"A pilot for a simple quoting process wouldn't make sense."
The client said that, not us. We were on a call this week with a manufacturer that runs eight divisions, and we'd just suggested starting the quoting rollout wherever the team was most fed up with spreadsheets. He cut straight past that. If the pilot proves the easy case, he said, it proves nothing.
He was right, and I've been on the other side of that decision often enough to know how tempting the easy division is.
Here is the case for starting simple, because it's a real case. You get a working tool in front of reps in weeks. Leadership sees quotes coming out of the CRM. The team learns the interface on low-stakes deals, and you build the goodwill you'll need when the hard divisions push back. Every project manager I know would take that.
The problem is what you learn. This company's industrial division sells truckloads of stocked SKUs against price lists already tied to each customer. That quote is a lookup. Any tool passes that test, including the spreadsheet they're trying to leave.
The architectural division is the other end of the spectrum. An architect sees a mesh facade on a building somewhere, wants their own version in a metal nobody has sourced, built into a sunshade that moves with the sun. Engineering gets involved. Detailing gets involved. Someone has to find out who can even supply the raw material. There is no SKU because the thing has never been made.
If your quoting tool can't hold that job, it's a price list with a better interface, and you'll find that out in month eight, after five divisions are live, when the sixth says no.
So the pilot went to the security division: complex enough to break a weak design, small enough to run as a pilot.
The second decision was where the custom code lives. HubSpot's native quoting (price books, quotes, the new contract records) handles the downstream half well. It does not, today, handle a manufacturer's spreadsheet logic: the dependencies between products, the rule that says a quote can't go out without a drawing attached, the SKU that gets added automatically in one division and never in another. The native roadmap will get there. This client couldn't wait for it.
The design our team landed on is a guided selling app built on HubSpot's UI extension cards. A rep answers a short series of questions, deliberately ordered, and the app outputs the products, quantities, list prices, discounts and dependencies the quote needs. From there, native HubSpot takes over: line items, price books, quote document, contract.
The reason for that split is maintenance. Everything downstream of the questions is HubSpot's problem to keep working. Only the questions and the logic behind them are ours, and that layer is small enough for one developer to own. The alternative, a fully custom quoting engine, would have been the client's most fragile system for the next decade.
The rep who has used the spreadsheet for ten years is the real acceptance test. If the new tool doesn't save that person time, they'll keep the spreadsheet open in another window and paste the answer in. So the design brief was blunt: no scrolling through hundreds of line items to find the one you need.
Two or three of the divisions quote fabrication through a specialised estimating tool. You feed it a STEP file, it flattens the part, checks it against the shop's machines and materials, and prices the cutting, bending and welding time. The team showed it off on the call and everyone on it was impressed.
Nobody is rebuilding that inside a CRM. It has an API, so the output flows into HubSpot for reporting and the deal record, and the estimating stays where the engineers already trust it.
The same call produced a request for takeoff software: read a 300-page drawing set, find every instance of a material tag, return the square footage. It would save the architectural team real hours. It's also not a quoting problem, and it went out of scope for the pilot on the spot. A pilot that tries to solve every division's every pain is a rewrite of the company, and rewrites don't ship.
Pick the complex division and the pilot costs more, takes longer, and produces a worse first demo. It also tells you, before the rollout budget is committed, whether the architecture holds. The easy division tells you that in month eight.
If you've run a CPQ or quoting rollout across divisions and started easy on purpose, I'd like to hear how it went. And if you started hard and regretted it, I'd like to hear that more.