Brightline Solar: from PDF quote form to 40-second estimate

Brightline Solar installs residential panels across Queensland and northern New South Wales. Good crews, honest pricing, ten years of five-star reviews — and a website whose quote journey was a downloadable PDF. A form you had to print, fill in, photograph and email. In 2024. The MD knew it was costing them work; what he didn't know was how much. Our estimate, after a fortnight of analytics archaeology and call-log listening: roughly two-thirds of high-intent visitors were evaporating between "Get a quote" and a human ever learning they existed.

This is how we replaced a dead document with a 40-second estimate engine, and what happened to the pipeline when we did. Figures are illustrative — every house, roof and market is different — but the shape of the result is real.

The challenge

Solar is a considered purchase with an impatient first step. A homeowner's interest spikes in a specific moment — a frightening quarterly bill, a neighbour's new array, a 40-degree weekend — and decays within days. Brightline's journey wasted the moment: the PDF asked for information the visitor couldn't possibly have at hand (roof pitch, switchboard photo, meter number), promised a callback "within 5–7 business days", and delivered a quote two to three weeks later. By then the moment had passed, or a competitor with a faster pencil had taken it.

Three specific failure modes surfaced in our research:

The form interrogated before it gave. Seventeen fields before any value returned. The psychology was all wrong — it read as homework, not help. We've written about this pattern generally in designing forms people actually finish: reciprocity first, interrogation later.

Every lead looked identical. A renter browsing out of curiosity and a cash-ready owner with a 45-square north-facing roof entered the same queue. Sales spent 60% of their time on calls that could never close.

The estimate lived in someone's head. One senior estimator, one spreadsheet, one deeply held fear of publishing numbers before a site visit. Meanwhile every competitor's ad screamed "instant quote!"

The constraint that defined the build: Brightline's sales team is exactly four people. The tool had to reduce their call load while raising close rates — not flood the queue with tyre-kickers.

The approach

Give the number first. Our entire architecture inverts the industry pattern. The visitor gets a real, defensible estimate — system size, annual saving, payback window — in under 40 seconds, before we ask for anything personal. Two inputs open the flow: address and quarterly bill. That's it. Address resolution runs against a geocoder, drops a pin on a map, and only then do we gently ask the visitor to nudge the pin onto their actual roof. The nudge game-ifies the one step that matters for accuracy.

Heuristics with honest labels. We built a shade-and-eligibility engine from mocked-but-principled rules — roof orientation from the building footprint, shading penalties by surrounding canopy, tariff assumptions by postcode band — and, crucially, we showed our working. Every estimate carries a confidence band and a "why this number" drawer that lists the assumptions in plain English. Solar has a trust problem; showing the maths was the single most commented-on feature in user tests. The same "impossible to look away, impossible to misunderstand" discipline we applied to Meridian Climate's council data explorer applied here to a sales tool.

Finance is content, not an afterthought. Rather than bolting a loan calculator to the end, the flow presents three routes side by side — cash, green loan, and a power-purchase-style plan — with total ten-year cost for each. Cash isn't pushed; the honest comparison is the persuasion. Roughly 40% of completers explore all three tabs, which told sales exactly what to prepare before the call.

Qualify with generosity. After the estimate, the form asks for the four things that actually predict a job: ownership status, roof age, decision timeframe, and whether anyone else is quoting them. Each question is framed as benefiting the visitor ("so we don't quote you for a roof that needs work first"). Lead score is computed inline; high scorers see a live booking calendar with real crew-visit slots, low scorers get a genuinely useful savings guide and a polite nurture track. The experiment design discipline we preach — pre-registered hypotheses, pre-agreed kill criteria — governed every A/B test we ran here, including the one that killed our beloved animated roof-tilt slider (it entertained, it didn't qualify).

A quote the office can stand behind. Estimates generate server-side, are stored with their assumptions, and expire after 30 days with a graceful re-estimate path. Sales inherits a lead record containing the actual numbers the visitor saw — no more "well, the website said" conversations.

The outcome

Nine months post-launch, illustrative results:

• Qualified leads up 71% against the same seasonal period, with "qualified" defined strictly: owner-occupied, roof viable, timeframe under six months.
• Cost per qualified lead down 33%, as paid spend was re-cut toward the keywords and suburbs the engine proved converted.
• Estimate completion: 64% of people who enter an address finish the flow — against 9% who ever returned the PDF.
• Sales time per closed job down ~40%. Reps open calls with the system size, shade assumptions and finance preference already known. The first call got 12 minutes shorter and noticeably friendlier.
• Quote-to-close window: 16 days → 6 days for calendar-booked visits, because the moment of intent is met while it's still warm.

You can play with a working slice of the engine — mocked data, same mechanics — in the Brassfern lab. If your own pipeline has a PDF-shaped hole in it, this is exactly the kind of growth engineering we do, and the brief form is mercifully short.

What we'd tell anyone building a quote engine

First: the estimate is the marketing. Spend your craft budget on the quality and transparency of the number, not the landing page around it. Second: qualification questions convert better after value is delivered, not before — the order of operations is the design. Third: give sales a record of exactly what the visitor saw. The tool isn't finished when the form submits; it's finished when the first call starts two moves ahead.