The 8 components of a digital showroom (in build order)

A digital showroom looks like a website. It is not built like one.

A website is a stack of pages. You write copy, you choose imagery, you set a navigation, you launch. You can ship a perfectly serviceable marketing site in a long weekend if you already know what you want to say.

A digital showroom is a stack of decisions about what the buyer can touch and what they cannot. It carries product data, configuration rules, real prices, a 3D model that has to render on a phone at the bus stop, a PDF generator, a CRM webhook, and an ad pixel that has to fire at the right moment or the marketing spend goes blind. The visible page is the smallest part of it. The back office is the project.

We thought the 3D was the hard part. The 3D was the easy part. The hard part was deciding which roof types are allowed at which post height and how the dealer markup carries through.

The mistake that kills most first attempts is treating the showroom like a website project. You ship a beautiful 3D viewer, the buyer plays with it for two minutes, then leaves because they cannot see a price, get a PDF, or hand it off. Or you wire a flawless pricing engine to a flat product card and the buyer never spends long enough on the page to use it. Or you build the whole thing and forget the ad pixel, so the campaign that was meant to feed it cannot tell whether it works.

The right way to think about it is a real showroom you can walk into. The catalog on the front shelf decides what the buyer believes you sell. The configuration depth decides what the buyer can change without asking a salesperson. The glass case at the back decides what gets gated behind a conversation. The back office decides what data leaves the room and where it lands. That mental model maps cleanly onto the build.

The 8 components, in build order#

Read these as a checklist for a planning meeting, not as a feature list to admire. The order is load-bearing. Skip number 1 and number 4 has nothing to render. Skip number 2 and number 3 prices configurations you cannot actually build.

1) Product catalog#

Your real SKUs, not a stock library of generic 3D models.

A digital showroom needs the same starting data as your warehouse. A list of every product line you sell, each with its real codes, real material options, real size envelopes, real accessory list, real lead times. If you sell three pergola ranges and a glass roof, the showroom carries four product lines with their real names and their real ranges, not four “demo pergola” placeholders adapted from the vendor’s library.

Two practical decisions sit here. Which product lines belong on the front shelf in v1 (start with the two highest-volume lines, not the catalog edge cases) and which custom options stay off the public showroom entirely (anything that needs a structural engineer or a site survey lives behind a “request a custom build” flow).

Get this wrong and every other component drifts. The configuration rules end up about a generic pergola instead of your pergola; the pricing engine prices a list you do not actually sell; the 3D model renders someone else’s product. Spend the week here.

2) Configuration rules#

What the buyer is allowed to change, what they are not, and what combinations are valid.

The rules engine is the part nobody sees but everybody feels. Which roof type works with which post height. Which screen colour is allowed when the structure is anthracite. Which size needs reinforcement. Which accessory is incompatible with which beam. Which lead time changes when a buyer adds integrated lighting.

The rules library is also the only part of the build that is genuinely yours. A 3D vendor can ship you a pergola model in a day, but they cannot tell you that your louvred roof needs a corner post at 4.5 metres. That comes out of the production manager’s head and goes into a structured rule set. If the rules engine is weak, the showroom prices builds that the install crew refuses on day one.

A good rule set is boring to write and dramatically reduces the wasted-quote rate downstream. A rule that fires “this combination needs an engineer” the moment a buyer crosses a span limit saves three calls and one site survey for every fifty configurations.

3) Pricing engine#

A real number, in the buyer’s currency, computed from your real cost rules, your real margin, your real installation modifiers by region.

This is where most cheap showrooms quietly give up and hand the work back to your sales rep. A “starting from” placeholder is not a pricing engine. A pricing engine takes the configuration, applies the BOM, applies the labour, applies the regional modifier, applies the dealer margin if you run a network, applies the tax, and returns a defensible number.

Three patterns to plan for. The first is a dealer markup, where the parent showroom carries factory pricing and each dealer site carries a per-dealer multiplier. The second is the regional modifier, where the same configuration costs more in Boca Raton than in Bratislava because of the install economics. The third is the discount lever, where the sales rep can override the public number with a justified discount on the priced PDF.

Without a real pricing engine you have a 3D viewer. Buyers can play with the model, but they cannot decide. A decision needs a number.

4) 3D model and rendering#

Parametric, mobile-first, and tied to the catalog from step 1.

Parametric means the model rebuilds when the buyer changes a value. Width from 3 metres to 4, the post count and the beam span recalculate. Roof type from louvred to retractable fabric, the framing redraws. Post finish from anthracite to oak, the texture updates across every visible surface. A scene-based 3D viewer (one model per preset) is not what we are talking about. That is a slideshow.

Mobile-first means the model has to load and stay smooth on a 4G handset, not just on the showroom Wi-Fi. The underlying technology is WebGL, which has been mature since roughly 2017 and runs natively in every modern phone browser. The work is in the model weight, the texture sizes, the draw calls, the loading sequence. A good 3D engineer can hold a parametric pergola under 3 megabytes and first paint under 2 seconds on a tired phone.

