Proving a new technology is ready to depend on: beyond the best result

Every new technology arrives with its best result: a sensor that caught every defect in the demonstration, a treatment system that removed a contaminant almost completely, a material with impressive test figures, software that predicted demand accurately on historical data. Those results are real achievements and a fair reason for interest. They are not proof that a business can depend on the technology.

Depending on a technology means building processes, staffing, customer commitments and money around it. That requires a different kind of evidence: that the technology keeps delivering under the variation, wear, contamination, maintenance, skills and support that exist where it will actually be used. The gap between “it works” and “we can rely on it” is where many technology investments disappoint.

This article explains what technology readiness means in practical terms: evidence that a technology will keep producing the required result in your operating conditions, over time, with the people, maintenance and support available. It covers the questions that a demonstration does not answer, how to define the operating conditions before choosing, how to test for the things that reduce performance and how to stage commitment so that money follows evidence.

A best result shows possibility, not dependability

Demonstrations and laboratory studies often explore conditions specifically to find the best performance. Real operations rarely sit at that point. Inputs vary, temperatures change, equipment wears, operators differ and maintenance is imperfect.

Published research in process engineering illustrates the point across very different technologies. Studies of catalysts, the materials that speed up chemical reactions, routinely track how activity falls over time as carbon builds up or the material degrades, because a catalyst that performs brilliantly for a day and poorly after a month can ruin a process’s economics. Studies of catalysts that can be regenerated and reused show that performance often falls a little with each cycle. Research on a granulated clay material for removing arsenic from water found that a fine powder form that worked well in a beaker was impractical in a flow-through system, that common substances in real water such as phosphate interfered with removal, and that capacity declined after repeated regeneration. Research on biological systems for treating waste streams found performance depended on interactions between the culture, the feed and the lighting, not on any one component.

The common thread is that readiness is a property of the whole system in its real conditions, not of the core technology at its best.

The chain of questions

A useful way to think about readiness is a chain of progressively harder questions:

1. Can it work? Does it produce the required result under defined conditions?
2. Can it work repeatedly? Do repeated runs reproduce the result within an acceptable range?
3. Can it tolerate variation? How does performance change with real variation in inputs, environment and load?
4. Can it be maintained? What degrades, how fast, and can it be restored predictably?
5. Can it be integrated? What utilities, controls, data, skills and interfaces does it need?
6. Can it be operated safely?
7. Can it scale? What changes when it is larger, faster or running continuously?
8. Can it create value? Does it pay its way once wear, maintenance, consumables and downtime are counted?

A demonstration usually answers the first question. Each later question needs its own evidence. The more expensive and harder to reverse the commitment, the more of the chain should be answered before committing.

Common misreadings

• Treating the best measured point as the expected operating point. Plan around sustained performance across normal variation, not the peak.
• Ignoring degradation because it happens after the demonstration. Fouling, wear, drift, contamination and ageing can turn a strong first week into a poor first year.
• Assuming component performance predicts system performance. An excellent sensor can be let down by lighting, mounting or data handling. A strong core process can be undone by the handling of its inputs and outputs.
• Treating technical feasibility as commercial readiness. Spare parts, local support, skills, standards or customer acceptance may not be ready even when the technology is.
• Defining affordability as purchase price. Consumables, energy, maintenance, downtime, training, monitoring and disposal can outweigh the purchase price over the life.
• Testing with clean inputs. Real inputs contain contaminants and variation that test samples may not.

Define the operating envelope first

Before choosing a technology, describe the conditions it must work in. This operating envelope includes:

• Inputs: composition, contamination, variation and how they change over days and seasons.
• Environment: temperature, humidity, dust, vibration, lighting, wash-down.
• Throughput: normal, peak and minimum loads, start-stop patterns.
• Utilities: power quality, water, compressed air, network connectivity.
• People: the skills of the operators and maintainers who will run it.
• Maintenance access: time windows, spare parts, specialist support.
• Acceptable downtime: how long the business can tolerate the technology being unavailable.

Writing this down first reverses a common pattern in which a promising technology is chosen and the business later discovers what it needs to support it.

Physical form and handling matter

Technologies that work chemically or electronically can still fail practically. Fine powders clog and are hard to separate. Viscous fluids are hard to pump. Hazardous reagents need storage, training and controls. Equipment that must be calibrated by a specialist from interstate may sit idle for days. The arsenic research above is instructive because the researchers treated the physical form of the material, granules rather than powder, as part of the engineering problem. Ask early how the technology will be fed, handled, cleaned, stored and serviced.

Test what reduces performance

A good trial deliberately includes the conditions most likely to reduce performance: real inputs with their normal contamination, the worst lighting or temperature, the busiest periods, the least experienced operator, the product variants that are hardest to handle. This is sometimes called interference testing. The aim is to find the dominant failure mode before the business depends on the technology, not after.

Look at variation as well as averages. A system with excellent average performance but large swings may be harder to manage than one with a slightly lower but predictable result.

Degradation, maintenance and recovery

Every technology degrades. The questions are what degrades, how fast, how you will know, and how capability is restored. For each technology, define:

• The degradation mechanism: wear, fouling, drift, contamination, ageing.
• Detection: how operators will know performance is falling before it causes a problem.
• Recovery: cleaning, regeneration, recalibration, replacement, and who does it.
• Cost: consumables, labour, downtime and disposal of spent materials.

The relevant business measure is often sustained output between interventions, or total cost per unit of good output including maintenance and recovery, rather than peak performance. Understanding why performance falls, not just that it falls, makes the technology far more controllable.

