Kitchen Exhaust Supplier Audits: A Weighted Scorecard Method

Most kitchen exhaust suppliers are chosen on three inputs: a demonstration, a datasheet and a price. The three things that decide the next five years of cost, namely duty-point performance in writing, spare parts supply and documentation quality, are usually the three nobody scores. A weighted scorecard repairs that gap, because it makes every mark point at a document you can attach to the purchase order.

What the Scorecard Is Actually For

A score is not a verdict on a company. It is a dated record of what you were shown, with a name against it. Ten months later, when a cell fails in month four of a promised eight-month life, that record is the difference between being misled and being under-informed. Buyers who skip the exercise tend to rediscover the same three failures in every project: the quoted airflow was a free-air figure that the installed duct run never delivered, the wear part nobody priced has a six-week lead time, and the compliance file arrived after the deposit rather than before it.

Three outputs are worth expecting from the process. First, a shortlist of two vendors with a written reason for excluding the others. Second, a list of clauses that come directly out of the scores, which is what turns an evaluation into a purchase order. Third, a baseline you can re-measure in twelve months, when claims meet operating hours. If your current comparison produces only the first output, it is a shortlist, not an evaluation.

When a Demonstration Tells You Nothing

Demonstrations are run at one fixed condition. Filters are clean, one appliance is fired, airflow is at its highest, the odour load is decorative, and the room temperature is comfortable. That is a fair test of two things only: whether the hood captures smoke at its face, and whether the controls respond the way the operator expects. Both matter. Neither is a performance guarantee.

A demonstration cannot tell you the airflow at the static pressure your duct run will impose, how the electrostatic stage behaves once the collector plates carry a grease load, how long the carbon bed lasts at your frying volume, whether the cell comes out without dismantling ductwork, or whether the documentation pack will satisfy the inspector. Those five questions are answered on paper or not at all, which is why the scoring discipline below is built around evidence classes rather than impressions.

Four Evidence Classes, Ranked

Every mark on the scorecard should be tagged with the class of evidence behind it. When two vendors score within a few points of each other, the class of evidence usually breaks the tie.

Rule: a performance claim supported only by class D scores zero, however confidently it is delivered in the meeting.

Gate Questions: Pass or Fail, Never Averaged

Before any weighting, run eight binary gates. A gate is not a criterion, so it never carries points. A weighted total can bury a fatal flaw: a bid that scores 74 out of 100 and cannot state its duty point is not 74 percent acceptable, it is unusable, because you have no way to check delivery on site.

1. The purification stage has a type-test report with a number you can request.
2. A declaration of conformity exists and names the standards applicable in the destination market.
3. The duty point is stated as airflow in cubic metres per hour at a stated external static pressure in pascals, not as a maximum airflow figure.
4. The electrostatic cell can be removed and refitted by one person with standard tools, through an access panel, without cutting ductwork.
5. Each unit ships with a serial-numbered test sheet.
6. A written critical spares list exists with part numbers and lead times.
7. Voltage and frequency are confirmed in writing for the destination, whether that is 230 V at 50 Hz, 120 V at 60 Hz, 100 V at 50 or 60 Hz, or a 127 and 220 V dual arrangement.
8. A pre-shipment inspection, whether yours or a third party’s, is accepted in the contract.

Record gate failures separately in the evaluation file. A vendor excluded at gate 3 in March is often the one who returns in September with a fan curve, and that is useful intelligence rather than a permanent blacklist.

The Weighted Criteria

The weights below sum to 100 and reflect lifetime cost rather than first cost. Adjust them for your own risk profile, but keep the total at 100 and commit to it before the bids arrive, otherwise the weights quietly bend towards whichever vendor presented best.

Scoring Each Criterion With a Question You Can Send

Vague criteria produce vague marks. Each item below is written as a question that can go into a tender clarification or an email, with the answer form that earns full marks.

Duty point, 20 points

Ask for the airflow in cubic metres per hour at the quoted external static pressure, the fan curve showing that point, and the motor power absorbed at the same point. An answer of maximum airflow with no pressure is a free-air figure, measured with nothing attached to the outlet. The same unit can lose a fifth of that figure on a real duct run. On a 2,000 cubic metre per hour hood, a 15 percent shortfall is 300 cubic metres per hour, which is roughly the capture margin for the last metre of canopy.

Full marks require three things together: a fan curve, a marked operating point, and a factory test sheet from a unit of the same configuration. Two vendors quoting identical maximum airflow can differ by 20 to 30 percent at 250 pascals, and that difference shows up as smoke at the hood edge on a busy Friday rather than as a line on a quotation.

