The showroom and the online store were never two businesses, and the sale and the service were never two departments. They were stages of one process, and the gaps between them had been treated as natural losses because nobody had examined them as design.
By 2025 the company ran on connected systems rather than on the memory of whoever answered the phone. Product data in one place, service delivery in one place, commercial activity in one place, and the joins between them built on purpose. That is not a technology achievement. It is the difference between a company that compounds what it learns about its customers and one that starts over with every enquiry.
The technology was the easy part in both halves. The valuable part was standardising the work first — which is why this case sits on top of the previous one, and why it could not have been done in the other order.
WHAT IT CHANGED
ON THE SERVICE SIDE, the quote moved from next-day to same-visit, and the decision from a three-to-ten-day gap to a conversation in the customer's living room. The close rate rose from 25–30% to 40–50%. Job volume increased roughly 10–15%. Customer feedback improved, for a reason that is not complicated: people prefer knowing.
ON THE RETAIL SIDE, roughly 20% of showroom visitors who left without purchasing later returned and bought — which measures precisely the thing that had been broken. Combined in-store and online retail sales rose approximately 55% in 2025 against 2024.
Both figures sit inside a business that was changing in other ways at the same time — new showroom, new staff, new procedures — and I would not claim either is attributable to a single feature. What the systems changed is that outcomes became measurable and repeatable rather than dependent on who happened to be running the job that week.
RESULT
The two halves were then joined deliberately rather than left as neighbours.
Airtable held the product and pricing data as the single source of truth, feeding both the website and the commercial CRM. HubSpot ran the commercial layer — connected to the website, the phone system, and the service CRM — so a call, a form submitted at midnight, and a referral out of a completed service job all arrived in the same place instead of in three people's inboxes. Housecall Pro continued to run service delivery, quoting and job control. RingCentral logged every conversation into both.
A customer could now scan a sample in the showroom, ask the site a question that evening, receive a preliminary estimate, book a measurement, be quoted in their own living room the following week, and watch the installation from their phone — and the company held one continuous view of that person the entire way through.
DECISIONS — CONNECTING THEM
Physical sample → NFC scan → standardised product data → transparent pricing → live product Q&A → visualisation in their own room → installation estimate → purchase or in-home consultation → retargeting
Two years later I did the same thing to the other half of the business.
I did not build this one. I specified the system, selected the vendors and ran delivery alongside our marketing director. The technical work was subcontracted and the project took roughly four months.
THE PRODUCT DATABASE CAME FIRST AND TOOK THE LONGEST.
Vendor information for approximately 2,500 samples was collected, cleaned and normalised into consistent SKUs, descriptions, box quantities and specifications, and held in Airtable as the single source of truth. Every supplier delivered a different shape of information and none of it was usable by a website. Nothing else could exist until that did.
THEN THE PHYSICAL-TO-DIGITAL LINK.
A Canadian NFC provider connected every showroom sample to its product page, so scanning a sample with a phone lands on exactly that product — no name to remember, no search to repeat.
THEN VISUALISATION AND PURCHASE.
Roomvo integrated so a customer could see a floor inside a photograph of their own room. Products purchasable directly through Shopify.
AND THE WEBSITE COULD ANSWER FOR ITSELF.
We connected a model trained on our own product database to live chat, so someone asking about a specification, a comparison or an installation requirement at eleven at night got a real answer from our catalogue instead of a contact form.
ONE DELIBERATE CONSTRAINT.
We never sold installation online. A wrong installation number costs more than a lost conversion, so the site published a preliminary estimate instead — typically landing within 10–25% of the final in-home quotation. That meant publishing a figure I knew was imprecise, in exchange for a customer who arrived at the consultation already knowing roughly what their project would cost. I would make that trade again.
AND THE SHOWROOM VISIT STOPPED BEING ANONYMOUS.
Once a customer scanned a sample and entered the site, tracking placed them into relevant retargeting audiences, so advertising returned them to the products they had actually considered rather than to flooring in general.
DECISIONS — THE RETAIL SIDE, 2024–2025
I tested this through 2021 and launched it in 2022.
A PRICED PROCESS LIBRARY.
Every operation the company performed — demolition by material type, installation by method, baseboards new or reinstalled, trim painting, staircases — described and priced per square foot inside the service CRM.
This is the part that could not have existed three years earlier. You cannot pre-price work until you have standardised it, and standardising it was the whole of the previous case. The production system was not only an operations improvement. It was the precondition for everything below. [CASE TWO →]
THE SALE SPLIT INTO TWO PARALLEL STAGES.
The salesperson goes to the house with an iPad Pro and a wide-angle lens and captures the job as they walk it: a photograph of every space, dimensions, what has to come out, what goes in, brand and specification, installation method, baseboards, trim, stairs. While they are still in the house, a second person at the office watches that information arrive and builds the quote live from the priced library.
That is the staircase line from case two, applied to selling instead of installing. One person captures and closes, the other prices, each owns a stage, and the work passes forward instead of waiting.
THE QUOTE ARRIVES BEFORE THE VISIT ENDS.
It goes to the customer by text and email and to the salesperson's iPad at the same moment, so the customer holds their own copy rather than reading someone else's screen. The salesperson closes in the room, with authority to discount and to arrange financing without calling anyone for approval.
AND THE QUOTE IS A PROJECT PLAN, NOT A PRICE.
It runs sequentially — a photograph and description of every process, from demolition through to final detail — walking the customer through their own project in the order it will actually happen. In a trade where the standard quote is a number and two line items, that justifies the price by showing the work, and sets the expectations for the job before the job begins.
THE CUSTOMER COULD THEN WATCH THEIR OWN JOB.
Every project ran through CompanyCam with a client-facing link: daily progress in photographs, notes, the plan for that day, and what was complete against what was still pending. Nobody had to call to find out what was happening in their own house.
AND EVERY CONVERSATION LANDED IN THE SAME RECORD.
The phone system was connected to the service CRM, so calls were logged automatically and conversation summaries attached to the customer's file. Anyone picking up a job could see what had been quoted, discussed and promised without asking whoever happened to take the call.
DECISIONS — THE SERVICE SIDE, 2021–2022
The service side was absent exactly where it mattered most — at the price and at the decision. The retail side was absent everywhere after the customer walked out of the building.
I fixed the service side first, because that was where the money was and because I already had what it required.
Analysis
Those are the same problem twice.
In both halves the company held information the customer needed, and the delay in delivering it had been accepted as a feature of the industry rather than recognised as a design choice. Nobody had decided a quote should take a day — it took a day because the person who measured was also the person who priced, and they could not do both at once. Nobody had decided a showroom sample should be anonymous — it was anonymous because nobody had connected it to anything.
Every one of those waits was removable. None of them were necessary. And each one was costing decisions.
THE REAL Problem
By the end of 2021 the company had six crews and I was still the constraint on every one of them. The work itself had been solved — that was the production system — but everything around the work still ran on conversations, memory and waiting.
A customer who wanted a floor installed had a salesperson visit, measure, leave, and send a quote the next day. From there the decision took three to ten days, often longer. During that gap the customer was comparing us to whoever else had measured that week, and we had no presence in the comparison.
In 2022 I opened a 4,500 square foot showroom holding around 2,500 physical samples. A customer could touch any of them and learn nothing about price, specification or installation without asking a representative. If they left without buying, the connection between that person and the sample they had spent twenty minutes deciding on was simply lost.
