TL;DR Self-serve and direct sales are go-to-market motions, not licensing decisions, and a motion does not get to select the value metric. When the pricing page and a rep hand the same buyer two different numbers, the argument that follows looks organizational and is architectural. Two teams are defending two regions of a pricing surface nobody ever drew. Commitment, term, and service wrap may legitimately differ between the motions. The unit being counted and the meaning of an edition may not. Two motions is not the problem. Two pricebooks is.
A prospect prices themselves on your pricing page on Tuesday. On Thursday a rep sends a quote. The two numbers do not match, and one of them has to lose. Every company running both motions has watched some version of this, and almost nobody names it correctly. The teams involved call it a people problem, a territory problem, a comp problem. It is none of those. It is a missing artifact, and that artifact sits underneath every decision in a SaaS pricing model. Naming it correctly is most of the fix.
Two Prices, One Buyer
The shape repeats. Sometimes a buyer finds both prices the same afternoon. Sometimes an existing self-serve customer learns what the negotiated edition costs and works out they are on the wrong side of it.
Internally it presents as a fight. The product team believes sales is undercutting the published price and training the market to wait for a rep. Sales believes the published price is capping the deal and handing the buyer an anchor before the first conversation. Both are describing the same artifact from two sides.
Here is the reframe: this reads as an organizational fight and it is a pricebook problem. The two teams are each defending a different region of a pricing surface nobody drew. No object above them says how the regions relate, so the argument repeats every quarter with different people in the room.
The Channel Is Not the Value Metric
Self-serve and direct sales are routes to the buyer. A route is not a licensing decision. It carries no authority to select the value metric, because the value metric is a property of what the product does for the customer, not of how the customer arrived.
Product-Led Growth (PLG) in particular gets treated as a monetization system rather than what it is, which is a channel strategy with a monetization question underneath it. Once a motion picks the unit, the company has two licensing models for one product. The self-serve edition counts one unit, the negotiated edition counts another, and every comparison the buyer makes between them stops being a choice and becomes arbitrage.
Why each motion pulls toward a different unit
The pull is real and comes from both sides. Self-serve rewards a unit the buyer can compute alone, at midnight, without a spreadsheet or a call. Direct sales rewards a unit that scales with an account as it grows, because that is what makes the second year worth more than the first. Neither is a reason to fork the metric the price attaches to, because there is one buyer and one product underneath both motions. A value metric selected to suit a channel is a metric selected for the seller’s convenience, and the buyer will eventually price both versions against each other.
What Is Allowed to Differ, and What Is Not
The slogan version of this advice is “align your motions,” which is useless because it does not say what alignment permits. The useful version is a boundary.
Commitment, term, and service wrap are fair game
A buyer who commits to more should land somewhere different from a buyer who commits to less. A buyer who signs for longer has given the vendor something the month-to-month buyer has not. A buyer who gets implementation help and a named contact has received more than the product. Those are real economic differences, and pricing them differently is ordinary discipline rather than inconsistency.
Peer-reviewed pricing theory has been clear on this for decades. The sound way to charge different buyers different amounts is to build a menu they sort themselves into, by how much they take and which capability level they need. The transfer here is negative: sorting buyers by which door they walked through is not self-selection, because the buyer can walk through the other door.
The unit and the edition are not
Two conditions must hold: the unit counted has to carry the same definition, and an edition name has to grant the same capability set no matter who sold it.
If a seat carries one definition in self-serve and a different definition under a negotiated agreement, the company has published two products under one name and will spend the next two years explaining the difference to buyers who found both. The same applies to packaging. Packaging groups capabilities into what a customer can buy, and that takes many shapes: modules, add-ons, a platform with apps sold around it, an all-in-one offer, or the familiar ladder of stacked editions. The ladder is the most common shape in B2B software and a perfectly legitimate one. Our objection is that it gets reached for by default rather than chosen. Whatever shape it takes, once the same name grants different capability sets in different motions the packaging model has forked, and the fork surfaces in every renewal where an account crosses the boundary. It is the same coherence problem that appears when a company prices multiple product lines without one architecture underneath, arriving through a different door.
