No Front Door

By Victor Muchiri, Chief of Staff at BuildVision · March 2026

Name an industry and there's probably a marketplace. Real estate has Zillow. Hospitality has Airbnb. Freight has Flexport. Used cars have CarMax. Even hiring has LinkedIn and Indeed. Someone built the front door — the single place you go to start a transaction — and captured an enormous amount of value by sitting there.

Construction doesn't have one. Not for finding subcontractors. Not for procuring equipment. Not for managing bids. Not for anything. There is no front door to the construction industry.

This isn't an accident. And it isn't because nobody has tried. Hundreds of venture-backed startups have attempted to build construction marketplaces over the past 15 years. Most of them are dead. The ones still alive have narrowed their scope so dramatically that they're no longer marketplaces at all — they're tools for specific workflows on specific project types.

Understanding why construction has no front door is the key to understanding what technology can actually do for this industry. And what it can't.

Why other industries got marketplaces

Marketplaces work when three conditions are true. First, the units of supply are roughly comparable. An Airbnb listing in Brooklyn and another in Williamsburg are different, but you can compare them on the same dimensions: price, location, rating, photos, amenities. You can put them side by side and make a choice.

Second, transactions are repeatable. A rider takes an Uber three times a week. A shopper buys on Amazon every few days. The marketplace handles the same type of transaction millions of times, and it gets better at matching supply and demand each time.

Third, the buyer doesn't need deep domain expertise to evaluate the options. You don't need to be a hotelier to book a hotel room. You don't need to be a logistics expert to ship a package. The marketplace abstracts away the complexity so that a general consumer can participate.

Construction fails all three tests.

Why construction breaks the marketplace model

Every project is a one-off. A 200,000 square foot hospital in Houston and a 50-unit multifamily in Denver share almost nothing. The mechanical systems are different. The codes are different. The labor markets are different. The equipment is different. The supply chains are local. You can't standardize the unit of supply when the unit of supply is an entire building that has never existed before.

Relationships are local and persistent. A GC doesn't pick a mechanical sub from a marketplace. They've worked with the same three mechanical subs for the past decade. They know their forepersons by name. They know who shows up when it rains. They know who handles change orders without a fight. That trust is earned over years of shared jobsites. A marketplace rating system doesn't capture it.

The specifications make it complex. When you procure a chiller for a hospital, you're not buying a commodity. You're buying a specific piece of equipment defined by a 40-page spec section written by a mechanical engineer for this particular building. The chiller needs to hit a specific cooling capacity, efficiency rating, refrigerant type, sound level, footprint, and electrical configuration. It needs to be compatible with the building automation system. It needs to ship within the project schedule. No two chiller purchases are identical because no two spec sections are identical.

The buying process is fragmented across specialties. Equipment procurement on a single project might involve mechanical, electrical, plumbing, and fire protection — each with their own engineers, their own spec sections, their own equipment schedules, their own manufacturer reps, and their own approval workflows. A single marketplace would need to understand all of these trades simultaneously. Nobody has managed it.

This is why horizontal platforms keep failing. They look at construction and see fragmentation, and they assume the solution is aggregation — put everything on one platform. But the fragmentation isn't a surface problem caused by lack of technology. It's a fundamental characteristic of an industry that builds custom products in custom locations using custom teams.

The graveyard of construction marketplaces

It's worth listing what's been tried, because the pattern of failure is instructive.

Sub-finding marketplaces. "Match GCs with subs, like Uber for subcontractors." These failed because GCs already know their subs. The problem isn't discovery — it's coordination. And the subs who need to be "discovered" through a marketplace are usually the ones without the track record to get work through relationships.

Material procurement platforms. "Buy construction materials online, like Amazon for rebar." These failed because construction materials aren't consumer goods. The specifications matter. The delivery logistics are project-specific. The credit terms are relationship-based. And the purchasing decision is made by someone who already has a supplier they trust and a phone number that works.

Bid management platforms. "Centralize the bidding process." These had more traction because they solved a real workflow problem. But they're not marketplaces — they're software tools. They don't aggregate demand or create a two-sided network. They make an existing process slightly faster.

