Why a Cosmetic Pen Stops Working Before It Is Empty
A pen that has stopped is usually not a pen that has run out. That is the first thing to establish when a complaint comes back from the field, and it is the thing that gets skipped, because the empty explanation is cheap and the real one requires a number nobody wrote down.
We hold specifications for 44 pen models across five families — 1.5–10 ml, bodies from Ø6.63 to Ø30.0 mm. Almost none of those specifications include a dose figure. That absence, more than any material choice, is what makes cosmetic pens fail in ways buyers find baffling.
Three numbers that have to agree
A pen’s working life is the smallest of three quantities, and all three are knowable before tooling is cut.
Fill volume is the one everybody specifies. It is on the catalogue page, it is in the mould drawing, and it is the number the brand puts on the carton.
Dose per actuation is the one that usually goes missing. On a twist pen it is the volume released per turn; on a click pen, per press. Either way it is a quantity that can be measured and written down — and either way it usually is not, replaced in the brief by controlled dosing, a phrase true of every pen ever made and binding on none.
Mechanism life is the one nobody asks about at all. Every advancing mechanism has a rated number of actuations before its travel degrades or stops. It is a component specification, it exists in the supplier’s data, and it is almost never in the purchase specification.
Take the fill volume and divide it by the dose per actuation, and you have the intended number of uses. Compare that to the mechanism’s rated life and you have the answer to the only question the end user cares about: does the pen empty, or does it quit? A mechanism rated below its dose count strands product at the bottom of the shaft, and the customer sees exactly that — a pen that quit with product still sloshing inside it.
This arithmetic takes about four minutes. It is not done, in our experience, on most pen projects, because the two numbers it needs are held by two different parties: the dose by whoever signs off the formula, the mechanism life by whoever supplies the component.
Priming is a cost, not a step
No pen delivers on the first actuation. The feed system has to be charged from the reservoir to the tip first, and every actuation spent doing that is product the user never applies. On a 2 ml pen that is not a rounding error; it is a measurable share of the fill.
Two things follow, and both are commercial rather than technical.
The first is patience. A user who presses or turns a pen several times before anything appears is forming a judgement about the product during those presses. Whatever the brief says about dosing precision, that user’s first impression is set by how long the pen took to wake up.
The second is re-priming. A pen left unused for weeks will need to be charged again, because product has moved back down the feed path or the tip has partly dried. This is where pens fail quietly — not dramatically, no leak on the carpet, just a consumer deciding the product is dud and not buying it twice. Ask your supplier what re-priming looks like after a defined idle period, and treat the answer as a specification, not a footnote.
The cap is either a seal or a lid
Of everything on a liquid pen, this is the one that decides whether it survives its own shelf life — and one you can neither photograph nor feel in a sample.
Covering the tip is not the same job as closing the tip chamber. A cover keeps dust off and does very little else; a closure shuts the flow path, so the formula is not simply open to air between uses.
When nothing closes, the tip dries and every use re-charges it through the same narrow path. The user reads that as a pen drying up while it is still more than half full — the complaint most often misdiagnosed as an empty pen, and the one most often answered by simply adding fill.
So do not ask about the cap; ask which interface closes, and ask in those words. A supplier who can answer has thought about it. One who answers by describing the shape of the cap has not.
Symptom to specification
Field complaints arrive as symptoms. Specifications are written as numbers. The mapping between them is short, and it is worth having on one page.
| What the user reports | Usual mechanism | Specification clause to fix |
|---|---|---|
| Stopped while still full | Mechanism rated life below the fill’s dose count; or the tip chamber is not sealed and the feed path has dried | Rated actuation life vs fill ÷ dose; sealing interface |
| Nothing comes out, or output is erratic | Feed system and formula viscosity outside their matched window — too thick to travel, or air drawn back into the path | Feed type and its stated viscosity range |
| Leaks in a bag or in transit | Seal and piston fit under pressure and temperature swing; orientation sensitivity | Leak and drop testing, named as separate line items in the report |
| Floods or over-delivers | A dose increment too large for the use case, on a mechanism that cannot retract | Dose per actuation, stated as a tested figure |
| Tip clogs or drags | Particle size and solids load against tip aperture | Tip construction checked against particle size, not against styling |
| Runs out faster than expected | Priming and re-priming losses not counted in the fill calculation | Priming actuation count, and re-priming after idle |
Read the third column as a list of things to demand in writing. Every one of them is a number or a named test. None of them is a matter of opinion, which is the point — a pen failure should be traceable to a clause, and if it is not, the specification was the defect.
What an inspection can actually catch
There is a version of this article that ends with insist on quality control. That is not useful, because inspections are only as good as what they were told to look for. Two records from our own files show the difference between an inspection that measures and one that certifies.
On a 2015 order of 18,100 plastic sprayers — a different format, same inspection discipline — sampling turned up untrimmed burrs on 80% of the units checked. The lot was marked Pending rather than passed and the factory had to put it right before anything shipped; the factory’s objection to a drop test sits in the same report, next to the finding it did not want to run. Nobody’s paperwork was flattered by that report.
The other record is a pass, and the more instructive of the two. A 2007 order of 130,464 aluminium jar sets was sampled to ANSI/ASQ Z1.4 Level II against limits of 10 major and 21 minor; it came back with 6 and 5. The note records that minor logo scratches sat inside the boundary, and the lot was released. That is what a passed lot looks like when someone writes the numbers down: not a clean sheet, but a lot that stayed inside a stated boundary.
For pens specifically, the line items that matter are mechanical rather than visual: leak testing, a drop test, actuation count to first delivery, dose check across a sample, and tip aperture against the formula. A report that grades appearance and does not test the mechanism has not tested the thing that will actually fail. The inspection records are published here →
What to put in the brief
If you take one thing from this, take the arithmetic. Before tooling, settle five things in writing:
- The formula’s viscosity range, from whoever signs off the formula. The feed system is chosen against this, not against the product category.
- Dose per actuation, as a tested figure with the test described.
- Rated mechanism life, as an actuation count. Then check it against fill ÷ dose.
- Priming and re-priming behaviour, including after a stated idle period.
- Which interface the closed cap seals, described rather than implied.
Four of those five are routinely absent from a first brief. They are also the four that decide whether the pen that leaves your warehouse is the pen the customer thinks they bought.
For the format families themselves and how they differ mechanically, see the five pen formats. The model data behind it, family by family, is published on the cosmetic pens page. If a pen is already failing in your line, the fastest route is to describe the symptom and name the model: the quotation form starts here.