Structured demand: how an SME buyer finds a deal before it lists

A German Mittelstand corporate-development team has been searching for eighteen months: precision CNC supplier to automotive, €3 to €8M revenue, EBITDA margin above 12%, owner willing to step out within 18 months, located in Flanders or North Brabant within logistical reach of their HQ. At the same time an owner in Mechelen who fits that profile is considering a sale within 12 to 24 months. He has spoken to his accountant. He called one M&A broker; the fee structure didn't work. He is about to put his deal on a Belgian marketplace.

These two will rarely find each other in today's system. Not because demand doesn't exist. Not because the business doesn't fit. Because there is no infrastructure that connects the buyer's intent to the right type of deal. That is what we mean by "there is no liquidity in the European SME succession market." Not that there is no demand. That demand is unstructured.

What is fundamentally wrong with the marketplace approach

An M&A marketplace is essentially an advertising model. Owners post deals, buyers browse. That works for low-friction transactions (used cars, apartments) where the searcher has time and a sub-optimal match is fine. For SME M&A, none of those conditions hold:

1. Buyers are institutional entities (corp dev, family office, PE, strategic) with specific complex criteria. They don't browse 800 listings.
2. Sellers want discretion (customers and staff cannot know). An open ad cuts against discretion.
3. Transaction value is high. A sub-optimal match means the deal collapses in DD or sells 30% under fair value.
4. Matching is multi-dimensional and partly qualitative. Sub-sector, size, geography, ownership, integration angle, time horizon, owner exit style. Ads cannot weight that.

Result: 70% of listings on SME marketplaces get no serious interest. 73% are mispriced (usually too high). And the most attractive buyers, typically cross-border strategics with the highest willingness to pay, rarely show up there.

What a matching engine does differently

Instead of showing listings, a matching engine channels demand. Buyers register their search profile with the same discipline they apply to their investment thesis: NACE/SBI level 4 or 5 sub-sector, size band, geographic priorities (first / second / no-go), recurring mix, owner-exit flexibility, integration angle, time horizon.

On the seller side, once a valuation runs through Upswitch, the deal can be auto-matched (anonymously) against active search profiles. Not every deal is opened for matching, that is the seller's explicit choice. When opened, the system identifies the ten to thirty buyers with overlapping search profiles in seconds.

Why this only works after the other primitives are in place

Tempting to start with matching, visible, pitches well. But without the underlying truth layer, a matching platform is just a better classifieds page. A German strategic buyer matched to a Belgian company will only engage if they trust the valuation. Without normalisation (P1), transparency (P2) and shared truth (P4), DD collapses on first contact. Upswitch builds matching as the last primitive, not the first. The 2026 matching engine is initially low-volume: only valuations that have run through our normalisation and transparency layers, only buyers whose search profiles are validated. We could scale faster by relaxing both, but that would destroy the structural value, trust.

How the matching process actually runs

1. Anonymised alert. Match found, both parties get an anonymised alert (sector, size band, region), no names.
2. Mutual interest confirmation. Both click "interested" without seeing each other.
3. NDA exchange. The system orchestrates digital NDA signing in 24 to 48 hours.
4. Full profile exchange. Buyer gets the IM and valuation; seller gets the buyer profile including prior transactions.
5. First call. Both parties review and, if both want, schedule a call, often via either side's M&A advisor.

End-to-end, anonymised alert to first call: about a week. Compared with the traditional model (an M&A broker working 8 to 12 strategic outreaches over 3 to 6 months), time efficiency is an order of magnitude higher. Not because the final conversation is better, that is comparable, but because cycles are shorter and match precision higher.

What this means for cross-border M&A

Today, Benelux SME deals are overwhelmingly local: 70 to 80% are Benelux-on-Benelux. Not because that is optimal. German, French and Dutch strategics often pay more for synergy. But cross-border coordination cost is too high today. Shared truth + structured matching flips that. The Antwerp owner becomes visible to the Munich buyer. The valuation reproduces in the buyer-advisor's environment. The cross-border deal that takes nine months can come back to three or four. That is what gets signed. The Upswitch Index per business type makes the cross-border comparison explicit. German and French buyers see for manufacturing, B2B services and agro-food the live country bands that ground the offer.

What we have seen so far

Our matching engine is in early access since Q2 2026. Early Benelux results: average time from valuation to first serious buyer call is 18 days vs. a market median of 75 to 90 days for traditional process. Match-to-LOI conversion is ~35% vs. 8 to 12% for marketplace leads. Not because our buyers and sellers are better, because match precision filters weak leads at the front. The endgame: a market where a Mechelen precision supplier and a Bavarian Mittelstand buyer find each other in 18 days instead of never. Liquidity follows truth, and the truth is that demand was always there. It just lived in a different dashboard.

The shared-truth layer that matching rests on is public. It is the Upswitch Index: per business-type EV/EBITDA, EV/Revenue and P/E bands, supported by a transparent methodology and published per country. A German buyer matched on a Belgian manufacturing or B2B SaaS deal sees not just our valuation, they see the exact band against which we valued. No Excel. One bench.