Cost per reported result is unit price divided by yield, and yield is never

A 96-well plate run that reports 88 results instead of 96 has a rerun rate of 8.3%, and that single number moves the cost per reported result more than any line on the quote. I have watched procurement teams negotiate a 9% discount on vials with caps and then lose 14% of the value to reruns caused by a septum that coring-plugged at injection 61. The unit price is what you pay. The cost per completed test is what you actually spent. If you are building a budget model for a chromatography lab, the second number is the only one that survives contact with a real sample queue, and the calculation behind it is what this page covers.

Cost per reported result is unit price divided by yield, and yield is never 100%

The naive calculation is price per piece divided by tests per piece. A $0.42 autosampler vial used once is 42 cents per test. That is arithmetic, not accounting. The working formula has four terms:

Cost per reported result = (consumable cost + prep labor + instrument time on failed runs + rework consumable cost) / (results accepted by the reviewer)

Each term is measurable. The denominator is the part most people guess at. In a clean method on a validated system, a 2 to 4% rerun rate is normal and expected (system suitability failures, a bubble in the syringe, a sample that arrived hot). Above 6%, you have a consumable or method problem, not bad luck. The gap between those two numbers is where the money is.

I use three yield inputs when I model a consumable change:

• Rerun rate: injections that produced no usable result and were repeated. Track it per 100 injections, not per batch.
• Reject rate: results that were produced but thrown out at review (integration, carryover, out-of-calibration). These cost you twice: the run and the review.
• Prep time per injection: minutes of technician time to get a sample into the autosampler, including weighing, dilution, filtration, and capping.

The prep number is the one people leave out. At a fully loaded technician cost of $45/hour, 4 minutes of prep is $3.00 per sample. That dwarfs a 42-cent vial. A consumable that saves 30 seconds of prep per sample is worth more than a 50% price cut on the vial itself. This is why “cheap” and “expensive” are almost meaningless words in a consumables bid.

What a rerun actually costs: the three buckets nobody itemizes

A rerun is not one cost. It is three, and they land on different budgets, which is why they get missed.

Bucket 1: the consumable you burned twice

Every rerun consumes a second vial, a second cap, a second septum puncture, a second aliquot of sample, and a second slug of mobile phase or carrier gas. On a GC method running helium at 1.2 mL/min, a 20-minute run is 24 mL of carrier. At industrial helium pricing that is real money per injection, and it doubles on a rerun. On the LC side, a 30-minute gradient at 1.0 mL/min burns 30 mL of mobile phase, and if you are running MS-grade acetonitrile at $180 per 4 L, that is $1.35 per injection in solvent alone. A 6% rerun rate adds roughly 8 cents per reported result in solvent. Small, but it compounds across thousands of injections.

Bucket 2: instrument time you cannot get back

An autosampler queue is a fixed-capacity resource. If your LC-MS runs 18 hours a day and you lose 6% of that to reruns, you have lost about 65 minutes of instrument time daily. If that instrument bills internally at $120/hour (a defensible loaded rate for a triple quad), that is $78 per day, roughly $19,000 per year per instrument. You do not see this on any invoice. You see it as a backlog and a decision to add a second shift.

Bucket 3: the review and the rework

A rejected result has to be caught by a human, documented, and re-queued. A data reviewer spending 6 minutes on a failed injection at $50/hour costs $5.00 per reject. That is more than ten times the price of the vial that caused it. When a septum cores and drops a plug into the inlet, the reject is not just one injection; it is the whole sequence after the plug, because the plug sits in the liner and every subsequent injection carries a fragment of it. I have seen a single cored septum invalidate 40 injections before anyone looked at the chromatogram.

This is the mechanism behind the sibling point that chromatography products are not interchangeable across grades. The grade determines the puncture count before coring, and the puncture count determines how many injections you get before the first plug.

Two grades, one method: a side-by-side cost-per-test model

Here is the comparison I actually run for customers. Two 2 mL autosampler vial-and-cap configurations on the same 200-sample-per-day LC-MS method. Both fit the same autosampler. The method uses a 22-gauge needle, 10 µL injections, and a 30-minute run. Technician time is loaded at $45/hour, instrument at $120/hour, reviewer at $50/hour.