Construction and materials, 15 points

Ask for the sheet gauge of hood and cabinet bodies, the collector plate material and thickness, the plate spacing in millimetres and the plate count per cell, the insulator material, the seal strip specification and the fastener type. Then ask which internal surfaces a cleaner can reach with a hand tool through the largest access panel.

Plate spacing deserves its own line in the file. A narrower gap raises collection efficiency but clogs faster and demands a better wash routine; a wider gap survives a rough cleaning regime at some cost in capture. Neither is wrong, but a supplier who cannot state the number has not designed for maintenance. Insulator material matters for the same reason: ceramic insulators tolerate heat and cleaning chemicals, and polymer substitutes track and fail in wet kitchens. Ask how insulators are stored at the factory, because tracking failures usually begin in a dusty store rather than in service.

Odour stage design, 12 points

Ask for carbon mass per stage in kilograms, bed depth in millimetres, the airflow passing that stage, and the replacement interval the supplier will put in writing, qualified by a stated grease load. Then compute contact time yourself as bed volume divided by airflow in cubic metres per second. The arithmetic takes a minute and separates a designed bed from a token tray.

Two further questions decide whether the interval is credible. What protects the carbon from grease, since a blinded carbon bed stops adsorbing long before it is saturated, and what is the disposal route for spent carbon in the destination market. A supplier who answers all three has a product; one who answers only the first has a brochure.

Spares and serviceability, 15 points

Ask for a critical spares list of roughly a dozen items with part numbers and lead times in weeks: cell, collector plate set, insulator set, seal strips, high-voltage power supply, control board, fan motor, bearings, filter media, carbon charge, access panel gaskets, door interlock. Then ask what happens in year six, when a cell model may no longer be in production.

Full marks require a five-year availability statement and a last-time-buy notice of at least twelve months. There is also an operating test that most buyers miss: compare the cell lead time with the cleaning interval. If a cell takes six weeks to arrive and the kitchen washes cells every three weeks, a single spare cell must sit on site, and that spare belongs in the quotation rather than in next year’s emergency budget.

Documentation and compliance file, 10 points

Ask for the declaration of conformity with the standards listed by number, the type-test report number, the wiring diagram, the maintenance manual in the local language, and the cleaning instruction with an interval. Add the noise data with the measurement distance and conditions, plus unit weight and lifting points for the installation contractor.

Score the file as delivered at handover, not as promised at tender. A missing declaration delays the inspection, and an inspection delay delays the first service, which on a lease with a fixed opening date is the expensive kind of delay.

Manufacturing and quality control, 10 points

Ask which fields appear on the serial-numbered test sheet, whether testing is 100 percent or sampled, how instruments are calibrated and how often, and how a serial number traces back to build date and inspector. In a factory making both light and heavy duty cabinets, also ask how the two are kept apart on the line, because a mixed build is where specification drift starts.

Commercial terms and references, 18 points

Score warranty scope, which wear items are excluded, response times for safety-related and odour-related faults, freight terms, payment structure and retention. Then score references, but only references at comparable duty: three sites in your market, with a similar cooking load and a similar operating pattern. A reference from a light-fume cafe says nothing about a fryer bank running fourteen hours a day.

The Factory Visit: Ten Observations in Two Hours

A visit is not a tour. Walk the line with a list and photograph each answer, because the scorecard needs evidence class C and your memory will not survive the flight home.

1. Sheet gauge: measure incoming material on the rack with a micrometer, do not read the label.
2. Cell assembly: is there a jig, and are plates aligned and gapped consistently across the stack.
3. Insulator storage: sealed, dry, dust-free, or open on a shelf near the welding bay.
4. Cabinet interior: welds dressed or rough, and whether weld spatter could trap grease.
5. Power supply burn-in: is there a rack, and for how many hours before packing.
6. Electrical test station: earth and leak tests performed and logged, or signed off on trust.
7. Instrument calibration: stickers with dates on the airflow and electrical test gear.
8. Packing: foam and edge protection, timber grade, and a loading plan that keeps units upright.
9. Finished goods: how many units, and how long they have waited, which tells you real lead time.
10. Complaint log: what failed in the last six months and what changed as a result.

Two hours on the floor will tell you more about spares and documentation than a week of correspondence, because both are visible as habits rather than as claims.

Normalizing Price So It Cannot Dominate

Price is the easiest number to compare and therefore the most dangerous one. Three steps keep it in proportion.

Step one, strip the scope. Duct length, freight, commissioning, spares kit, number of filtration stages, controls and interlock scope all vary between bids. Bring every quotation to the same scope before any figure is scored, and put the differences in a written assumptions register that each vendor signs.

