The CPQ glass floor: why small manufacturers still quote in Excel

I have spent 25 years building sales configurators for manufacturers, first the big, expensive kind and lately the kind that runs on a language model. One question keeps coming back from the smaller ones: if CPQ pays for itself so reliably, why does almost nobody our size have it? The honest answer is on the price list.

What is the CPQ glass floor?

The industries where CPQ became standard practice share three traits. Their products are variant-rich: a crane, a switchgear cabinet or a ventilation unit can be built in thousands of combinations, and only some of them can actually be manufactured. Their commerce is quote-driven: nothing is sold from a shelf, every order begins as a specified, priced document. And they had the organisational muscle to carry what configuration software used to demand.

The first two traits are just as true for a 15-person manufacturer of trailers, gates or greenhouse systems. Their quote is the bottleneck too. They just never got the third trait, so they quote from an Excel sheet, a PDF price list and the one person who knows which options do not fit together. That gap between the need and the ability to buy is the glass floor.

What does CPQ cost a small manufacturer?

Take the smallest serious project there is: one product family, one market. On a traditional platform, the implementation alone typically takes four to five months and costs roughly €40,000 to €100,000 with a partner, before the first quote is sent. Then come the licences, priced per seat, with annual minimums and an enterprise sales process. The full breakdown is on our CPQ pricing page.

For a 500-person manufacturer that is a rounding error. For a 15-person one it is the marketing budget, and the minimums bite before the first quote is sent. Seat minimums, annual commitments and an enterprise sales process only pay off on large accounts, so the pricing model quietly rules out small projects. Nobody decided that small manufacturers should not have CPQ. The price list decided it for them.

Why does the cost that is not on the invoice matter more?

The licence is the visible part. Three costs never show up on a quote from a vendor:

• Calendar time. Four to five months from kick-off to go-live for a single product family, and a small company feels every one of those months.
• Your best product expert. The person who knows the rules spends most of the project in workshops. In a company of 15, that person is also quoting, supporting dealers and fixing production problems.
• Upkeep for life. A live model needs about half a full-time person every year, internal or partner, or it quietly degrades. We cover this in our list of mistakes to avoid when implementing CPQ.

Add those up and the market outcome is exactly what we see: there are essentially no small CPQ projects. It is not that small manufacturers do not need one. It is that the delivery model was never built for them.

Has this happened before?

It has, in the category next door. In the Siebel era, CRM was enterprise software: seven-figure projects, dedicated administrators and price lists that assumed hundreds of seats. Small companies kept their customers in spreadsheets, not because they had no customers, but because the category was not priced for them. Then the delivery model changed, first to SaaS and eventually to HubSpot's free entry tier, and CRM became standard equipment even in a five-person firm. Nobody asks any more whether a small company can afford a CRM.

CPQ is CRM's neighbouring category, sold on the same price-list logic. In my view it is running the same journey roughly ten years behind.

What breaks the floor?

What broke the floor in CRM was not a cheaper Siebel. It was a different way of delivering and pricing the capability. The same shift is now under way in configuration.

AI-native CPQ puts a language model in front of a constraint solver. The model handles the conversation and the general reasoning; the solver holds the product rules and refuses any configuration that cannot be built. Because far less has to be hand-built around the rules, the work moves from months to tens of hours, and the half-time upkeep of a traditional model goes with it. sailsrep.ai, the platform I work with, aims for about 40 hours of work to a working product and a launch on the customer's website in around two weeks. When a model costs tens of hours instead of months, the vendor no longer needs ten seats and a dedicated administrator to make the deal pay, and entry pricing without seat minimums becomes possible. When the entry ticket stops assuming you are big, the floor stops being a floor.

That does not make traditional CPQ wrong. For a multi-country programme with deep engineering rules, ERP integration and hundreds of sellers, a mature platform such as Tacton is still what we implement, and the cost is justified by the scale. The point is narrower: the same capability should not stop at the factory gates of the companies that could not absorb an enterprise project.

What should a small manufacturer do now?

Start where the money is, not where the complexity is:

1. Count your quotes. How many a week, how long each takes, and how many come back with an error. That is your business case, in your own numbers.
2. Find the decisions behind most of them. You do not need every variant on day one. Model the choices and constraints behind the bulk of your normal quote volume, and route genuine exceptions to engineering.
3. Pick one product family. A focused first scope goes live faster and is easier to maintain. Expand when real demand justifies it.
4. Measure quote time, before and after. If it does not drop from days to minutes, something is wrong with the scope, not with the idea.

If you want the longer version of how a project runs, our guide to the CPQ process walks through it step by step.

Full disclosure: cpq.se is a Tacton reseller and implementation partner for small and mid-sized manufacturers, and a CPQ advisor to enterprise organisations. I also work at sailsrep.ai. We have a stake in both sides of this argument, which is exactly why we are confident about where the floor is.

Frequently asked questions