This is the visible work the marketing team will ask about most. Resist the urge to start here. Without the catalog, rules, and pricing in place, a beautiful 3D model renders nothing that the buyer can act on.

5) Intent capture#

The 3 to 5 short questions that qualify the buyer without breaking the configuration flow.

The temptation is to ask everything. Postcode, install address, planning permission, budget, timeline, decision-maker, project scope, finance. Ask all of it and the buyer leaves before the PDF. Ask none of it and the sales team gets a stack of configured PDFs from people in the wrong country, the wrong budget bracket, or the wrong timeline.

The discipline is to pick the 3 to 5 questions that change how the sales rep approaches the call. Country and region (changes the install economics). Timeline (decides which queue the lead goes into). Project status (a planning-stage homeowner is a different conversation from a contractor with planning permission already in hand). Optional budget (only if your AOV varies by more than 3x; otherwise it is friction with no payoff).

Intent capture sits between the configuration and the PDF, not at the top of the funnel. The buyer earns the right to be asked by spending five minutes on the configurator. Asking before they have committed any attention loses them.

6) PDF output#

The priced spec that becomes the first call.

The PDF is the moat. Two days after a buyer leaves your showroom, a slower competitor sends them a handwritten quote in an email. By then the buyer already has a designed document on their desktop with your logo, your spec, your price, your lead time, and a link back to their configuration. The competitor is now negotiating against a real document. The buyer is not.

A serious PDF carries the configuration line by line (every option the buyer chose, with the part code if you sell to a trade buyer), the 3D render of the final build, the priced totals, the lead time, the next-step CTA, and a unique configuration link so the buyer can come back and adjust without starting over. The PDF lands in the buyer’s inbox as an attachment and in the seller’s CRM as a payload.

If you are templating the PDF in v1, build it to match your brochure exactly. The PDF is the first piece of brand the buyer takes home. A configurator that emails a screenshot is not a configurator.

7) Analytics and ad pixel#

GA4, Meta CAPI, Google Ads, and your tag manager, wired to the same five events.

Without analytics the showroom is invisible to your marketing budget. You will spend money on ads, the ads will drive traffic, the traffic will configure, and the dashboard will show you click-through-rate to the configurator landing page and then a black hole. Cost-per-qualified-lead, cost-per-configured-PDF, cost-per-submitted-inquiry. None of those exist unless the showroom fires named events at the moment they happen.

The five events to instrument from day one. configurator_started when the buyer first interacts with the model. configurator_priced when the pricing engine resolves a complete spec. lead_submitted when the buyer submits intent capture. pdf_downloaded when the priced PDF is generated. submitted_high_intent, fired as a server-side conversion, when the configuration has spent more than the median session time in the model and reached a complete spec.

Server-side conversions matter because iOS 14 attribution eats client-side pixels for breakfast. A showroom that only fires browser events undercounts conversions by 30 to 50 percent on iOS traffic, and Meta’s optimization algorithm reads that as “this audience is cheap.” The CAPI lambda or equivalent server-side endpoint is not optional in 2026.

8) CRM webhook#

Where the lead lands, in what shape, with what fields ready for the sales rep to act on inside 60 seconds.

The CRM webhook is the handoff. The configurator fires a JSON payload to your CRM the moment the buyer submits, carrying the configuration, the priced spec, the intent answers, the UTM source, the device, the country, and a link to the PDF and a link to come back into the same configuration. The CRM creates the deal, sets the stage, assigns the rep, fires the first follow-up email, and timestamps the inbound.

Three integration patterns. The first is a direct webhook to a native CRM API (HubSpot, Pipedrive, Salesforce). The second is an iPaaS bridge (Zapier, Make, n8n) when the CRM is something custom or older. The third is a file drop for dealer networks where each dealer runs their own back office.

Without the CRM webhook the sales rep checks email manually, copies and pastes the spec into the deal record, and misses the 5-minute follow-up window that turns a configured lead into a real conversation. Built right, the rep opens their CRM, sees the configured deal at the top of the pipeline, opens the priced PDF, and starts the call with “I see you configured a 4 by 5 metre louvred build with anthracite posts on Wednesday.” The conversation begins where it used to end.

Build vs buy vs template#

Three paths. The math decides which one is right for you, not the marketing on any vendor’s homepage.

1. 1 day Premade template Existing category model, your products mapped on top, live on a public URL. The standard build. Custom premade builds extend to about 5 to 10 days when a brand needs bespoke styling or rule edits.
2. 4-6 mo Custom freelancer 30,000 to 60,000 dollars typical, plus monthly maintenance. The vendor learns your category on your dime; the showroom ships when their backlog clears. Most miss the first launch date.
3. 12-24 mo In-house build Around 150,000 euros of engineering time before first buyer. Requires a 3D engineer, a CRM engineer, a pricing engineer, a DevOps engineer, and a product manager who can hold all four in their head.
4. €20K-50K Generic 3D vendor Configure One, Threekit, and similar generalist 3D platforms charge this for the setup that adapts their engine to your category, then monthly on top.

