Part 2: From Criteria to Confidence – Assessing Without the Guesswork (or the Grudges)

Part 2: From Criteria to Confidence – Assessing Without the Guesswork (or the Grudges)

From the Desk of Henning Lubbe Education Lead | SCASA Board Directors


Where We Left Off

Two weeks ago we built the foundation. We asked what a training programme actually is, we set objectives, we climbed Bloom's ladder from "name the parts of the portafilter" to "design a workflow for the morning rush," and we mapped theory to practical checklists.

I've received some feedback, thank you for the positivity. But, I am waiting to hear from more of you.

Because here's the uncomfortable truth: you can have a beautiful set of criteria and still assess badly. Criteria tell you what to look for. The principles of assessment tell you how to look – and how to make sure that when you say "competent," it actually means something.

Let's get to it.

The Principles of Assessment: Your Non-Negotiables

Here in South Africa, if you want your training to count towards anything accredited – SAQA-registered unit standards, QCTO occupational qualifications, your SETA's skills programmes – your assessment has to comply with a recognised set of principles. Nobody gets credits for showing up. Credits are awarded for verified evidence of competence, gathered properly.

But even if accreditation is the last thing on your mind and you're just training your own people on the bar, these principles are the difference between "I think she's fine" and "I know she's competent, and here's the proof."

Fair – Nobody gets disadvantaged for something unrelated to the skill. If your theory test is written in dense technical English and your learner's first language is isiZulu, you're testing English, not coffee. And no ambushes – learners know upfront what's coming and how it'll be judged.

Valid – The assessment measures what it says it measures. A written test on the history of the espresso machine is not evidence that someone can pull a consistent shot. Every item traces back to a criterion from Part 1. Can't draw the line? Cut the item.

Reliable – Two assessors, same learner, same day, same result. Same assessor, Monday and Thursday, same result. "Makes good coffee: yes/no" isn't reliable. Good according to whom?

Current – They can do it now. Not three years ago in a different café. Skills fade, standards move, machines change. And check your own currency – if you're still assessing against a 2015 recipe card, we need to talk.

Authentic – It's their own work. Obvious, until three learners hand in the same paragraph on Ethiopian processing, straight off the same website.

Sufficient – Enough evidence to be confident. One good shot at 07:00 on a quiet Tuesday proves very little. Three shots, realistic conditions, different variables – now we're talking.

Practicable, cost-effective and efficient – It has to actually work in your bar, your roastery, your packing floor, with your people and your time. A tool that needs two assessors, a refractometer and four hours you don't have will never get used. And an unused tool assesses nothing.

Stick these seven on the wall. Every tool you build gets checked against them.

 

The Human Element: Getting Emotion and Bias Out of the Cup

Now for the part that stings a bit, because it's personal. Assessment is done by humans, on humans – and humans are gloriously unreliable instruments.

The research on assessor behaviour is depressingly consistent. We rate people we like more generously. We rate people who remind us of ourselves more generously. We get harsher at the end of a long shift. We anchor on first impressions – one bad opening shot and the next four have to be spectacular just to break even. We're patient with the confident and impatient with the quiet, regardless of whether confidence has anything to do with competence. (It usually doesn't.)

Honestly? I think bias is the single biggest quality problem in training and assessment in our industry. Not lack of knowledge. Bias.

You'll recognise it:

  • "He's been here eight years, of course he's competent." Tenure isn't evidence.
  • "She's got great energy with customers." Lovely. Is her milk texture right?
  • "That's my dream team – I trained them, I know they're good." You are the least objective person in the building when it comes to your own trainees.
  • "The supplier's technician obviously knows the machine." Does he? Who checked?

When we assess people we trained ourselves, we're marking our own homework. Every gap we find in them is a gap in us. The urge to round up is enormous. I've felt it. You've felt it.

The fix isn't to remove the humans. It's to remove the human's discretion wherever discretion isn't needed:

  • Descriptors, not opinions. Not "tamping: good/bad." Rather: "Tamp is level; puck shows no channeling on knock-out; pressure consistent across all three shots." The assessor ticks what they saw. What "good" means was decided when the tool was built – not on the day, not in the mood.
  • Blind where you can. Sensory work – espresso, roast profiles, cupping – done without knowing whose cup is whose.
  • Two sets of eyes or moderation for anything high-stakes: promotions, certification, accredited credits.
  • Write it down, don't remember it. Memory is a bias engine.

Introducing: Subject Matter Experts (SMEs)

(New terminology. You're going to hear this a lot from us.)

One of the initiatives we'll be building through SCASA is a framework for recognising and regulating Subject Matter Experts – the people who are actually qualified to assess in a specific technical field.

