About
Why noisefloor.systems exists.
A semiconductor vendor's field team can look after a few dozen accounts properly. The market its parts could go into has thousands of products in it. Most of the decisions that matter, which part goes into which product, are made where the manufacturer is not looking, and it learns about them when a quote request arrives or when it does not.
noisefloor.systems was founded to give those manufacturers some of that sight back: to do the reading a field team has no time for, and to hand back something an engineer or a salesperson can check and act on.
The level below which a signal goes undetected.
Every engineer knows the term. The noise floor of a measurement is the level of everything that is not the signal. Whatever sits below it is there, and the instrument does not show it.
A market behaves the same way. A new hire, a product revision, a regulatory filing, a line in a datasheet: each is public, each is small, and on its own each disappears in everything else that is published that day. The work is lowering the floor far enough to see them, and being honest about what is still not visible.
Evidence with its source attached, observed facts kept apart from assumptions, and a plain statement when the public record does not answer the question.
Manufacturers are the last to know their part is being evaluated.
At CELUS I worked with semiconductor and component manufacturers on one side and the engineers designing products on the other. What stayed with me was how little the manufacturers could see.
Much of their business runs through distribution. The distributor takes the enquiry, ships the samples and books the order. The manufacturer often sees a line in a sales report weeks later, without the product it went into or the engineer who chose it. They do not know when their part is being evaluated. They do not know who is designing. They find out that a competitor won a socket when the product ships.
It is most pronounced in the long tail. A manufacturer knows its twenty largest customers by name. The thousands of smaller companies designing products with its parts, or with a competitor's, reach it only as order lines through a distributor, if at all.
None of this is anyone's fault. It is how the channel is built. But it means a company that knows its own product better than anyone is planning its sales year half blind.
- Field engineers cover the largest accounts. The rest of the market gets a website and a distributor.
- The long tail is where that hurts most. Thousands of small and mid-sized OEMs each buy modest volumes, together they are a large part of the market, and no manufacturer can afford to visit them one by one.
- Engineers evaluate parts without contacting the vendor. By the time a quote is requested, the choice is made.
- A smaller manufacturer's part is filtered out before evaluation because the engineer has not heard of it.
- A design win takes one to three years to turn into volume, so a missed evaluation is missed revenue long after.
- Marketing and sales teams are small. There is no time to read a market product by product.
- The CRM records companies and deals. It does not record which design at which customer a part is in.
Not everything. What is inside a product and when a team will choose are often not public. But a good part of the picture is, in datasheets, filings, job postings and teardowns, and reading it systematically gives a field team somewhere to start that is better than a list of company names.
Robin Malter.
Before noisefloor.systems I managed GTM strategy, operations and automation at Altium, now part of Renesas, and at CELUS, an electronics design-automation company. In both roles I worked closely with component manufacturers, semiconductor companies and the OEMs that develop electronic products.
I build the workflows, review what they produce, and decide with the client what a finding is worth. Where a technical judgement is needed, a field application engineer makes it.
Software does the reading. People make the decisions.
Collecting documents, extracting what they state, comparing it against a client's criteria and sorting the results is repetitive and structured. That part is automated, and every step keeps the source it worked from.
Deciding which market to enter, which account deserves an engineer's week, and what to say to them is not. That part stays with people: the client's team, an application engineer, and me.
- Every claim carries its source.
- Missing evidence is recorded as missing, not filled in.
- Technical fit and commercial timing are judged separately.
- Records are about products and companies. Personal data is not collected from public traces.
Munich. Working with semiconductor vendors.
- Semiconductor vendors, especially mid-size and smaller ones with thin field coverage
- One service: design-in intelligence, in three steps
- English and German
What design-in intelligence is · A worked system, described as it is
Start a conversation