A buyer's in-house specification exists to protect a process. It may tighten a limit because a downstream step cannot tolerate more, add a test that the released specification does not carry, or set a narrower particle-size window because a compaction step is sensitive to it. Those additions are not criticisms of the released specification; they are statements about a different process.

The mistake is to treat the two documents as competing versions of the same thing. They are not. One describes what is supplied; the other describes what can be used. The gap between them is the subject of the table.

The four columns that do the work

A gap table with the right columns takes an afternoon to build and saves the same time repeatedly. Four columns are enough.

The first is the test item. The second is the limit as released. The third is the buyer's limit. The fourth is where the work happens: a classification of the gap and the action that follows from it.

Item	Released limit	Your limit	Gap and action
Content (dried basis)	As stated on the certificate	Your range	Compare only after confirming both are on the same basis
Named related substance	Individual limit	Your limit, if any	Decide whether to add a limit or accept the released one
Loss on drying	Limit plus method	Your limit	Confirm the drying regime matches; otherwise the numbers are not comparable
Particle size	Reported distribution	Your window	Qualify the window as a first-article measurement if your process needs it
Residual solvents	Named solvents and limits	Your list	Add any solvent your process or market requires to be controlled
Microbial limits	Limits by route of use	Your limits	Check the route assumption behind each set

The fourth column is where the value sits, and it is the column most comparisons omit. A gap that can be closed by adding a first-article test is a different problem from a gap that requires a change in the released specification, and a gap that exists only because the two documents use different bases is not a gap at all.

Where the gaps usually appear

Four areas produce most of the differences in practice, and each has a characteristic cause.

Calculation bases are the first and the most common. Content on a dried basis compared against content as-is, or purity by area normalisation compared against a reference-standard assay, produces an apparent gap that is an artefact of definitions. This is worth resolving first, because it removes items from the table rather than adding them.

Particle size is the second. The released specification may report a distribution where the buyer's process needs a window, and a reported value is not a control. Turning a reported measurement into a controlled one is a real change and should be agreed as such.

Residual solvents come third, and the cause is usually market rather than process. A buyer serving a market with a particular expectation will want solvents controlled that the released specification does not name, and the right response is to add them for that buyer's orders rather than to argue about the default set.

Microbial and endotoxin limits are the fourth, and the cause is the assumed route of use. Related substances limits and the microbial panel are set against an assumed application, and where the two documents assume differently the limits will differ without either being wrong.

What "meets specification" does not cover

Conformance to a specification is a statement about the tests in that specification. It is not a statement about anything the specification does not test, and this is where buyers are most often surprised. A certificate can show complete conformance while the material differs in a property that matters to the buyer's process and appears nowhere on either document.

The gap table is the instrument for finding those. Every item on the buyer's own specification should appear in it, including items that the released specification omits, because an omission is a gap of a particular kind: one that will never produce a failing result and will therefore never be noticed until it causes a problem downstream.

How to run the exercise

Start from the buyer's own specification rather than from ours, because the purpose is to protect the process. For each item, record the released limit and the source document it came from, then record the buyer's limit and its basis, then classify the difference.

Four classifications cover almost everything. Equivalent items, where the two agree once the bases are aligned. Tighter items, where the buyer's limit is stricter and the exposure is the buyer's. Additional items, where the buyer tests something the release does not. And differing bases, which is not a gap but a definition problem to be closed before anything else is analysed.

Where an EP grade claim or a material without a monograph is involved, add a fifth class for monographs: items where a pharmacopoeial comparison is possible and items where it is not. The distinction matters because a monograph comparison is a comparison against a public document, while a released specification is a comparison against ours, and the two carry different weight in an audit.

What to do with the result

A completed gap table usually produces three actions. Some items are reclassified as equivalent once the bases are aligned, which reduces the apparent work. Some items become first-article measurements, taken once to qualify the material and then held. A small number become specification discussions, where the buyer asks for a tighter limit or an additional test on that buyer's orders.

That third category is the one worth handling deliberately. A specification agreed for one buyer's orders is a real commitment on both sides, and it should be recorded where both companies can find it rather than living in the correspondence. Comparing EP and USP limits is a useful rehearsal for that conversation, because it exercises the same discipline of matching definitions before matching numbers.

A short worked example

Consider a buyer whose in-house specification and our released specification differ in three places. The content limit is the same figure on both documents. The buyer's is written on a dried basis; ours states the basis next to the value. The comparison is equivalent once the bases are matched, and no further work is needed.