Step two, convert to five-year consumables and energy. Four numbers per bid are enough: fan power absorbed at the duty point, carbon mass and replacement interval, cell washing interval and cost of a wash, and scope excluded from the price. A worked comparison makes the point. Two bids for a heavy-fume kitchen look 12 percent apart on the equipment line. Bid A carries 25 kg of carbon replaced every three months, which is 100 kg a year, with four service visits. Bid B carries 40 kg replaced every six months, which is 80 kg a year, with two visits. Add the two extra visits and the 20 kg of extra carbon and the gap narrows to single digits, at which point the duty-point evidence, not the price, should decide.

Step three, cap the price weight. Give price 10 to 15 points inside the same 100, and score total five-year cost rather than unit price. A price weight above 25 reliably produces the cheapest compliant-looking bid rather than the cheapest outcome.

Reading the Score: What the Bands Mean

Re-Scoring at Twelve Months: Claim Versus Actual

The first year of operation is the only real test, and it is cheap to run if you captured a baseline. Measure five quantities and compare them with what was scored.

The difference between claim and actual is your supplier development agenda for year two, and your leverage at renewal. It is also the most useful document you can hand a new supplier when you ask them to beat the incumbent.

Seven Scoring Errors

1. Averaging gates into the total, which lets a fatal compliance gap hide behind strong references.
2. Scoring the salesperson. Fluency in the meeting is not evidence of anything on the list.
3. Treating familiarity as merit. A brand you have used before earns points only where you can cite a measured outcome.
4. Accepting the word complies as an answer. Ask for the document, the number, or the units.
5. Forgetting destination conditions: voltage, frequency, ambient temperature, documentation language, and altitude where relevant.
6. Scoring the unit but not the interface: duct connection detail, electrical terminal and isolator, make-up air provision, condensate or grease drainage.
7. Anchoring on the cheapest bid, after which every other bid is read as expensive rather than as different.

From Score to Clause: Six Clauses That Mirror the Criteria

Evaluation that stops at a spreadsheet changes nothing. Each high-weight criterion should reappear as a clause, and each clause should name the evidence due at handover.

The One-Unit Version: Twelve Questions in Thirty Minutes

Buyers replacing a single hood do not need a tender process, but they need the same information. Score each question 0, 1 or 2, for a maximum of 24, and treat 18 as the threshold to proceed.

1. What is the airflow at the static pressure this installation will impose?
2. Can you send the fan curve with the operating point marked?
3. What is the cell plate spacing, and how many plates per cell?
4. What are the insulators made of, and how are they replaced?
5. How many kilograms of carbon are in the odour stage, and what interval do you warrant?
6. What is the lead time for a replacement cell, in weeks?
7. What does a spare cell cost relative to the unit, as a percentage?
8. Which document proves the purification efficiency, and what is its number?
9. Is the declaration of conformity valid for my market and my voltage?
10. Who measures the installed airflow, and to what tolerance?
11. What is excluded from the price, in writing?
12. Can I speak to two customers running the same menu load?

Where Suppliers Usually Come Up Short

Run this process across a shortlist and the gaps cluster in the same three places: duty-point evidence, spares lead times and documentation released before payment. That pattern is not accidental. A supplier who can answer all twelve questions in writing is usually the one whose architecture is simple enough to document, which is why the ductless all-in-one format tends to score well on serviceability: one cabinet carries the electrostatic cell, the filter stage and the carbon bed, and the cell comes out through a front panel without touching ductwork. Where the building can support a full duct run and the volume is high, the ducted ESP purifier range keeps the same cell architecture inside a duct spool, and the ductless range hood family covers the light and heavy fume cases with a paired odour cabinet. Where there is no fixed cookline at all, ventless catering carts are scored on exactly the same twelve questions, which is a useful test of whether a vendor has one evaluation standard or several.

What full marks look like on paper is unglamorous: a light-fume all-in-one such as the N-series returns a fan curve, a serial test sheet, a stated carbon mass and a cell lead time in weeks; a heavy-fume unit such as the ND-series does the same at a higher duty point, with the paired Odor Removal Tower declaring its carbon charge separately. Three documents, one page each, and the comparison becomes arithmetic rather than opinion.

A Scorecard Is a Contract in Draft

Every criterion you score should reappear as a clause, and every clause should reappear as a document at handover. When those three line up, the question of which supplier was better stops being a matter of remembering a demonstration and becomes a matter of reading a file. Buyers who keep that file through the first year of operation usually find the second purchase takes an hour, because the only new information is how the first year actually went.