Make the argument buyer-side once, because it settles the internal debate faster than any principle. In the patterns our library holds, buyers who can see both prices do not object to a difference driven by what they committed to. They object to a difference they cannot explain to their own finance team.
Where Does Your Pricing Architecture Actually Stand?
A few questions return your pricing architecture score and show which of your licensing, packaging, and pricing decisions needs attention first. Real diagnosis, not a mailing-list toll.
One Pricebook, Two Regions of the Same Surface
The pricebook is the single artifact both motions read from. It carries the list prices, the SKU structure, and the value metric each SKU is priced against. Above it sits the pricing surface, a multi-input control surface that produces a net price for any commitment a customer might make, across volume, product, Customer Group, and channel. The surface is constructed, not observed, and most companies do not have one.
Self-serve occupies one region of that surface and direct sales occupies another, and the two regions have to be continuous. Where they are not, the discontinuity is exactly where the arbitrage lives and where the internal fight starts. A team that has never drawn the surface cannot see the discontinuity, so it experiences the consequence instead: a buyer who found the seam.
Continuity, not identical prices
Continuity does not mean the two regions price identically. It means the difference between two adjacent points is explainable by something the buyer can see. Peer-reviewed work on nonlinear pricing schedules supports the same intuition from the other direction: a well-formed schedule prices adjacent commitment levels so that moving between them is a smooth economic decision rather than a cliff. Cliffs teach buyers to game the edge.
Where the two regions meet is a design decision, not a rule anyone can hand you. Several variables bear on it: the commitment level at which a buyer genuinely starts wanting a negotiation, what the negotiated motion adds beyond the product, and what it costs to serve each region. Those variables pull against each other and they are specific to the company. Anyone who supplies the number without knowing your business is selling you someone else’s answer. If you are inside this conflict now, talk to an expert and describe the seam.
One new pressure raises the stakes. LLM inference gives generative capability inside the self-serve region a real per-use serving cost, sharpening the temptation to fork the metric between editions exactly when forking costs the most.
When the Internal Fight Is the Diagnostic
The two teams are not disagreeing about strategy. Each is defending a region no artifact reconciles, and the disagreement has nowhere to land.
Research in industrial B2B markets documented this shape long before software had the problem. When two functions hold different theories of what a list price is for, the one closest to the customer settles the question unilaterally through discounting, and the corporate pricing decision is reversed at the deal. The finding that transfers is not that people disagree. It is that without a routine to resolve the conflict, the resolution happens anyway, in the deal record, where nobody has to defend it out loud.
The software analog is a rep working around the published price, or a product team setting the self-serve price without the negotiated base in view. Same mechanism, different instrument. It surfaces later as pricebook deviation, the systematic gap between what the pricebook specifies and what closed deals record. In the patterns our library holds, companies running two motions off two pricebooks show their widest spread right at the boundary.
Escalating the argument does not settle it, because the argument is not the problem. Building the artifact does.
The Questions That Settle It
Five questions, in order. The answers belong to your business rather than to a framework.
Do both motions count the same unit? If not, ask which one is the value metric and what the other one is measuring instead.
Does an edition name grant the same capability set in both motions? If it does not, the packaging model has already forked.
What does a buyer see at the boundary? Take the largest commitment self-serve accepts and the smallest a rep will quote, and ask whether a buyer holding both would call the difference explainable to a CFO.
Which artifact do both teams read from? If the answer is a pricing page for one and a quoting tool for the other, with no object above both, the friction has nowhere to resolve.
What does the negotiated motion add? If the answer is a discount and a signature, the boundary is not real and the surface has a hole in it.
One shortcut tempts almost everyone here. Revenue tooling advertises the ability to maintain a separate price list per motion: one for self-serve, one for enterprise, one for reseller. That capability is useful at the execution layer, but it is also the mechanism that produces two pricebooks. Runtime flexibility is not an architecture decision.
Two motions is not a problem to eliminate. Plenty of companies should run both. Running two motions off two pricebooks is the problem, and it does not resolve inside either team, because neither team owns the object that reconciles them. Somebody has to hold the pricebook above both motions, with the standing to make the call and the architecture to make it defensible. If your product team and your sales team are arguing about a price right now, describe the conflict to a pricing expert.
