Two pods both claiming thirty decibels are not comparable unless both were tested to the same standard.

Ask which standard produced the number, not just what the number is

The single question that makes a spec sheet meaningful. Suppliers across this market give the same warning: compare decibel reduction figures only among suppliers who obtained their ratings to ISO 23351 from a certified external testing facility, because in-house testing methodologies vary widely with sometimes arbitrary, inflated or fake results. One puts it plainly, that any figure not obtained to the standard is most likely inflated. The practical version is a request rather than a judgment: ask for the ISO 23351-1 test report and the name of the testing facility, and treat a brand that cannot produce third-party test data as unverified on acoustics rather than as bad. Unverified is the accurate word and it is also the more persuasive one in a procurement conversation.

The wrong standard gives a bigger number, which is why this is hard to spot

Worth understanding because it means the misleading figure can be completely genuine. Other acoustic standards are real and widely used: ISO 16283-1 measures sound insulation between rooms in buildings and Dw rates structural elements like walls and doors. Guidance notes that these may give higher decibel ratings for a pod while being less relevant, since they address structural sound insulation rather than speech privacy. So a supplier quoting a Dw figure is not necessarily inventing anything, they are answering a different question with a more impressive answer. The flattering number is the less applicable one, and the only way to tell is to ask which test produced it.

The standard measures the leak, not the wall, and a pod is mostly holes

This is why a purpose-built standard beats a borrowed one, and the reasoning is concrete. The test measures outward speech level, meaning how much of the occupant’s voice leaks outside the pod, rather than the acoustic properties of individual panel materials, which one guide describes as making it a far more accurate predictor of real-world speech privacy than any partition-focused standard. Picture what a pod actually is: a door with a seal, a ventilation path, a cable entry, and glass in a frame. A panel-based test evaluates a wall sample and misses every one of those. A whole-assembly test cannot miss them, because they are part of the assembly being measured.

Twenty-five decibels is the floor, and it depends on your ambient noise

Two numbers to carry, and the second one is the part people miss. The stated threshold is that for an office pod to ensure speech privacy in a typical office environment, it must achieve a minimum reduction of twenty-five decibels measured to ISO 23351. But that guidance is explicitly conditioned: as long as the office produces roughly forty decibels of background noise, Class A+, A and B will all deliver speech privacy, with most offices averaging forty to sixty. So the requirement is not a property of the pod alone, it is a relationship between the pod and the room. A genuinely quiet office provides less masking noise, which makes speech privacy a harder problem rather than an easier one, so measuring your own ambient level before setting a target is worth half an hour.

And there is a ceiling, so more is not automatically better

Useful for resisting an upsell and it comes from a supplier rather than a skeptic. Asked whether higher reduction is always better, one answers not automatically, since thirty to thirty-five decibels of reduction is enough to bring most open-plan noise down to a private conversation level, and going higher adds cost without a proportional benefit in a typical office setting. Another makes a subtler point in the same direction, that the goal is an environment which is quietly productive rather than eerily silent, and that complete acoustic isolation would remove a psychological comfort worth keeping. Between the twenty-five decibel floor and the low thirties there is a real band worth paying for. Above it you are buying a specification rather than an outcome.

Know the classes, and know that nobody is at the top

Useful context for reading a data sheet. The classification runs from A+ downward, with A+ defined as above thirty-three decibels of speech level reduction and Class A given as thirty to thirty-three. Sources differ on where the scale ends, one describing it as running to D and another to E, so treat the bottom of the range as uncertain. And one detail worth knowing before anybody claims it: as of the sourcing here, there are no known manufacturers with an A+ rating. An independent comparative study exists, with Turku University of Applied Sciences measuring one manufacturer’s pod alongside ten other best-selling booths to the standard, the best of those coming in at 30.3 decibels. That study was published by the manufacturer whose pod won it, which is worth knowing while reading it.

Two words that mean the figure is not a measurement

A short tell that saves a lot of reading. Guidance names it directly: a common mistake is accepting an "up to thirty-five decibels" claim without a test method attached, because up-to figures are lab best case rather than what you get with a door left slightly ajar. Notice where that observation lands, which is the door. In a product whose performance is decided by seals and openings, the gap between a best-case laboratory figure and daily use is largely a question of whether the door closed properly, which is a maintenance and installation issue rather than a specification one. So "up to" is a signal to ask for the report, and a well-fitted door is part of what you are buying.

Most buyers compare these in the wrong order

The procurement point, stated bluntly by a supplier: most buyers skip the acoustic verification entirely and go straight to capacity and price, which is backwards, since a pod that seats six but reduces noise by fifteen decibels instead of thirty to thirty-five fails the actual job it was bought to do. That is worth sitting with, because capacity and price are easy to compare and acoustics are not, so the comparison naturally drifts toward whatever is legible. The fix is sequencing rather than effort: establish a minimum DS,A figure against your own ambient noise, eliminate anything that cannot document it, and only then compare seats, footprint and cost among what remains.