Case study · 3D design, engineering, and simulation software

Turning the European startup ecosystem into qualified engineering-software accounts

Finding European startups that build physical products is a qualification standard sales databases can't express. We researched what each company actually does, then decided.

Download detailed case study (PDF)PDF · 5 pages

Case study · Enterprise engineering software

Dassault Systèmes, 3D design, engineering, and simulation software.

13 European markets researched
3 possible fit classifications
Regions researched
The Nordics, Benelux, UK & Ireland, and the Baltics
Fit states
Yes, no, or unknown — preserving uncertainty reduces false confidence
Core question
Does the company plausibly need sophisticated 3D design, engineering or simulation software?

The problem

A startup list was not the market

Dassault Systèmes wanted to identify young European companies that could plausibly need sophisticated 3D design, engineering and simulation software.

Why a normal startup list wasn't enough

  • Early-stage companies under five years old, with roughly €5 million or less in funding and revenue, formed the boundary of the search universe.
  • Most startups were irrelevant: a software startup, marketplace or consulting business could pass every company filter with almost no need for engineering software.
  • Two companies can share identical firmographics — five years old, €3 million raised, twenty staff, “technology” — while one builds workflow software and the other designs and manufactures a robotic product. Filtering could narrow the market; it could not finish the qualification.

The question was not “is this a startup?” but “is this the kind of startup that actually designs, engineers or builds something?”

What we built

A research model for defensible engineering-software relevance

  1. Define the structured boundaryRemove obvious non-fit companies by geography, age, funding and revenue — creating the search universe, not yet the qualified list.Define
  2. Research what the company actually doesDoes it develop a physical product? Design or prototype hardware? Is engineering central, with engineering or design talent employed?Qualify
  3. Classify fit with three statesYes — evidence the company fits. No — enough evidence it doesn't. Unknown — not enough evidence, which is not the same as no.Qualify
  4. Research ecosystem affiliationsStartups live in networks: accelerators, incubators and VC portfolios reveal additional companies to research and paths into the market.Contact
  5. Identify the relevant peopleFirst determine the right company; then ask who matters inside it — technical founders, heads of engineering, design and product-development leaders.Launch

No single source contained every answer: startup databases established funding, LinkedIn revealed team composition, company websites revealed the physical product, and accelerator portfolios revealed affiliation. The qualification decision emerged from the combination.

AI inside the research

Narrow questions, not one ambiguous instruction

Prompts evaluated company information against the ICP by breaking the decision into narrower questions: does the evidence indicate the company develops its own physical product? What evidence supports that conclusion? Is there prototyping or manufacturing evidence? Is the evidence strong enough to classify?

That makes AI a research component rather than a source of unsupported certainty — and keeps contact research downstream of account judgment, because contact information has little value if the account is wrong.

Result

Account intelligence, not a startup export

  • The ICP became executable research logic: physical-product inference from public sources, ecosystem research as a discovery mechanism, stakeholders identified once fit was clear.
  • An ICP is not finished until a system can determine who fits it.
Right company → Right reason → Right people → Right message.
Next caseTempo Software: from a Jira signal to product-aware outreach →

Related reading

A practical guide to the research and system-design decisions behind work like this.

How to find B2B prospects that sales databases miss

Let's discuss your GTM challenge.

A short, no-obligation introductory conversation.

Schedule a conversation