Most CPQ projects start with the product catalog. List the products, define the bundles, set up the rules. But the pricing engine — the part that determines whether the system actually works — usually gets treated as a configuration task rather than a design decision.

That’s where things go wrong.

Pricing in a B2B company isn’t a number in a field. It’s a set of business rules, relationships, and exceptions that have evolved over years, often without being written down. Force that into a CPQ tool without first understanding the logic underneath and you end up with a system that’s technically live but commercially broken. Sales reps work around it. Finance doesn’t trust it.

Before you configure anything, understand how pricing actually works

Every company I’ve worked with has someone who knows the pricing — usually one person. They carry it in their head: which customers get which discounts, how volume breaks work, when to override standard pricing, which partner tier gets what margin.

The first step before any CPQ work isn’t configuration. It’s extraction. Sit down with that person and document what actually happens. Not what the price list says. What happens when a real deal gets quoted.

You’ll find consistent things. The published price list is a starting point, not the truth. There are customer-specific agreements nobody has written down. Volume discounts follow a structure that lives in the rep’s head. And there’s a layer of discretionary discounting governed by nothing except what feels right. All of this needs to be mapped before a single pricing rule gets created in the system.

The price waterfall

The concept that makes pricing manageable is the price waterfall — a sequence of steps, each adjusting the price based on a specific type of rule.

List price is the published baseline. Contracted price overrides it when the customer has a pre-negotiated agreement. Volume discounts apply automatically based on quantity and should live in discount schedules, not applied manually by reps. Discretionary discount is what the rep applies during negotiation, within defined guardrails. Channel margin applies as a separate layer if the deal goes through a partner. Net price is what flows to the invoice.

flowchart TD
    LP["<b>List price</b><br/>Published baseline"] --> CP["<b>Contracted price</b><br/>Pre-negotiated agreements"]
    CP --> VD["<b>Volume discounts</b><br/>Automatic, by tier"]
    VD --> DD["<b>Discretionary discount</b><br/>Rep-applied, within approval limits"]
    DD --> CM["<b>Channel margin</b><br/>Partner / distributor layer"]
    CM --> NP["<b>Net price</b><br/>What flows to the invoice"]

The power of the waterfall is transparency. Without it, nobody can answer basic questions: what’s our average discount, which reps discount the most, which product lines have the thinnest margins. With a waterfall, these are standard reports.

Four decisions to make before touching CPQ

How many price books do you need? One price book per scenario becomes unmaintainable. A better approach is a single standard price book with segment- or channel-specific pricing handled through rules and contracted pricing.

Where do volume discounts live? In discount schedules, not applied manually per quote. Decide whether discounts apply per product or across the whole quote — per product is simpler and more predictable.

What are the guardrails on manual discounting? Define the lanes, enforce them in the system, route exceptions through approvals. Without this, every deal is a negotiation between rep and manager over email, with no audit trail.

How does channel pricing work? Partner discount is a separate waterfall step, not mixed into customer-facing discounts. You need to know what the end customer paid and what the partner earned, independently.

Product catalog design affects pricing complexity

Separate what you sell from how you price it. A product should represent what the customer buys, not every pricing variation. Bundles need explicit logic: does the bundle have a single price, or components priced individually with a discount on top? Subscription products need term logic — proration, co-termination, renewal pricing. These decisions need to be made before configuration starts.

What a well-built pricing engine gives you

Quotes go out faster. Reps don’t need to check with anyone for standard pricing, volume discounts apply automatically, approval routing is built in. Margin becomes visible — management can see discount patterns and spot erosion before it becomes a trend. Channel consistency improves. Renewal pricing has a foundation in the original deal structure. And every pricing decision is tracked with an audit trail.

The pricing engine can be ready in weeks. Getting the organization to trust it takes months — reps still call the pricing person to double-check, managers still review quotes within policy. It’s not a training problem, it’s a trust problem. Plan for it.

A pricing engine isn’t a configuration task. It’s a design decision. Get the logic right first. Document what actually happens. Then build a system that reflects that reality, with structure and guardrails built in.

The technology is the easy part.