EMP BUDGET

What an Environmental Monitoring Programme Costs a South African Food Plant, Line by Line

By Mthokozisi Nkosi, Food Safety Specialist & Lead Auditor, ASC Food Safety · 25 min read

An environmental monitoring programme (EMP) costs what your own numbers multiply to: sites per sampling event, times events per month, times 12, times your laboratory’s price per test, plus materials, courier, sampler time and the follow up a positive result triggers. In the illustrative medium ready to eat plant below, laboratory tests were about 69% of an R518 110 (about USD 31 835 · EUR 27 614 · GBP 23 644 · AED 116 558) year. ASC’s Automated EMP tool (R1 450, about USD 89 · EUR 77 · GBP 66 · AED 326, per site per year) and RA19 template (R690, about USD 42 · EUR 37 · GBP 31 · AED 155) are the smallest lines and make the rest of the spend count.

How do you calculate what an environmental monitoring programme costs?

Multiply the number of sites you swab per sampling event by the events per month, then by 12, then by your laboratory’s quoted price for that test. Do it separately for each zone and each test, then add materials, courier, sampler time, investigations after positives, training, documentation and software. That total is your annual EMP budget.

A single monthly figure with no arithmetic behind it cannot be checked. “10 zone 1 sites, weekly, for Listeria spp.” can be checked by your finance manager and by your auditor.

The core formula for each test line is:

• Annual tests = sites per event x events per month x 12
• Annual test cost = annual tests x your laboratory’s quoted price for that test
• Samples to buy = the sum of annual tests across lines, unless your laboratory runs more than one test from one sponge (ask them; it changes the materials line)

What drives the size of the bill?

Four decisions set almost the whole bill: how many sites you monitor in each zone, how often, which organisms you test for, and how many positives you find and chase. Each should come from a documented risk assessment of your product and process, which is what certification schemes and export markets expect to see.

A plant making ready to eat food that supports the growth of Listeria monocytogenes will spend more than a dry bakery, and it should.

The requirements behind the spend

FSSC 22000 Version 7 clause 2.5.7 requires a risk based environmental monitoring programme for the relevant pathogens, spoilage and indicator organisms, a documented procedure to evaluate the effectiveness of controls against contamination from the manufacturing environment, data with regular trend analysis, and a review at least annually and more often when triggers occur (FSSC 22000 Scheme Version 7). Version 7 was published in May 2026, and audits against Version 6 are allowed up to 30 April 2027 (FSSC upgrade process). Clause 2.5.1 (a) expects verification analysis to be done by a competent laboratory. BRCGS Food Safety Issue 9 clause 4.11.8 and SQF Edition 9 Food Manufacturing clause 2.4.8 also set environmental monitoring requirements; SQFI expects Edition 10 audits to start on 2 January 2027 (SQFI).

If you export, the importing market adds its own expectations. Article 5(2) of EU Regulation (EC) No 2073/2005 requires producers of ready to eat foods that may pose a Listeria monocytogenes risk to sample processing areas and equipment for L. monocytogenes, and names ISO 18593 as the reference method for that sampling (EUR-Lex). In the United States, 21 CFR 117.165(a)(3) requires environmental monitoring for an environmental pathogen or an appropriate indicator organism where contamination of a ready to eat food is a hazard requiring a preventive control, and 117.165(b)(3) requires the procedure to name the organisms, sites, timing, frequency, tests and laboratory (eCFR). That list is, in effect, your list of cost drivers.

At home, Regulation R638 of 22 June 2018 under the Foodstuffs, Cosmetics and Disinfectants Act 54 of 1972 is the food hygiene law, enforced by municipal environmental health practitioners. It does not require an environmental monitoring programme by name, but regulation 6(4)(b)(i) says a food contact surface must not carry more than 100 viable micro-organisms per cm² before food touches it, sampled with the SANS 5763 swab technique (R638, gov.za). That makes surface counts a legal line in every food plant’s budget. Codex CXG 61-2007 on the control of Listeria monocytogenes is not law in South Africa on its own, but its Annex I on environmental monitoring is a widely used reference. For design basics, read our microbiological risk assessment and EMP guide; this article stays on the money.

What are the cost lines in an EMP budget?

An EMP budget has ten recurring lines: sampling materials, laboratory tests, courier and cold chain, sampler time, investigative sampling, product hold and testing, training, documentation, software and outside expertise. The cost of not monitoring never appears on the budget until it appears on the balance sheet.

