Home specimen collection: the real unit economics

Consumer lab testing looks like a software business with a shipping problem. It is the reverse: a logistics business with a software layer. The price of the assay is rarely what decides whether a product is viable — the cost per completed result is.

Four collection models, four cost shapes

Home kit (self-collected). Cheapest at scale, highest failure rate. A dried blood spot or saliva kit costs a few dollars to ship, but a meaningful share of kits are never returned, and a further share arrive with insufficient or haemolysed samples. You pay for the kit twice: once to ship, once to recollect.

Draw centre. Highest reliability, lowest flexibility. A partner site draws the sample with trained staff, so rejection rates fall sharply. But the patient must travel, which cuts conversion for convenience-driven products, and the site charges a collection fee that is often larger than the assay price.

Mobile phlebotomy. Best experience, worst unit economics at low volume. A phlebotomist travels to the patient. Per-visit cost dominates, so this model only works when the basket value is high — a $189 panel, not a $29 single test.

Hybrid. Offer a kit by default, escalate to a draw when the kit fails or the assay requires serum. Most successful consumer products land here.

The number that matters: cost per completed result

Work the arithmetic on a $99 panel:

Line	Kit model	Draw centre
Assay (lab cost)	$34	$34
Collection	$6 (kit + outbound shipping)	$22 (site fee)
Return logistics	$8 (prepaid label)	—
Rejection / non-return rate	18%	4%
Effective cost per completed result	≈ $58	≈ $58

The two models converge, which is the point: cheap collection plus high failure equals expensive collection plus high reliability. Teams that quote the assay cost as their COGS understate the real number by roughly 70%.

Where margin actually comes from

Panel design. Bundling raises basket value without proportionally raising collection cost. One draw for twelve analytes costs almost the same to collect as one draw for three. This is why panels, not single tests, carry consumer lab businesses.

Routing. Sending every order to the same laboratory is simple and wasteful. Routing by region, assay availability and current capacity cuts both turnaround and cost. A 24-hour result is worth more to the customer than a 72-hour one at the same price.

Recollect policy. A free recollect feels generous until you measure it. The right design charges nothing to the customer but tracks recollect reasons obsessively — most rejections are caused by three or four fixable steps in the collection instructions.

Payment mix. Card processing on a $99 charge costs meaningfully more than a USDC settlement, but crypto introduces its own friction and refund complexity. The practical answer is to offer both and let the customer choose; the margin difference is real but secondary to conversion.

What to instrument from day one

You cannot manage this model on invoices. Instrument four events per order: kit shipped, sample received, result final, and — the one teams forget — collection failure with a reason code. Without the reason code you will keep paying for the same mistake.

Then track the funnel as rates, not counts:

• Ship → return rate (by kit type, by region)
• Return → accepted rate (by assay, by collection method)
• Accepted → final rate, and median hours in each state
• Recollect rate, and cost per recollect

The uncomfortable conclusion

Home collection does not remove cost, it moves it. What it removes is the visit, and with it the ceiling on how many people will complete a test. A product that converts 4% of interested users through a draw centre may convert 20% through a kit, even with an 18% failure rate — and that difference in completed results is worth far more than the delta in cost per sample.

Design for completed results, not for cheapest shipment.