Equipment rental marketplaces. "The Airbnb of heavy equipment." These had the closest thing to a marketplace use case, because rental equipment is more standardized. But even here, logistics, insurance, maintenance state, and the relationship between the rental company and the contractor limit how much a platform can abstract away.

The common failure mode is the same every time: the startup assumes that construction wants what other industries have — a central platform — and doesn't account for the fact that construction's fragmentation is load-bearing. It exists for reasons, and those reasons don't go away because you built an app.

So what actually works?

If marketplaces don't work, what does? The answer is more specific and less glamorous: tools that structure the data within existing workflows and relationships.

Construction doesn't need a front door. It needs connective tissue — a way to make the existing network of relationships, specifications, and transactions more productive without trying to replace them.

Equipment procurement is the clearest example. Today, procurement starts from zero on every project. A 400-page bid package arrives. Someone pages through the spec book and mechanical drawings to find the equipment. They build a spreadsheet. They email manufacturer reps. They track quotes in their inbox. They compare options in another spreadsheet. The equipment gets bought, the project closes, and all that data disappears into a project folder nobody will ever open again.

The next project starts from zero. Same GC, same equipment types, same manufacturers — but no memory of what happened last time. No structured record of pricing, lead times, or approval rates. No way to use past purchases for better terms on future ones.

This is the fragmentation that actually matters. Not the lack of a marketplace — the lack of structured data. Every project is an island because the data from each project is trapped in unstructured formats: PDFs, emails, spreadsheets that live and die with the project.

What changes when you structure the data

When equipment procurement data is structured — when every spec, schedule, quote, and selection is captured in a format that can be queried, compared, and analyzed — things change in ways a marketplace never could.

Memory across projects. You quoted a Trane chiller on a hospital project six months ago. Now you're bidding a similar project. Instead of starting over, you pull up the previous quote, check if the specs are comparable, and use actual pricing data from a recent transaction. Your bid is tighter because it's based on real numbers, not estimates.

Pattern recognition across your portfolio. You've procured air handling units on 40 projects this year. Structured data tells you that Carrier AHUs average 12-week lead times, while Daikin averages 8 weeks. It tells you that approval rates for proposed alternates are 60% higher when you include performance comparisons. It tells you that three-quote coverage correlates with 8-12% lower equipment costs. These patterns are invisible when the data lives in email threads.

Stronger vendor relationships. When you can show a manufacturer rep your purchasing volume across 30 projects — specific models, quantities, annual spend — you're a strategic account, not a one-off buyer. That visibility is impossible when every project's procurement data is scattered across different inboxes and drives.

Value demonstration to owners. An owner asks: "Why should we use your team for procurement?" Today, you say something about experience and relationships. With structured procurement data, you say: "On the last 15 comparable projects, our equipment procurement process saved an average of 11% compared to initial budget, and our average lead time performance was 3 weeks ahead of schedule." That's a different conversation.

None of this requires a marketplace. It doesn't require anyone to change their relationships or adopt a new platform for finding subs or manufacturers. It requires structuring the data that already flows through your procurement process — data that's currently being generated and immediately discarded on every project.

The connective tissue construction is missing

The reason construction has no front door isn't that the industry is backward. It's that construction's complexity resists the simplification that marketplaces require. You can't reduce a $50M hospital project to a listing. You can't rate a mechanical subcontractor like an Uber driver. You can't comparison-shop chillers like hotel rooms.

But you can structure the data. You can build systems that remember what happened on past projects and apply that knowledge to future ones. You can create procurement workflows that capture every quote, every selection, every approval in a format that accumulates value over time.

That's the connective tissue construction is missing. Not a marketplace — a memory layer. A way to make every project smarter than the last one because the data from all of them is structured, searchable, and useful.

Construction doesn't need a front door. It needs the data infrastructure that makes its existing network of relationships, expertise, and transactions compound over time instead of resetting to zero on every project.

That's what we're building.