The particle-size item differs in kind rather than in value. Our specification reports a distribution; the buyer's sets a window on the upper end because a compaction step is sensitive to the coarse fraction. That is a genuine gap, and the resolution is a first-article measurement rather than a limit change, because what the buyer needs is assurance about the material they receive rather than a permanent change to the specification.

The third item is a residual solvent that the buyer's market requires to be controlled and our specification does not name. That gap is closed by adding the test for that buyer's orders, which is a real commitment on both sides and belongs in the agreement rather than in the correspondence.

Three apparent differences produced one no-op, one qualification activity and one specification change. A pass-fail comparison would have produced the same three differences with no indication of which was which.

Keeping the table alive

A gap table is built once and maintained, and maintenance is where most of them fail. Three events should trigger a review. A revision of the pharmacopoeial monograph the material is compared against, because a monograph revision can change a limit without either company having done anything. A change in the buyer's process, because a tighter limit usually encodes a process requirement and process changes move those requirements. And a change in the released specification, which should be communicated rather than discovered.

Where the material has a monograph, the European Pharmacopoeia publishes its revisions on a schedule of its own, and a specification built against one edition can be out of date without anyone noticing. Keeping the edition stated in the table is the simplest way to spot that.

Where the material has no monograph, the table has no external anchor and the released specification does all the work. In that case the review triggers are internal, which makes them easier to miss and worth writing into a calendar rather than a memory.

Where this connects to audits and questionnaires

Supplier audits and quality questionnaires almost always ask for the specification and the certificate, and rarely ask what the buyer's own specification says. The gap table is what lets a buyer answer the next question, which is whether the two are compatible. It also converts a recurring conversation into a document: once the table exists, a change of personnel on either side does not restart the discussion from the beginning.

The audit dimension is worth stating plainly. An auditor who sees a material received against a specification that differs from the released one will ask how the buyer established that the difference does not matter. A gap table with classifications answers that question directly, and the framework that governs specification setting is the reference an auditor will have in mind when asking it.

For materials intended for regulated markets the documentation expectations are set out in the relevant guidance documents, which treat the specification as a controlled document rather than as a reference sheet. That is the standard the table should be built to, even where the material is not yet destined for such a market.

Where the exercise most often stalls

Gap tables stall in the same place. The buyer's specification exists, the released specification exists, and the two use terms that look the same and are not. Content, purity and assay are the usual culprits, because each can be defined more than one way and the difference is invisible on a single line of a document.

The way through the stall is to write the definition next to the limit on both sides before comparing anything. That single step converts most apparent gaps into equivalences and reduces the size of the table considerably. Where the definitions still differ after that, the difference is real and the remaining work is worthwhile.

A second stall appears when the buyer's specification contains items that have never been tested because the material has always been bought from one source and the item was inherited from an earlier document. Those items should be classified as unevaluated rather than as met, because an unevaluated item is a gap of exactly the kind the table exists to expose.

A third appears when the two specifications are written for different applications of the same material. Where that is the case the table should say so at the top rather than at the bottom, because the comparison is between two different requirements and not between two versions of one requirement.

Frequently asked questions

Is it a problem if a supplier's specification is broader than mine?

Not automatically. What matters is whether the items that differ are items your process depends on. A broader limit on a property that does not affect your route is a difference on paper; a broader limit on a property your process is sensitive to is a real gap.

Why not simply ask whether the material meets my specification?

Because the answer is on the certificate and does not tell you what happens when it does not. A gap table tells you which differences will generate failures, how often they are likely to, and what the fallback is when they do.

Which specification takes precedence?

The one written into the supply agreement. A specification is a commercial document as well as a technical one, and where two documents differ the contract decides. The gap table exists so that the difference is known before a dispute rather than during one.

What if my specification has no entry for a test the supplier reports?

Then the test is informational for you rather than controlled, and the table should say so. It is worth deciding deliberately whether to add a limit, because an uncontrolled line that is reported on every certificate tends to become an expectation by habit.

How often should the gap table be revisited?

Whenever either specification changes, and at least at each supplier qualification review. Monographs are revised on a cycle of their own, so a table built against one edition can be out of date without anything having happened at either company.

Does a tighter in-house limit create a compliance obligation?

It creates an obligation on you. A specification tighter than the released one means material can conform to the release specification and still fail yours, so the incoming test programme has to be able to detect that. The gap table is where that exposure becomes visible.