Here's the problem we're solving. Right now, pretty much anyone in coffee can call themselves a trainer or an assessor. A senior barista assesses the juniors. A machine supplier's technician "certifies" a café's staff. A roaster judges a green lot. A sourcing manager evaluates a farm's processing. Some of these people are outstanding. Some are guessing with confidence. And there's no consistent way to tell which is which.

SMEs will be recognised by field, because coffee isn't one skill – it's a dozen:

Field

Scope of Assessment

Bar & Beverage

Barista skills, espresso, milk, brew methods, workflow, service at the bar

Roasting

Profile development, sensory evaluation of roasted coffee, defect identification, roaster operation and maintenance

Green Coffee & Sourcing

Grading, defect counts, moisture and density, processing method evaluation, farm-level knowledge

Technical & Equipment

Machine and grinder installation, maintenance, calibration, diagnostics – supplier trainers live here

Supply Chain & Inputs

Packaging, chemicals and cleaning products, food safety compliance, storage and logistics

Farming & Production

For those working with local producers – agronomy, harvesting, processing, drying

 

The principle is dead simple: a Subject Matter Expert assesses within their subject. A brilliant roaster isn't automatically qualified to judge latte art. A machine technician isn't automatically qualified to run a cupping. And – this is the bit people miss – being excellent at something doesn't make you qualified to assess it. Assessing is its own skill. An SME will need both: proven expertise in the field and training in assessment principles, tool design, bias management and moderation.

We'll share the framework over the coming months. If you're a supplier, technician, roaster or trainer who wants to help shape this – reply to this email. I mean that.


Building the Tools: Criteria In, Descriptors Out

Right. You've got criteria from Part 1. You've got seven principles. You've promised to fight your own bias. Let's build the actual instruments.

Rule One: Every item maps to a criterion

Take each criterion and ask: What would prove this? Then: What tool collects that proof? No tool for a criterion? You're not assessing it. Items on a tool that don't map to any criterion? You're wasting everyone's time.

Rule Two: Descriptors, not judgements

A descriptor tells the assessor exactly what to look for at each level. Here's espresso extraction in three columns:

Criterion

Not Yet Competent

Competent

Exceeds Standard

Extraction appearance

Blonde and watery from the start, or dark and dripping with visible channeling

Even flow, tiger-striping, colour moves dark → caramel → blonde within target time

As competent, plus learner reads the cues and adjusts proactively across shots

Taste balance

Clearly sour (under) or bitter, burnt, ashy (over); harsh metallic or hollow finish

Balanced acidity and sweetness, no dominant bitterness, clean finish

As competent, plus learner describes the profile and links it to grind, dose and yield

Yield & time

Off recipe by >20% on yield or ±8 sec on time

Within tolerance on 2 of 3 shots

Within tolerance on 3 of 3, with explanation of any variance

Workflow

Skipped purge, uneven distribution, dirty portafilter, wrong cup

Purged, distributed, tamped level, clean, correct vessel

As competent, and efficient under simulated rush

 

See what happened? The assessor is no longer asking "was that good?" They're asking "which box does what I saw fit into?" That's the shift from opinion to evidence.

Rule Three: Match the tool to the evidence

Not everything is a checklist:

  • Observation checklists – movements, sequence, hygiene, workflow. Watch the tamp. Watch the dial-in. Watch the purge.
  • Sensory scoresheets – taste, aroma, tactile. For roasters: under-developed (grassy, sour), correctly developed (sweet, clear), or over-roasted (burnt, ashy, carbon)? For green coffee: colour and uniformity, defects, smell (fresh, fermented, musty, phenolic), hand-feel for moisture and density. This isn't image-based assessment – it's multi-sensory, and the tool has to capture all of it.
  • Written or oral theory – for understanding and analysis. Why does channeling happen? What affects development time?
  • Portfolios and projects – for creating. Design the workflow. Write the packaging SOP.
  • Documentation review – for compliance. Batch records complete? Roast logs filled in? Temperature log signed?

Rule Four: Assess what surrounds the skill

Assessing someone packing roasted beans? The beans going into the bag are only part of it:

  • Food safety – hands washed, gloves, hairnet, apron, surfaces sanitised, allergen controls.
  • Paperwork – batch number, roast date, best-before, traceability.
  • Soft skills – when the sealer jammed, did they flag it or push through and seal 40 bad bags? Did they ask when unsure?

We love assessing the glamorous bit – the shot, the roast, the rosetta – and ignore everything that keeps the business legal and running. Go beyond the cup.

Rule Five: Pilot, then be critical

Run the tool on two or three people before you trust it. Then look hard: Did assessors read the descriptors the same way? Too long? Did an item give everyone the same result (probably useless)? Did an item give wildly different results between assessors (probably ambiguous)? Fix it. Tools are living documents.