Read that table twice. The Grade B consumable costs 87% more per unit and still wins, but only by about $1.33 per reported result, or 2%. The win comes almost entirely from the 5.4-point drop in rerun rate and the 0.6-minute prep saving, not from the vial price. On a 200-sample day that is $266/day, roughly $65,000 a year. That is the number to put in front of a procurement manager, not the unit price.

Now change one input: a method where the needle punctures the same cap 8 times (a batch of standards injected repeatedly). Grade A cores at injection 41 and starts producing plugs. Grade B holds to 90. The rerun rate on Grade A climbs to 14%, and the cost per reported result for Grade A jumps to about $71.40 while Grade B stays near $65.60. The gap widens to $5.80 per result. The grade decision is not about the vial. It is about how many times the needle hits the same septum.

Prep time is a consumable property, not a technician property

This is the claim I get the most pushback on, and I will defend it. The time it takes to get a sample into the autosampler is set by the geometry and the closure, not by how fast the analyst moves.

• A cap that needs a crimper adds 8 to 12 seconds per vial versus a screw cap. At 200 samples that is 27 to 40 minutes a day.
• A septum that is too soft (below 40 Shore A) deforms under the needle and grabs it, so the analyst re-seats the vial. That is a hidden 5 to 15 seconds on the samples that fail.
• A vial with a non-uniform neck ID causes the autosampler gripper to mis-seat, and the run aborts at injection. One abort per 50 vials is 2% of your day gone.
• A cap with a liner that is not bonded shifts during transport and covers the septum hole, causing a partial injection. The analyst notices at review, not at prep, so the rerun is a full 30 minutes.

The dimensional discipline behind this is the same one that governs fit across the whole line; if you want the deeper treatment, the note on why vials with caps fail on dimensional grounds rather than force is the place to start. For the cost model, the takeaway is simple: measure prep time per sample on both grades, in minutes, with a stopwatch, for one full shift. Do not estimate it. The difference between 4.2 and 3.6 minutes is $0.45 per sample, and that is $32,000 a year at 200 samples a day.

The reject rate is where the honest bids separate

A rerun is visible. A reject is not, because the result looked fine until a reviewer caught it. Rejects are the expensive half of the yield term, and they are almost always traceable to a consumable lot rather than to the method.

Field case 1: the drifting internal standard

A contract lab running EPA Method 8260 on a purge-and-trap GC-MS saw the internal standard response fall 11% across a 60-injection sequence, then recover on the next sequence. The method was fine. The instrument was fine. The measurement that identified it: the analyst ran a 20-injection sequence using vials from a different lot and the drift disappeared. The cause was extractable siloxanes from a batch of septa that had been cured off-specification. The lot was not visibly different. The fix was a lot-level extractables screen (a blank run of 10 vials from the lot, looking for siloxane peaks at m/z 73, 147, 207, 281) before the lot went into the queue. The 11% drift was producing rejects at roughly 4% of injections, which at $66 per result was about $2.64 per injection lost. The screen costs 40 minutes per lot. It pays for itself on the first lot it catches. The chemistry of that failure is covered in detail on the extractables note; here it matters only as a line item in the reject bucket.

Field case 2: the crimp that loosened at 4°C

An environmental lab stored crimped 40 mL VOA vials at 4°C between prep and analysis, per the method holding time. Rejects ran at 6% on the volatile organics panel. The measurement that identified it: the lab weighed 20 vials before and after 24 hours at 4°C; 6 of them lost more than 0.15 g, which is a leak. The crimp torque was set by feel. The fix was a crimper with a torque setting of 18 to 22 in-lb and a daily check against a reference vial. Rejects fell to 1.2% within a week. The U.S. EPA SW-846 holding-time requirements do not specify a crimp torque, but the leak test does the job. At $66 per reported result, the 4.8-point reject reduction saved about $3.17 per injection on that panel.

Both cases share a shape: a consumable property (extractables, seal integrity) produced a reject that looked like a method problem, and the fix was a measurement on the consumable, not on the instrument. That is the whole argument for modeling rejects separately from reruns.

Puncture count is a spec, and it belongs in the bid

Septum puncture count is the single most useful consumable spec for cost-per-test modeling, and it is the one most often left off the datasheet. Ask for it. A PTFE-faced silicone septum in the 1.2 to 1.4 mm range, 45 to 55 Shore A, typically holds 80 to 100 punctures with a 22-gauge needle before coring begins. A bonded single-layer silicone septum of 1.0 mm often cores at 35 to 45. The number depends on needle gauge and bevel; a 26-gauge needle with a sharp bevel extends the count by 20 to 30%.