Fit matters more than sophistication

Appropriate technology is sometimes misread as meaning simple or low-tech. A better meaning is technology that fits its context. A sophisticated automated system can be entirely appropriate where power, connectivity, diagnostics and specialist support are reliable. The same system can be fragile on a remote site or in a business without the skills to support it. A simpler option with slightly lower performance can be the better choice if it is robust, easy to inspect and maintainable by the people who will actually run it. Context, not prestige, should decide.

A readiness review

For any significant technology commitment, use a short review. It does not need a perfect answer at every step. Its purpose is to show where uncertainty remains and whether the next commitment is proportionate.

Fund evidence in stages

Rather than a single decision between continuing to investigate and buying in full, stage the commitment so each step retires a specific uncertainty. An early step might test function with real inputs. The next tests durability and maintenance. A pilot tests integration and economics. Full adoption follows when evidence supports it.

Staging creates clear points at which to stop. If the critical failure mode turns out to be uneconomic to manage, the business can stop before costs and commitment become large. Be wary of the phrase “the remaining issues are just engineering” when those issues decide reliability, safety or cost.

A worked example

This is an illustration. A small food manufacturer packs about 20,000 jars a week and checks labels by eye at the end of the line. A supplier demonstrates an automated camera inspection system that detects nearly every label defect in samples the business provides. The price is attractive, and the owner is keen to buy.

Using the readiness review, the owner first writes down the operating envelope. The line runs in a room with skylights, so lighting changes through the day. The line is washed down every afternoon, and the room is humid. The business packs 12 products, two of which have glossy labels. Line speed varies with product. Nobody on staff has used vision systems before.

The owner negotiates a four-week paid trial on the real line, with the supplier agreeing to credit the trial fee against the purchase. Each day, staff feed a set of jars with known label defects through the line to measure detection.

In the first week, the system catches 94% of the known defects and wrongly rejects about 3% of good jars, around 600 a week. Each false reject needs a manual check of about 30 seconds, about 5 hours of labour a week. The causes are found quickly: afternoon sun through the skylights, condensation on the lens after wash-down and reflections from glossy labels.

The fixes are a shroud over the camera, a lens-cleaning step added to the wash-down checklist and separate inspection settings for the two glossy products. By the fourth week, detection is 98% and false rejects are about 0.5%, around 100 jars a week or about 50 minutes of checking.

The owner buys the system on conditions: training for two staff in adjusting inspection settings, a spare lens kit held on site, a defined support response time and a weekly check using the known-defect jars. The trial cost a little and took a month, but the business now knows how the system behaves in its real conditions, what it needs to stay effective and what to monitor.

How this applies to a small Australian business

Small businesses often buy technology from vendors whose demonstrations are compelling and whose support is located elsewhere. Practical steps:

• Write down your operating envelope before talking to suppliers.
• Ask for references from businesses with similar conditions, and speak to them.
• Negotiate a trial on your site with your inputs, ideally with the cost credited against purchase.
• Test worst cases, not just typical conditions.
• Ask about local support, spare parts and training, and get commitments in writing.
• Count whole-of-life costs, including consumables, maintenance, downtime and staff time.
• Plan monitoring so you know when performance falls.

The articles on proving a process is ready for production and the best long-term technology may not be your next investment cover related decisions.

Signals worth watching

• The same best result quoted repeatedly, with no information about variation.
• Trials much shorter than the expected life of the equipment.
• Maintenance assumptions with no demonstrated procedure.
• Performance that depends on narrow input specifications.
• Consumables or support costs left out of the business case.
• Unexplained performance decline.
• Pilots with no named owner from operations.
• Operators developing workarounds.

Common mistakes

• Buying on the demonstration without a trial in real conditions.
• Choosing the technology before defining the operating envelope.
• Ignoring degradation, maintenance and consumables.
• Testing with clean, ideal inputs.
• Comparing purchase price instead of whole-of-life cost.
• Choosing sophistication that the business cannot support.
• Committing in full when a staged approach was possible.

Frequently asked questions

Will suppliers agree to trials? Many will, especially for significant purchases, though terms vary. A paid trial credited against purchase is a common compromise. If a supplier refuses any form of trial or reference check, weigh that in your decision.

How long should a trial run? Long enough to cover the main sources of variation and at least one maintenance cycle. For many small-business technologies, that means weeks rather than days.

What if the technology is new and no one has used it in conditions like ours? Then you are partly funding development. Stage the commitment, keep the old method available and make sure the contract reflects the uncertainty.

Questions to ask

• What has this technology proven beyond its best result?
• What condition in our operation is most likely to reduce its performance?
• What degrades first, and how will we know?
• What maintenance, consumables and skills does it need, and can we provide them?
• What changes between the trial and full operation?
• What evidence would justify the next commitment, and what would stop it?
• Can our business run this without the supplier on site?

Bringing it together

A best result is evidence of possibility. Readiness is evidence of dependability. Define your operating envelope before choosing, test the conditions that reduce performance, understand how the technology degrades and how capability is restored, and choose technology that fits your context rather than the most sophisticated option. Review readiness across function, repeatability, variation, stability, maintenance, integration, safety, scale and economics, and stage your commitment so that money follows evidence. The question is not whether it works. It is whether your business can depend on it.

Source: KEVOS notes, drawing on published process-engineering research on catalyst stability, water treatment media and biological treatment systems. Examples and figures in this article are illustrations.