Moderation: Checking the Checkers

I promised this in Part 1.

Moderation is verifying that the assessment was fair, valid, reliable and consistent – and that the assessor's call was sound. It's not re-assessing the learner. It's assessing the assessment.

In accredited environments it's mandatory – a registered moderator samples decisions, reviews evidence, confirms or challenges. In your own business, run a lighter version:

  • Internal moderation – a second trained person reviews one in five completed records. Do the observations support the result? As the Head of Training or a Business owner, schedule some time for reviewing and training meetings.
  • Assessor calibration – get your assessors together, same three espressos, score independently, compare. Disagreement isn't failure – it's data.
  • Learner feedback – did they understand what they were being assessed on? Did it feel fair? The learner is a witness.
  • Records – keep everything. If you ever go for accreditation, this history is gold. If there's ever a dispute, it's your defence.

Why Are We Training in the First Place?

Step back a second. Tools are only useful if you know why the training exists. In practice, every intervention in a coffee business is one of these:

  1. Induction / onboarding – baseline competence. Heavy on observation and product knowledge.
  2. Retraining due to a performance gap – something went wrong. Assess against the specific gap. Watch your bias here – the person already feels judged.
  3. Continuous training – consistency and lifting the standard over time. Periodic re-assessment, calibration, refreshers. This is where most businesses fall over – train once, assume forever.
  4. New product / new equipment – something changed. Short, sharp, frequent.

Know which one you're doing. It changes the tool.


Product Knowledge: The Skill Everyone Forgets to Assess

A specific plea, because I see this in every hospitality business I walk into.

Product knowledge is not optional, and it's not just for the front of house.

If you run a hotel with a café, a spa, a restaurant and rooms – everyone on the property should be able to answer a guest's question about any of it. Yes, the gardener too. "Is there a spa?" "Yes, behind the pool, open 9 to 6, I'll ask reception to book you in." That's not a barista skill or a gardening skill. That's a business skill, and it pays the bills.

For coffee businesses: does your team know what's in the drink? Origin, process, roast level, what it tastes like? When a new seasonal single origin lands, does it get a staff cupping and a two-minute briefing – or does it just appear in the hopper and the team learns about it from a customer's question?

Product knowledge suits low-stakes, high-frequency assessment beautifully:

  • A five-minute Friday quiz on the current menu. Make it fun. Prizes allowed.
  • A "product of the week" card in the back with three facts everyone should know.
  • A monthly "explain this drink to me like I'm a customer" role-play, scored on a simple rubric.

Light, continuous, low-anxiety. That's how training sticks without becoming tedious – and yes, it's still assessment. You're still gathering evidence and still writing it down.


Paying For It: The Levy Isn't a Tax, It's Your Money

Quick but important. If your business pays the Skills Development Levy (1% of payroll, above the threshold), you can claim a portion back through your SETA – mandatory grants for submitting your Workplace Skills Plan and Annual Training Report, and discretionary grants for specific programmes, learnerships and skills programmes.

Most small and medium coffee businesses I speak to have never claimed a cent. That money funds accredited training. Accredited training needs assessment that meets the principles above. See how it connects?

Or you choose to fund training yourself, out of operating budget, because you get that it's an investment. Either way – the assessment has to be robust, because whoever's paying wants to know it worked.

SCASA will be putting together SETA guidance for the coffee sector. Watch this space.


Your Homework for Part 2

Take your task from Part 1 – espresso or bean packaging – and:

  1. Audit it against the seven principles. One honest sentence per principle on how you meet it – or don't.
  2. Rewrite three checklist items as descriptors in the three-column format.
  3. Add the "surround" items – food safety, paperwork, soft skills – your original list missed.
  4. Name your bias risk. Who assesses your team? Did they train them? What's one thing you could change?
  5. Run a mini-calibration. Two people, same three shots (or three bags), score independently, compare. Send me what you found.

I will assist with every submission.


Where This Goes Next

Part 1 was criteria. Part 2 was principles, tools, bias and moderation. Together they give you an assessment system that stands up to scrutiny – yours, your team's, and eventually an accreditor's.

What's still missing is the standard itself – the minimum competence the industry actually agrees on. That's the big conversation, and it's coming. Subject Matter Experts are the first structural step towards it. Consensus between a café in a rural town, a roastery in Gauteng and a hotel in Cape Town won't happen overnight. But it starts with all of us assessing properly – so when we compare notes, we're comparing evidence, not opinions.

Keep training. Keep assessing. Keep building.

 

Henning Lubbe Director: SCASA Board

Skills Development Specialist

Email: henning@kushindaacademy.com

SCASA: info@scasa.coffee

 

Back to blog

Leave a comment