Technology review, interrogated.
Also called: Technology validation · Maturity review · Tech risk pass · Component due diligence
A hard pass over your chosen technologies before you commit, checking maturity, supply, obsolescence and single points of failure rather than novelty.
Each technology you picked gets interrogated for hidden risk before committing. Is it mature, can you buy it in two years, what happens when one part dies? The datasheet is a sales document, not a guarantee. There’s a worked review below, and a checklist you can run.
What a technology review is
This is validation, not exploration. Earlier in Stage 03 Innovate you researched technologies to find candidates: what could heat the chamber, what could sense the temperature, what could close the loop. That work asks “what is possible?”. A technology review asks the harder, more boring question: “is the thing I picked actually safe to bet a product on?”
The two passes feel similar and are not. Exploration is optimistic by design, it wants options on the table. A review is sceptical by design, it wants to find the reason each option will let you down before a customer does. In my experience the failure mode here is treating the datasheet as truth. A datasheet is a sales document with a tolerance table attached. It tells you what the part does on a good day in a lab; it tells you nothing about whether the maker will still ship it in eighteen months.
So you take each load-bearing technology choice and you put it under four headings: how mature is it, can you reliably buy it, will it go end-of-life, and what breaks the whole product if this one thing fails. The output is not a longer list of options. It is a verdict on each choice: commit, commit-with-mitigation, or go back and pick again.
Resist the pull toward the newest part. A clever new sensor with a flashy datasheet and three months of production history is a liability dressed as an advantage. Boring, mature, second-sourced parts win products. The novelty belongs in the proposition, not the bill of materials.
Here is what that review looked like for the proofing box we ran through it, so you can see the shape of a real verdict rather than a generic template.
Notice the review didn’t reject the design. It found three quiet liabilities and fixed them while fixing was still cheap. That is the entire job.
Where teams get it wrong
The same review, done two ways. The weak version trusts the paperwork; the strong version goes looking for the part that will let you down.
- “The datasheet says it works, so it works.”
- Picking the newest part because the spec looks best.
- One sensor, no independent safety backstop.
- Never checking lifecycle status before tooling.
- Verify stock, lead time and second sources at two distributors.
- Favour mature, proven parts over the newest in the catalogue.
- Add an independent cut-out for any safety-critical path.
- Pull lifecycle status for every part before committing to tooling.
The weak column is faster and feels productive. It just moves every one of those risks downstream to a point where they cost real money to fix.
How it fits the bigger picture
Technology review is activity 07.10.03 in the framework, inside Stage 07 Engineer. It sits after the early technology research from Stage 03 Innovate, which proposed the candidates, and feeds directly into materials research (07.10.04), where the chamber, shell and thermal materials get the same hard validation pass the components just had.
What it can do
It catches the part that will go end-of-life, the connector on a six-month lead time, and the single sensor that can cook the product, while every one of those is still a cheap edit on a drawing rather than an expensive recall. It turns a hopeful bill of materials into a defensible one.
What it can’t do
It can’t tell you the technology is the right choice for the customer, that was settled earlier, and it can’t validate the physical materials around the electronics. That is what materials research does next. It also can’t remove risk entirely; it can only make the remaining risk known and chosen rather than discovered later.
See the full 10-stage process →
Try it yourself
List every load-bearing technology in your design. For each one, answer four questions honestly: How mature is it, really? Can I buy it reliably in two years? Is it flagged end-of-life anywhere? What happens to the whole product if this one part fails? Anything you can’t answer with evidence is an open risk, not a settled choice. Write a verdict against each: commit, commit-with-mitigation, or pick again.
Want a structured first pass before you go this deep? Start the Free Sprint → and the GPT will help you frame the technology risks worth chasing.
Your technology-review checklist
Project notes: the part that nearly aged out
▸ From the notebook · optional reading
A half-day review with the Manchester PCB partner that caught a display driver about to go end-of-life, before a penny of tooling was cut.
3 min read · click to open
Dan was ready to freeze the design. The proofing box electronics worked on the bench, the temperature held at 26°C, give or take half a degree, under 30W, and he wanted to release the board for tooling. I asked for half a day first: “Before we commit, we walk the whole bill of materials with the people who actually build it.”
What the walk turned up
We sat down with the PCB partner in Manchester and went part by part. Three things surfaced that the datasheets had quietly hidden.
The OLED driver. Flagged “not recommended for new designs” on the manufacturer’s lifecycle page. It still sold, so it looked fine on the bench, but designing it in meant a redesign was coming within a year or two. We swapped it for a current part with a published longevity commitment. A drawing change at that point; a re-spin and re-certification if we’d found it after launch.
A connector on a 30-week lead time. Fine for a prototype, fatal for a launch. We pushed for a second-sourced equivalent so a single distributor’s stock-out couldn’t halt production.
One sensor running everything. The same sensor read the chamber and drove the heater. A stuck-high fault would cook the dough and overheat the box, exactly the failure BS EN 61010 exists to prevent. We added an independent thermal cut-out on its own path. Cheap. Non-negotiable.
What the review cost vs saved
- Cost. Half a day, plus three small part swaps. No change to the £149 price or the £38–55 bill of materials.
- Saved. One near-certain re-spin and re-certification, an avoidable production stall, and a safety gap that should never have reached a kitchen.
The newest, cleverest sensor on the table never made it in. The boring, proven one did. That is usually how a good review ends.
— Engineer stage, project notes, 2026
— Next in Engineer → Materials research