Calibration discipline
Measuring equipment runs on a documented calibration interval with expiry tracked per instrument. A panel cannot be released from an out-of-calibration device — the pipeline blocks it rather than flagging it afterwards.
An analytics lab built around evidence, not paperwork.
SPMLABO operates from Mwiki in Kasarani, Nairobi, serving manufacturers, exporters, distributors and buyers who need a product claim settled with data.
Our premise
Most product disputes are really data disputes.
A consignment gets rejected, a batch underperforms, a competitor alleges your label is wrong. In almost every case the argument is not about opinion — it is about what was actually in the material, measured how, against which limit.
We set up SPMLABO to close that gap. Samples are logged under chain of custody at intake, mapped to a defined analytical method, and run on instrumented lines that write results straight into the record. The report you receive carries the method, the calibration state, the tolerance used and the variance measured — so the conclusion arrives with its own supporting evidence attached.
The name reflects the operating model: SPM — statistical product metrology — delivered through automated analytics rather than manual transcription.
facilityMaina wa Maguru Road
suburbMwiki, Kasarani
cityNairobi, Kenya
hoursMon–Sat · 08:00–18:00 EAT
contact[email protected]
record modelinstrument-direct, versioned
Quality & compliance
The controls behind every issued number.
Our quality system rests on four practices. We describe them plainly because "we follow good practice" is not something an auditor can verify.
Chain of custody
Every sample carries a unique identifier from receipt through each bench step to report issue. Receipt condition is recorded, and a retained duplicate is held so a disputed result can be re-measured rather than merely re-argued.
Method traceability
Where a recognised standard method exists for the analyte, we use it and cite it. Where we have had to adapt one, the deviation is stated in the report so a reader can judge its effect instead of guessing.
Versioned reporting
Reports are immutable once issued. A correction produces a new version with a stated reason for the change, so the history of a result stays visible instead of being quietly overwritten.
The cut test
A defect count is a judgement, so we make it repeatable.
Slaty, mouldy, insect-damaged, germinated and broken beans are each scored from a counted subset on a gridded board, against the grade limits stated in the report. The board layout, the count and the acceptance limit all appear in the issued record, which is why a buyer can re-open the same report months later and check the arithmetic.
How we work
Four commitments we hold ourselves to.
The verdict must be defensible
A result you cannot explain to an auditor is a result with no value. Every verdict we issue is traceable to a method, an instrument state and a named tolerance — all of which we will show you on request.
Automation removes transcription, not judgement
Instruments write directly into the record, so numbers cannot be mistyped between bench and report. Interpretation, however, stays human: a scientist signs every report and is accountable for it.
We say when analytics cannot help
If the claim you want to make is not testable with the sample you have, we tell you at intake. Declining work honestly is the reason buyers keep sending us the next batch.
Turnaround is a specification like any other
A correct report that arrives after your shipment is released has failed. Timelines are committed at intake and a delay is communicated before it becomes a surprise.
Location
Find us in Mwiki, Kasarani.
Sample drop-off is available during working hours on Maina wa Maguru Road. Dispatched samples are logged the day they arrive.