Sampling materials

Sponges suit large, uneven surfaces such as frames, floors and drains; stick swabs reach small gaps. ISO 18593:2018 describes horizontal methods for surface sampling with contact plates, stick swabs, sponges and cloths (ISO), and the EU names it as the reference method for processing area sampling. Neutralising broth is cheap insurance, as the waste section explains. Our training site has a practical surface swabbing guide on sponges, swabs and neutralising broth.

Laboratory tests

Pathogen detection gives a present or absent result per sampled area. Indicator counts such as Enterobacteriaceae or aerobic plate count show how clean a surface is and suit cleaning verification. The common reference methods are ISO 11290-1:2017 for detection of Listeria monocytogenes and Listeria spp. (ISO), ISO 6579-1:2017 for Salmonella detection (ISO), ISO 21528-2:2017 for Enterobacteriaceae colony counts (ISO) and ISO 4833-1:2013 for colony counts at 30 °C (ISO). Ask which method sits behind each price and whether it is on the laboratory’s SANAS schedule of accreditation.

A presumptive positive usually needs a separately billed confirmation. Typing of isolates, including whole genome sequencing (WGS), is an occasional line. Courier costs depend on distance and dispatch frequency, and sampler and data time are real money even when nobody invoices them.

What do laboratory tests cost in South Africa?

Many South African food laboratories quote on request rather than publishing current tariffs, so a written quotation for your own tests, devices and volumes is the only price worth budgeting on. The price per test moves with the organism, the sampling device, the method, the accreditation scope, the turnaround and how many samples you send.

Published price lists date quickly, so the worked example below uses labelled assumptions rather than any laboratory’s tariff. Two quotes with similar totals can buy very different results, so compare them line by line:

Accreditation scope matters as much as price. FSSC 22000 clause 2.5.1 (a) expects analysis that verifies food safety to be done by a competent laboratory using validated methods, with accreditation such as ISO/IEC 17025, and a cheap result from a test outside the laboratory’s scope can cost you an audit finding.

Worked example: what does an EMP cost a medium ready to eat plant?

Our illustrative medium ready to eat plant runs 1 212 Listeria spp. tests, 240 Enterobacteriaceae counts and 48 Salmonella tests a year. With the assumed prices below, the programme costs R518 110 (about USD 31 835 · EUR 27 614 · GBP 23 644 · AED 116 558) a year, about R43 176 (about USD 2 653 · EUR 2 301 · GBP 1 970 · AED 9 713) a month, and laboratory tests are 69% of it. Swap in your quotes and site counts.

The plant is hypothetical: one site making chilled ready to eat products, with a high care line, a raw side and dispatch. Each “event” is one sampling round, and weekly sampling is counted as four events a month.

Step 1: routine laboratory tests

Rand amounts in this table convert at roughly USD 61, EUR 53, GBP 46 or AED 225 per R1 000 at September 2026 rates.

Step 2: everything else

Rand amounts in this table convert at roughly USD 61, EUR 53, GBP 46 or AED 225 per R1 000 at September 2026 rates.

Step 3: the total and what it tells you

Adding the lines: R355 800 (about USD 21 862 · EUR 18 963 · GBP 16 237 · AED 80 043) + R63 000 (about USD 3 871 · EUR 3 358 · GBP 2 875 · AED 14 173) + R21 600 (about USD 1 327 · EUR 1 151 · GBP 986 · AED 4 859) + R22 500 (about USD 1 382 · EUR 1 199 · GBP 1 027 · AED 5 062) + R24 000 (about USD 1 475 · EUR 1 279 · GBP 1 095 · AED 5 399) + R17 400 (about USD 1 069 · EUR 927 · GBP 794 · AED 3 914) + R6 250 (about USD 384 · EUR 333 · GBP 285 · AED 1 406) + R3 500 (about USD 215 · EUR 187 · GBP 160 · AED 787) + R690 (about USD 42 · EUR 37 · GBP 31 · AED 155) + R1 450 (about USD 89 · EUR 77 · GBP 66 · AED 326) + R1 920 (about USD 118 · EUR 102 · GBP 88 · AED 432) = R518 110 (about USD 31 835 · EUR 27 614 · GBP 23 644 · AED 116 558) a year, or about R43 176 (about USD 2 653 · EUR 2 301 · GBP 1 970 · AED 9 713) a month. Laboratory tests are 68.7% of the total, materials 12.2%, and data capture alone is 4.6%. Training, the risk assessment, the software and the consultation together come to R7 560 (about USD 465 · EUR 403 · GBP 345 · AED 1 701), or 1.5%.

That split is the point. A plan built on a weak risk assessment, run by untrained samplers and trended by nobody still costs R355 800 (about USD 21 862 · EUR 18 963 · GBP 16 237 · AED 80 043) in laboratory fees. It just tells you less.

What does a positive result cost?