Pricing ranges drawn from typical market quotes for outdoor living configurator builds in 2025 to 2026.

The premade path works when your category has a known template. Pergola, veranda, garden room, awning, carport, glass roof, outdoor kitchen, fencing, sauna, modular cabin. The 3D model exists, the rule library exists, the pricing patterns exist; your brand, your products, and your pricing map on top. Implementation went from about 6 months in 2022 to about 1 day for premade builds and 5 to 10 days for custom ones, which is the single largest cost shift in the category in the last three years.

The custom-freelancer path is the right call only when your product has no category cousin. A bespoke industrial machine, a specialised lab instrument, a category nobody else sells. If a premade exists for your line, paying a freelancer 30,000 to 60,000 dollars to rebuild it from scratch is paying twice.

The in-house path is rarely the right call. The honest case for in-house is when your product configurator is itself the product (you are selling configurators, or your IP is a unique rules engine the market does not have). Otherwise, 150,000 euros and a year of engineering time spent on a 3D viewer is 150,000 euros and a year not spent on the parts of the business that actually compound.

The generic 3D vendor path looks attractive on a demo (their engine renders anything) and falls apart on month two of implementation, when the project team has to teach the vendor what a louvred roof is, how dealer markup actually works in your category, and why a 4.5-metre post is not optional. The bill arrives anyway.

What to skip in v1#

The temptation in v1 is to ship every feature you have ever wished a showroom had. The discipline is to ship the spine and resist the rest until you have buyers using it.

The honest answer to “what should I cut to ship faster” is almost always “everything you added in the third planning meeting.” The first meeting picked the spine. The second meeting found a few real edge cases. The third meeting started decorating. Cut the decorations.

Mobile is not a checkbox#

Over half of outdoor living traffic is on a phone, and the share rises every quarter as the Pinterest and Instagram referral mix grows. A showroom that loads in 800 milliseconds on a fibre connection and 8 seconds on a tired 4G handset is, for the median buyer, an 8-second showroom.

Test on the conditions buyers actually use. A mid-tier Android handset (most of the world is not on the latest iPhone). A 4G connection with a real signal, not the empty fast-lane of a developer’s home Wi-Fi. A real ad-traffic flow, with the buyer arriving on the configurator landing page from a tracked URL with a UTM and the configurator pre-selected to a real product.

Three numbers to hold against the build. First paint under 2 seconds on the median 4G phone. First model interaction (the buyer can rotate, resize, or change a value) under 4 seconds. Full configuration including a render and a priced spec under 6 minutes on average. Below those, the funnel decays exponentially at every step.

If the vendor demo runs only on a desktop or on showroom Wi-Fi, ask them to open the same configurator on your phone, on your own data, before you sign anything. That single check has killed more bad vendor decisions than any other.

How you know it’s working#

Four metrics. Track them weekly, judge them quarterly.

Configurator engagement rate. Sessions that interact with the model divided by sessions that land on the page that hosts it. A healthy number sits between 25 and 45 percent. Below 20, the page is not framing the showroom well; the buyer scrolls past without engaging. Above 50, the page may be filtering too hard at the top and you are losing volume that would have configured.

Configured-to-PDF rate. Sessions that complete a full configuration and generate a priced PDF, divided by sessions that engaged with the model. A reasonable target is 30 to 50 percent. Below 25, look at where the funnel breaks: is the configuration too long, is the pricing engine breaking on edge cases, is the intent capture too aggressive.

Lead quality score. A simple composite. Configuration time, completeness of intent capture, country match (for in-region categories), price band (do they fall inside your AOV), and whether the rep can reach the buyer in the first 24 hours. Most CRMs let you compute this on the fly. The aggregate quality score should rise quarter over quarter as the showroom and the rule library tune; if it does not, the marketing mix is sending the wrong traffic.

Cost per qualified lead. Ad spend in the period divided by qualified leads generated, where “qualified” means the lead matches the quality score threshold. This is the only marketing number that ties the showroom to the P&L. Industry baselines run 100 to 300 dollars per qualified lead for outdoor living categories in 2025. Showrooms running for 90 days commonly cut that by 50 to 80 percent because the configurator filters the volume before the rep ever sees it.

These four together tell you whether the showroom is doing the job. A showroom with a great engagement rate but a dismal lead quality score is a marketing problem (wrong audience). A showroom with a great quality score but a low configurator-to-PDF rate is a build problem (the pricing engine or the intent capture is breaking the flow). The metrics tell you where to look, not what to do.

A digital showroom is built like a real showroom, not like a website. The catalog, the rules, the pricing, the model, the intent capture, the PDF, the analytics, and the CRM webhook are the eight things that have to work for the buyer to walk in, look around, decide, and leave a configured order behind. Ship the spine, instrument it, watch the four numbers, tune the rule library, and let the funnel compound.

If the next decision is which vendor to buy from rather than which architecture to build, the /compare/ section is the right place to start, the configurators gallery shows live category templates, and the companion piece on what a digital showroom actually is sits one step before this build guide.