The mobile product configurator test: 7 checks before you sign

It is 9 in the evening. The dinner plates are still on the patio table. The homeowner is half-watching a streaming show and half-thinking about the pergola conversation that started in the garden centre last Saturday. The phone is in their hand because the phone is always in their hand. They tap a Google search, a sponsored result, a landing page, then an “open the designer” button. Three seconds pass. Five. A grey rectangle hangs on the screen where the pergola is supposed to be. By eight seconds, they have closed the tab and gone back to the show. By morning, they will have forgotten the search ever happened.

That single moment is what every outdoor living vendor pitch is competing for. Not the sales rep demo on a Mac in a showroom. The buyer with a phone, in the evening, on the patio, deciding inside ten seconds whether to keep tapping or close the tab.

By Monday morning, 40 percent of our inquiries already came in as configured 3D quotes. The first conversation was different from day one.

That 40 percent number lives or dies on the phone. The dealer who wins it has a configurator that loads on the patio. The dealer who loses it has a configurator that demoed beautifully on the vendor’s desktop and stuttered the first time a real homeowner opened it on cellular.

Why mobile broke product configurators#

The first generation of 3D product configurators were built for showrooms and trade booths. Big screens. Reliable wired internet. A salesperson standing next to the buyer to walk them through the controls. Texture sets in the hundreds of megabytes. Render loops that ran the CPU hot. Hover states for every secondary action. None of that survived the migration to a buyer-facing website on a phone, but the engines under the hood were already in production by the time outdoor living moved its traffic to mobile.

Three things break in the move. First, the payload. A desktop-first 3D engine commonly ships 30 to 50 megabytes of textures, geometry, and runtime before the first frame appears. On a hardwired connection that is invisible. On a 4G uplink in a residential area in the evening, with the household streaming on the same connection, that download is an 8-second wait the buyer does not give. Web performance research has been steady on this point for a decade. Core Web Vitals from web.dev set the Largest Contentful Paint threshold at 2.5 seconds. Configurators that miss it lose buyers before the page resolves.

Second, the CPU. A 3D engine that runs hot eats the phone battery and triggers thermal throttling within minutes. The pergola that looked smooth at the start of a session turns into a stutter at minute three. Worse, the phone is hot in the buyer’s hand. People do not buy from products that physically hurt to use.

Third, the interaction model. Touch is not a mouse. There are no hover states. There are no right-clicks. Pinch-to-zoom collides with the native browser pinch. Drag-to-rotate collides with native scroll. A desktop configurator that gets ported to mobile without rebuilding the controls feels like fighting the phone every time the buyer tries to look at the underside of the roof slats. That fight is the moment they leave.

None of these problems are unsolvable. All of them are choices the engineering team makes at the start of the build. The trouble for outdoor living dealers is that most generic 3D vendors made the choices in the other direction five years ago, and selling the work to redo them now is harder than pretending it does not matter.

What “mobile-ready” really means#

Vendors say “of course it works on mobile.” Almost every demo does, technically. The question is whether it works the way the buyer needs it to work. Five concrete things separate a configurator that loads on the phone from a configurator the phone buyer will finish.

Performance budgets, written down and held to. The mobile build has a separate target for first paint, first interactive moment, and steady-state frame rate. A real number, in seconds, that the engineering team commits to and the QA team measures. Without that number the build drifts toward whatever the desktop tolerated, which is always too slow.

Touch targets the buyer can hit. Every interactive element on the configurator hits the 44 by 44 pixel minimum from the WCAG touch target guideline, with breathing room between adjacent targets. A “change post finish” tap that the buyer keeps missing because the swatch is too small is the same thing as the buyer leaving the configurator.

Gesture-based rotate and zoom. The buyer wants to see the pergola from a different angle. They want to walk around it with a finger drag. They want to pinch to zoom on the louver detail. Every gesture has to do exactly what the phone has trained the buyer to expect, with no hover-state fallback and no clumsy on-screen controls that occupy half the visible canvas.