A positive result costs the investigative sampling around the site, confirmation tests, any product held and tested, cold storage while it waits, and staff time for root cause and corrective action. A positive on a food contact surface costs the most because product is involved; one found early in zone 3 is cheap by comparison.

Codex CXG 61-2007 Annex I describes responses ranging from no action, through intensified cleaning and source tracing with more environmental testing, up to holding and testing product (Codex CXG 61-2007). Each step adds cost, so define alert, action and escalation levels in advance.

Investigative (star burst or vector) sampling

Sample outward from the positive site: adjacent surfaces, connected equipment, the floor and drain below, and nearby traffic routes, each tested individually. Cost it as extra samples x price per test, per round, until your clearance rule is met. We allowed 15 samples per event; a complex slicer can need more.

Hold and test

If a zone 1 positive places product on hold, you pay for product tests, cold storage, delayed dispatch and disposal of anything that fails.

Typing of isolates

When a site keeps returning positives, typing can show whether one strain has settled in or new strains keep arriving. The first can point to a harbourage point that cleaning is not reaching; the second to ingress through people, materials or traffic. Ask your laboratory about typing and its price before you need it.

Where is EMP money wasted?

Money is wasted when tests are run where they cannot find anything useful: weekly pathogen tests on zone 4, swabs taken only on freshly cleaned surfaces, composites used for follow up, devices without neutraliser, and results that nobody trends. Each costs the same per test as a useful sample.

Rand amounts in this table convert at roughly USD 61, EUR 53, GBP 46 or AED 225 per R1 000 at September 2026 rates.

Why swabbing only clean surfaces wastes money

Swab when the organism has had time to work out of harbourage points. Guidance from the Grocery Manufacturers Association says zone 1 to 3 sampling should take place at least three hours after production begins and should be varied across times and shifts (GMA, 2014). Post clean indicator swabs still verify cleaning, but that is a different purpose.

Composites: a saving that can backfire

Codex Annex I says that under certain circumstances samples can be composited without losing the required sensitivity, but a positive finding then needs more testing to locate the positive sample. GMA guidance allows up to five sponges in one sample, only in mature programmes where positives are rare, and never mixing product contact with non product contact samples. The saving disappears the moment a composite is positive.

No neutraliser, no result you can trust

A sponge picks up sanitiser residue with the bacteria, and without a neutraliser that residue can keep acting in transit. Li and colleagues described an effective environmental sampling method as one using a transport broth that neutralises sanitiser residues and keeps stressed Listeria monocytogenes cells viable; in their trial, D/E, Letheen and HiCap broths performed similarly with no quaternary ammonium compound or 400 ppm of it on the surface, but Letheen broth did not neutralise 8 000 ppm (Li et al., 2020, BMC Microbiology). The GMA guidance also says a buffer containing a neutralising agent should be used if sanitiser residues are present. Match the broth to the chemistry in each area and record it.

Paying for tests nobody trends

FSSC 22000 clause 2.5.7 (c) asks for regular trend analysis. A result sitting in a laboratory email is a paid for test that has not done its job yet.

Where should EMP money go?

Spend where the organism is most likely to be found before it reaches food: zone 2 and zone 3 sites during production, investigative sampling after every positive, and the time to review trends every month. Zone 1 testing verifies your controls; zones 2 and 3 are where you catch problems early. Budget investigations in advance, or plants re-swab only the positive site, get a negative and close the case while the harbourage point remains.

Zones 2 and 3 during production

Equipment frames, conveyor undersides, drains, floor junctions and door seals near exposed product are typical harbourage points. A positive there costs a cleaning response and some vector samples, not a product hold. Testing zones 2 and 3 for Listeria spp. as an indicator is common practice. United Fresh guidance calls Listeria spp. an appropriate indicator for L. monocytogenes, while noting a Listeria spp. finding does not mean L. monocytogenes is present (United Fresh, 2018).

Drains and harbourage

Drains are sampling sites and cleaning points, and they can also harbour pests such as drain flies and cockroaches, so a drain that keeps returning positives may need pest control and maintenance attention as well as more cleaning.

What does it cost not to run an EMP?

South Africa has a measure of it. The 2017 to 2018 listeriosis outbreak caused 1 060 cases and 216 deaths where the outcome was reported and was traced to a ready to eat processed meat product and its manufacturing environment. The manufacturer’s parent company reported R365 million (about USD 22.4 million · EUR 19.5 million · GBP 16.7 million · AED 82.1 million) in recall costs, net of insurance claims, for six months, and researchers put the wider economic cost far higher.

