The situation
Quoting at our manufacturing company used to be a multi-step grind. To send a single quote, someone had to:
- Copy a master sheet
- Create a new folder
- Make edits across multiple pages of the sheet
- Pull pricing from a different reference
- Manually format the customer-facing output
Realistic turnaround was a discovery call plus 45 minutes of post-call work. If a customer wanted three configurations to compare, that was three rounds of the same dance — and we mostly just didn’t offer that, because nobody had the time.
We looked at buying a fully spec’d CPQ platform. We were close. But every version we evaluated required a heavy migration of our existing logic into someone else’s data model, on someone else’s pricing schedule, with our team learning a new tool on top.
The math didn’t work. So we built our own.
What we built
The focus here is data and spreadsheet mastery, not AI for AI’s sake. The win came from collapsing a sprawling system into a clean one.
- A single input/output page — the salesperson enters the customer’s needs in one place, and the quote comes out in one place. No more navigating across sheets and folders.
- A pick-sheet system — instead of typing, the salesperson selects from validated options. Adding a new product or option is a structured, low-error process.
- Apps Script automations — handle the copying, calculating, and document generation that used to be manual.
- An HTML data layer — turns the streamlined sheet into something that can be deployed across teams and, when ready, even rebuilt customer-facing.
The system supports multi-quote proposals — basically a configurator. A customer who wants to compare three configurations gets all three in one document, generated in the time it used to take to make one.
What changed
| Before | After | |
|---|---|---|
| Time from call to quote sent | Discovery call + 45 min | Quote built during the call, sent the moment it ends |
| Multi-quote proposals | Rare — too much effort | Standard — happens by default |
| Edits after sending | Painful, error-prone | Trivial |
| Onboarding a new salesperson to quoting | Days | Hours |
The deeper shift isn’t the time savings. It’s that the quote stopped being a follow-up artifact and became part of the conversation. That changes how the sales call itself runs. Customers feel heard. Decisions land while we’re still on the line.
What I almost did instead
The CPQ-versus-build decision is the part of this story I think most owners need to hear.
I’m not anti-platform. There are real cases where buying makes sense. But every CPQ vendor we talked to was solving a problem one or two sizes bigger than the one we actually had — and charging accordingly. The migration alone would have absorbed more time than the build.
The instinct is always to assume someone else has solved the problem better than you can. Sometimes they have. Often they haven’t — they’ve solved a similar problem for a different size of company, and you’d be paying for the parts you don’t need.
The lesson I take into client work: if the gap between your situation and the platform is small enough to bridge by building, building usually wins. It costs less, you own it, and the system fits your operation instead of the other way around.
The spreadsheet-mastery point
The unsexy truth: most “AI quoting” pitches you’ll see on LinkedIn won’t help you until your underlying data is clean. We didn’t start with the AI layer — we started by making the data sane. Once that was done, the AI work that came later actually had something to stand on.
That’s the pattern I keep coming back to. Foundation before roof. It’s true for quoting tools. It’s true for everything else.