No hover-only UI. Hover does not exist on a phone. Every tooltip, every secondary control, every preview-on-pointer interaction needs a tap equivalent or it disappears entirely on mobile. Vendor demos commonly miss this on the first review.

Sub 3-second time-to-first-render on a mid-range Android on 4G. The aspirational target is 2 seconds. The acceptable target is 3. The threshold at which the buyer leaves is roughly 4. Measure it once, write the number down, and refuse to ship until the build hits it.

The 7 mobile checks to run before signing#

A vendor demo on a desktop tells you nothing about how the configurator behaves where it matters. Run these seven checks on every vendor on your shortlist before you sign a contract. They take about 20 minutes. They will save 12 months of regret.

1. 01

   Open it on your phone, on cell data, not Wi-Fi

   Disconnect from the showroom Wi-Fi. Drop your phone to 4G or 5G. Open the configurator from a cold start. Count seconds. If the model is still not rendered at 5 seconds, the configurator is not mobile-ready, no matter how the rep frames it. The buyer in the garden centre car park is on cellular too.

2. 02

   Measure first paint and first interaction time

   Use Chrome on Android with the dev tools open, or Safari with the Web Inspector. Look at Largest Contentful Paint and Time to Interactive. Web.dev's guidance is 2.5 seconds and 100 milliseconds respectively. Configurators that miss those numbers lose buyers between the tap and the model.

3. 03

   Configure the whole product with one thumb

   Put the phone in one hand. Walk through every step. Change the size, the roof, the posts, the finish, the screens, the accessories. If any step needs two hands, a stylus, or a desk to lean on, the configurator was not built for the way buyers actually use phones in the evening.

4. 04

   Test in landscape and portrait

   Rotate the phone. The configurator should rebuild cleanly, keep the buyer's selections, and use the new aspect ratio for the 3D canvas. Common failure modes: the model resets to the default build, the right-side panel disappears, the controls overflow the screen edge. Each one is a deal lost on the rotation.

5. 05

   Check the PDF generation on mobile

   Submit the configured build. Watch what happens. The PDF should generate server-side and arrive in the buyer's inbox within roughly 30 seconds, with a thumbnail preview the buyer can open on the phone without downloading. If the page hangs while the PDF builds client-side, the buyer leaves before the file resolves.

6. 06

   Verify the submit fires the tracking pixel and CRM webhook on mobile

   Open the network panel in the mobile browser before you submit. After submission, confirm three things fire: the Meta pixel, the Google Ads conversion, and the CRM webhook. Each one. If they fire on desktop and silently fail on mobile, the marketing team is underreporting the channel and the sales team never sees the lead.

7. 07

   Hand the phone to a non-technical friend and watch them try

   The single highest-signal test in the list. Pick someone on your team who has never seen this configurator before. Open the link, hand them the phone, ask them to design a pergola. Do not coach. Do not narrate. Watch where they get stuck. The first 90 seconds will surface every flaw a slide deck hid.

The first three are gates. Fail any of them and the configurator is not viable on the phone, full stop. Check four and five tell you whether the vendor understood the device the buyer actually owns. Six and seven tell you whether the configurator was built once for the showroom or built twice for the buyer.

A longer walk through the broader vendor decision sits in the 3D configurator software buyer’s guide, with five buying paths and the trade-offs between them.

What desktop demos hide#

The vendor rep demoing the configurator is using a Mac, plugged into ethernet, in a controlled office with a curated sample build that loaded once at startup and is now cached. The buyer evaluating the configurator is using a four-year-old iPhone, on holiday in Spain, on a hotel Wi-Fi shared with 200 other guests. The gap between those two contexts is the entire reason mobile-ready configurators are rare.

