Every marketplace pitch deck opens with the same two names. Uber solved transport by aggregating drivers at network scale. Airbnb solved travel by aggregating spare rooms at network scale. Both are held up as the template: win liquidity, win the category, defend with scale.
Quietly, far from Silicon Valley’s conference stages, a different kind of platform is winning categories too — and it looks nothing like either template. It is thin, covering one narrow home-services vertical, employing no drivers, hosts or operators, carrying no inventory, and often running on a team small enough to fit around one table. It simply routes a buyer’s request to a handful of independent local providers and takes a referral fee when a deal closes.
That model is unglamorous, rarely pitched, and almost never covered — yet in fragmented, low-trust, opaque-pricing markets it is proving genuinely defensible. The question worth asking isn’t whether the horizontal giants can be replicated. It’s when a narrow vertical marketplace actually wins, and why most attempts don’t.
The horizontal playbook doesn’t travel well
Network-scale marketplaces are capital-hungry by design. They need liquidity on both sides before the product works at all — subsidising supply, demand, or both, for long enough to reach critical mass. That calculus assumes deep capital markets, high card penetration, and a demand side willing to transact inside an app. Commentary tracking these platforms — including business coverage in Forbes — increasingly notes that the horizontal, network-scale model is only one branch of the category, not its full definition.
Move the same playbook into a market with thin venture capital, patchy card infrastructure, and mostly informal, relationship-driven supply, and the economics buckle. Liquidity is expensive to build and easy to lose, and a well-funded local incumbent can often out-subsidise a foreign entrant before it reaches scale. This is precisely the gap the thin vertical model exploits: it needs no liquidity subsidy, because it isn’t trying to own the transaction — only the moment the buyer decides whom to call.
The four preconditions for a vertical marketplace to win
Not every home-services category is a candidate. The pattern that recurs across the ones that do work is a short checklist rather than a single insight.
The supply side has to be fragmented, with no single brand already dominating buyer trust. Pricing has to be genuinely opaque — the kind of purchase where two providers quote wildly different figures for what looks like the same job. The purchase itself has to be high-consideration and infrequent enough that buyers feel real anxiety about getting it wrong; nobody builds a comparison habit around something bought weekly. And demand has to be mobile-first, arriving through organic search rather than an app download — where data cost is a real constraint, the install step is where most attempted marketplaces quietly die.
Where all four line up, a thin platform can win a category cheaply, through search intent alone, without ever employing a single tradesperson.
According to Investopedia’s definition of a vertical market, the defining trait of this structure is specialisation within one narrow buyer-and-supplier niche rather than breadth across many — exactly the discipline these platforms hold to.
What the model looks like when it works
A live example is worth more than the framework alone. Consider a South African platform operating in exactly one vertical: precast and vibrecrete boundary walling and palisade fencing — a category most outside observers would never think to build a business around.
The platform holds no inventory and employs no installers. It is paid a disclosed referral fee by the independent installers it introduces to a homeowner, and it is free to the person requesting quotes. Visitors set the scale of the search themselves — a buyer already leaning toward one installer can request a single quote, while someone still comparing price and timeline can request more, typically up to three or, where more installers are active in the area, up to five. It is, in effect, a South African marketplace for comparing precast boundary-wall quotes, built to the same thin, no-inventory logic as the framework describes.
It clears all four preconditions: fragmented independent installers, no dominant national brand, pricing the average homeowner cannot benchmark unassisted, an infrequent, anxiety-laden purchase, and demand arriving almost entirely through search rather than an app. That fit is precisely why the model works there — and, per the framework, why it should travel to comparable categories rather than stay a one-off.
The data moat: why pricing transparency, not scale, is the defensibility
What makes this kind of platform hard to copy is not the routing technology, which is trivial to replicate. It is the pricing data it has accumulated and published.
The precast example illustrates the mechanism. Publishing an indicative installed range — roughly R620 to R1,400 per metre — alongside named cost drivers (decorative finishes adding a premium over plain panelling, coastal areas adding one for salt-air treatment, difficult site access raising excavation costs) does something a horizontal marketplace’s scale never achieves in a niche category: it answers the exact query a confused buyer types, and gives that buyer enough context to negotiate confidently once quotes arrive. A competitor can copy the interface in a week; rebuilding years of accumulated, search-ranked pricing context is a slower moat, and a much harder one to buy.
The counter-case: why most verticals don’t clear the bar
None of this makes the model a sure thing, and the honest version of this thesis has to include the categories where it fails.
Thin take-rates are the first problem: a referral fee on a low-ticket, high-frequency service rarely covers the cost of acquiring the lead in the first place. Category ceilings are the second — some verticals are simply too small in transaction volume to support even a lean team once the founders’ own time is priced in. Supply-side disintermediation is the most persistent risk: once a provider and buyer have transacted once through the platform, nothing stops them going direct next time, and thin marketplaces without a genuine reason to stay in the loop leak volume steadily. And in markets with real regulatory exposure — building codes, licensing regimes, consumer-protection rules — a platform that merely routes leads can still inherit reputational risk for work it never touched.
The framework filters for viability. It does not filter for certainty.
The takeaway for builders and investors
The next wave of proptech winners in emerging markets is unlikely to look like the last one. Rather than another listing site trying to out-scale an incumbent, the more interesting opportunity sits in categories nobody has bothered to formalise: fragmented, opaque-priced, infrequent, mobile-first — and small enough that a founder can own the search results before anyone with more capital notices there was a category worth owning.
For anyone evaluating a pitch that looks like this, the four preconditions are the diligence checklist, and the counter-case is the reason not every fragmented category is a business. The ones that clear all four rarely look impressive on a slide. They just quietly work.