Those case and death figures come from WHO’s regional office for Africa, citing the NICD; the outbreak ran from its declaration on 5 December 2017 to 3 September 2018 (WHO Africa). WHO’s March 2018 outbreak notice reported that 91% of the strains were L. monocytogenes sequence type 6, found in polony, a ready to eat processed meat product, and in the manufacturer’s processing environment, and that the company and three of its retailers exported to 15 countries in the African region, all of which issued recalls (WHO).

National Institute for Communicable Diseases (NICD) whole genome sequencing work published in 2019 found that outbreak strain isolates from the Enterprise Foods production facility and its products were highly related to the clinical isolates (Smith et al., 2019). In its results for the six months to 31 March 2018, Enterprise Foods’ parent company Tiger Brands reported costs associated with the recall of R365 million (about USD 22.4 million · EUR 19.5 million · GBP 16.7 million · AED 82.1 million), net of insurance claims submitted to date and excluding ongoing trading losses (Tiger Brands interim results, 2018). A cost study in Food Control estimated fatality costs of over US$260 million, hospitalisation costs of about US$10.4 million and productivity and export losses of more than US$15 million, and did not include costs such as plant sanitisation (Olanya et al., 2019).

Set the illustrative R518 110 (about USD 31 835 · EUR 27 614 · GBP 23 644 · AED 116 558) a year against those figures. An EMP guarantees nothing, but it gives you the chance to find a harbourage point while the fix is a drain repair and a round of vector swabs.

Which EMP cost line gives the biggest return?

The software line. At R1 450 (about USD 89 · EUR 77 · GBP 66 · AED 326) per site per year, the Automated EMP tool is 0.3% of the worked example, yet it replaces the spreadsheet work of scheduling, logging, alerting and trending. If it cut data time from 8 to 3 hours a month in your plant, the time saved would be worth R15 000 (about USD 922 · EUR 799 · GBP 685 · AED 3 375) at R250 (about USD 15 · EUR 13 · GBP 11 · AED 56) an hour.

That is an illustrative assumption: 96 hours at R250 (about USD 15 · EUR 13 · GBP 11 · AED 56) is R24 000 (about USD 1 475 · EUR 1 279 · GBP 1 095 · AED 5 399), 36 hours is R9 000 (about USD 553 · EUR 480 · GBP 411 · AED 2 025), a difference of R15 000 (about USD 922 · EUR 799 · GBP 685 · AED 3 375). The bigger value is harder to price: a positive that triggers an alert email instead of sitting in an inbox, and a repeat site that shows on a chart before it shows in an audit finding.

What the tool does

• Builds the RA19 style risk register with 5 x 5 inherent and residual scoring, and sets frequency and a four week, shift rotated schedule with a swab budget
• Sets alert, action and escalation levels, hold and release rules, a star burst (vector) positive drill, CAPA and an action tracker
• Logs every result (date, site, zone, test, result, lab reference), with CSV paste import
• Draws trend charts: monthly positive rate by zone, repeat site list, positives by area, and count results against your own alert and action limits (aerobic plate count defaults: alert above 10 and action above 100 per cm², the R638 ceiling)
• Grades each result as it is saved and emails flagged results to up to ten addresses: a pathogen on zone 1 is an Action for Listeria spp. and an Escalation for Listeria monocytogenes, Salmonella or Cronobacter; zone 2 detections are Actions; zone 3 and 4 detections are Alerts; a repeat at the same site within 12 months moves up one level
• Reminds you of the annual review, flags long runs without positives and repeat detections, and logs other clause 2.5.7 (d) triggers
• Exports an RA19 style Excel workbook for auditors

Compare it with spreadsheets in EMP software vs spreadsheet, and see common EMP audit nonconformities.

How do you build your own EMP budget?

Start from the risk assessment, count sites and frequencies per zone, get written laboratory quotes, then run the formula for each test line. Add materials, courier and people time, set aside money for investigations and holds, and review the budget against your trend data every year.

1. Complete the risk assessmentScore each zone, site and organism with RA19 or the Automated EMP tool.
2. Fix the site list and frequencyRecord sites per event and events per month for every zone and test.
3. Get two written quotesFrom ISO/IEC 17025 accredited laboratories, per test, per device, with confirmation, courier and turnaround stated.
4. Run the formulaSites x events x 12 x price for each line, checked by a second person.
5. Add the hidden linesMaterials, cold chain, sampler hours, data hours, training, documentation and software.
6. Ring fence investigationsEstimate positives from your history and cost what each would trigger.
7. Review with the dataAt the annual review, move money to the sites that matter and record why.

Frequently asked questions

Published by ASC Food Safety, South African food safety and quality consultants. This article is general guidance and not a substitute for site specific advice.