Three patterns repeat on every desktop-first demo. The 3D model loads in under a second because the rep opened it 40 times this week and every asset is in the browser cache. The interactions feel fluid because the laptop is running a 12-core CPU with a discrete GPU. The PDF arrives instantly because the network round-trip across an office ethernet connection is roughly 4 milliseconds. None of those three conditions exist on the buyer’s phone.

The fix is not to make the desktop slower. It is to evaluate every vendor on the worst device a real buyer is likely to own. The rule we use internally is “if it works on a three-year-old Android phone on holiday-Wi-Fi, ship it.” Anything that needs better hardware than that is leaving buyers on the table.

We thought the demo was great. Then we put it on our own website and watched our heatmap. Nobody scrolled past the spinner. The model never loaded fast enough for the buyer to wait.

The point of the mobile evaluation is not to be hard on the vendor. It is to surface, before contract signing, the exact gap between the rep’s demo machine and the buyer’s pocket. Closing the gap is fixable. Discovering the gap nine months into a six-figure implementation is not.

Mobile traffic patterns in outdoor living#

The mobile-first argument is not a UX preference. It is a marketing data point. The data we see across outdoor living dealers, in five countries, with both paid and organic traffic, runs roughly like this.

1. 55% Mobile share of outdoor living sessions Median across 40+ outdoor living dealer sites we run analytics on. Range: 48 percent in B2B-heavy Germany to 68 percent in residential-first UK.
2. 8–10pm Daily mobile session peak Two-hour evening window accounts for roughly a third of weekly mobile traffic. Desktop peaks midday at the office.
3. 62% EU mobile sessions on Android Mid-market premium products skew Android in central Europe; UK and US premium tilts iPhone. The mid-range Android is the worst-case device, and it sets the floor.
4. 40% Caribbean Blinds inquiries via 3D configurator The 3D configurator share at one of the UK's best-known pergola dealers. Most of those submissions are mobile.

Aggregate mobile usage across outdoor living dealer sites, plus the named anchor from /customers/stories/caribbean-blinds/.

Three patterns matter for the configurator decision. The mobile share is over half and climbing. The evening peak does not overlap with the desk-based research window the marketing team plans for. The device skew points at the mid-range Android as the worst-case scenario, not the latest iPhone in the showroom drawer.

The Caribbean Blinds proof point is the one to anchor on. 40 percent of inquiries arriving as configured 3D quotes is not a small slice of leads. It is a structural change to how the sales team’s Monday morning works. That 40 percent number depends on the configurator working on the device most of those buyers are holding when they submit. The dealer who ignores the mobile build is volunteering to be a smaller fraction.

A wider view of the device behaviour sits in the pergola configurator guide, which walks through the configurator anatomy and the demo checks on the configurator decision in general. The mobile question is one input among several in the wider buyer’s decision.

When mobile-ready is not the priority#

A short anti-fit list. Mobile-first matters most when the buyer is a homeowner, the AOV is between 5,000 and 50,000 euros, and the traffic mix is consumer-led. Three contexts where the rule loosens.

For everything else in outdoor living, residential cabinetry, mid-market modular construction, and the rest of the buyer-facing configurator world, the phone is the device. The desktop is a secondary surface. The build budget allocates accordingly or the project underperforms.

The closing argument#

The mobile evaluation is not the last item on the vendor checklist. It is the first item. If the vendor cannot ship a configurator that loads on a four-year-old Android on hotel Wi-Fi, the rest of the feature list does not matter. The buyer who would have bought is in the streaming feed.

For the dealers running this evaluation for the first time, the practical advice is short. Pick the vendor’s demo. Open it on your own phone. Open it on cellular. Run the seven checks above. If anything fails check one, two, or three, move on to the next vendor. The other features can be argued over. The phone cannot.

The 3D configurator software buyer’s guide covers the wider vendor decision, the pergola configurator guide covers the configurator anatomy in one product category, and the configurators gallery shows what mobile-first looks like in real dealer sites. The buyer in the evening, on the patio, on the phone is the one to design for. Everything else follows from that.