Translate that into your method. If your autosampler punctures each cap once, puncture count barely matters and the cheaper septum is rational. If you inject standards 6 to 10 times from the same vial, puncture count dominates, and a septum rated at 40 will core inside your sequence. The cost of that coring is not one rerun; it is the plug in the liner and the 30 to 40 injections that follow it. I have a rule of thumb: if the same cap sees more than 5 punctures, buy the higher-puncture septum and stop arguing about unit price.

For the wider product families and the specs behind them, the Aijiren technical pages list septum hardness and puncture data by part, and the Restek technical resources are a good cross-check when a supplier’s datasheet is thin. For column and inlet consumables, the Agilent GC columns and accessories pages and the Thermo Fisher chromatography and sample handling catalog both publish the dimensional and material specs you need to run the model.

A decision checklist for the bid evaluation

When two suppliers quote different grades, run this checklist before you compare prices. Each line is a number you can get or measure.

• Puncture rating at your needle gauge: ask for it in writing. If the supplier cannot give a number, treat the part as unrated.
• Septum hardness: 45 to 55 Shore A for most autosamplers. Below 40, expect needle grab and mis-seats.
• Liner and cap dimensions: neck ID and cap height within the autosampler’s tolerance (typically ±0.2 mm). Out-of-tolerance parts cause aborts, not reruns, and aborts are more expensive.
• Extractables data: a lot-level blank run, not a generic certificate. Ask for the siloxane screen at m/z 73, 147, 207, 281.
• Rerun rate on a 100-injection trial: run both grades on the same method and count. This is the only number that matters and it takes one shift.
• Prep time per sample: stopwatch, one shift, both grades.
• Seal integrity: for crimp closures, a 24-hour weight-loss check at your storage temperature; for screw caps, a torque spec (typically 8 to 12 in-lb) and a daily reference check.

If a supplier cannot supply four of these seven, the unit price is not comparable, because you cannot compute the denominator. A bid without a puncture rating and a rerun trial is a bid on the numerator only.

Where the model flips: when the cheap grade actually wins

I am not going to pretend the premium grade always wins. It does not. The model flips when the puncture count per cap is low and the method is forgiving.

If your autosampler punctures each cap once, your method has a system-suitability check that catches failures early, and your rerun rate on the standard grade is already under 3%, the premium grade cannot recover its 87% price premium. In that case the standard grade is the rational buy, and the money is better spent on a lot-level extractables screen than on the vial. The crossover in my models sits at roughly 4 punctures per cap and a standard-grade rerun rate of about 5%. Below both, buy standard. Above either, buy premium and stop negotiating the unit price.

There is one more input that flips the model hard: holding time. If your method holds samples for 7 days at 4°C before analysis (common on volatile organics), seal integrity matters more than puncture count, and a cap that loses 0.1 g over a week will produce rejects no matter how cheap it is. In that case the grade decision is driven by the weight-loss test, not the puncture rating. This is where a supplier who will run the 24-hour weight-loss check on your storage temperature earns the premium, and one who will not should be dropped from the bid. The vial supplier comparison pages are useful for finding suppliers who publish that data rather than promising it.

The decision rule

Compute cost per reported result as (consumable + prep labor + instrument time on failed runs + rework consumable) divided by accepted results, and require every bid to carry a puncture rating, a septum hardness, a lot-level extractables screen, and a 100-injection rerun trial. Buy the grade whose rerun rate and prep time, measured on your method, produce the lower cost per reported result, not the lower unit price. If the standard grade holds a rerun rate under 5% and the cap sees fewer than 4 punctures, buy standard and put the savings into lot screening. If either threshold is exceeded, the premium grade is cheaper per reported result even at 87% more per unit, and the unit price conversation is over.

References & further reading

Related resources from our lab consumables network:

• Chromatography Products at chromatographyselling.com
• Aijiren at aijirentech.com
• Vials With Caps at vialswithcaps.com
• Vial Supplier at vialsupplier.com

External references and standards:

• Agilent chromatography consumables and columns
• Thermo Fisher chromatography and sample handling
• U.S. EPA SW-846 test methods for waste and materials
• Restek technical resources for chromatography