What's the relationship between a founder's go-to-market motion (PLG, sales-led, or hybrid) and the appropriate level of discount authority to delegate to sales leadership in 2027?

The founder who gave one VP the same authority in two different businesses

Consider a founder who has run this twice. The first company sold a developer tool at $19 and $49 per seat per month off a public pricing page. No quota-carrying reps, conversion happened inside the product, and the entire commercial apparatus was a billing system and a pricing page. The second company sells a compliance platform at $180,000 average contract value into regulated enterprises, with nine-month cycles, procurement, security review, and a redline negotiation on every single deal.

The founder hired the same VP of Sales into the second company that had joined the first one late. In the first company, that VP arrived and immediately built what they knew: a five-tier discount-authority matrix, a deal desk staffed at 1.5 headcount, CPQ approval routing, and a weekly deal review. Within two quarters the company had a governance apparatus reviewing transactions that were, overwhelmingly, someone entering a credit card on a pricing page. The deal desk's actual workload was roughly four negotiated deals a quarter. The approval chain added days of latency to a motion whose entire value proposition was that a developer could be paying within eleven minutes of signup. Worse, the machinery started getting used: because a deal desk existed, the company began doing custom deals, and by the end of the year a business that had never negotiated was negotiating perhaps fifteen percent of new ARR — not as a strategic decision anybody made, but because the apparatus existed and apparatus wants to be used.

In the second company, the same VP's build was exactly right and the founder resisted it for eighteen months. The founder's identity was "we're a product company," discounting "felt gross," and so approvals happened in a Slack channel — a rep would post a deal, the founder or a sales manager would type "yeah, fine," and that was the governance. Across those eighteen months, average realized discount drifted from roughly twelve percent to roughly thirty-four percent. Nobody noticed, because nobody was measuring it. When a large prospect's procurement team benchmarked three of the company's existing customers and found three materially different effective per-seat prices for near-identical scope, the inconsistency became a negotiating weapon.

Two companies, one leader, opposite errors: an over-build on the PLG motion and an under-build on the sales-led one. The instructive part is that neither error was about the quality of the governance design. The matrix the VP built in company one was a good matrix. It was simply derived from the wrong motion. The relationship at the center of this question is not that some motions need "more" governance and some need "less" in a general sense — it is that the motion determines *whether negotiation exists at all*, and discount authority is only meaningful where negotiation exists. Delegating discount authority in a business with no negotiation is delegating a power nobody will use, except that they will start using it. Withholding it in a business where every deal is negotiated does not prevent discounting — it just moves the discounting somewhere unmeasured.

The founder's first job, therefore, is not to design an authority structure. It is to answer honestly what motion the company is running *today* — not two years ago, not aspirationally — and then derive the appropriate authority level from that answer. Every remaining decision in this entry is downstream of that one.

How motion determines who can say yes to a discount

The mechanism connecting motion to authority runs through one variable: how price reaches the buyer. In PLG, price reaches the buyer through a page. In sales-led, price reaches the buyer through a person. That single difference cascades into everything else about authority.

When price reaches the buyer through a page, the buyer self-selects a tier and pays the posted number. Every customer at a given tier pays the same amount, and that uniformity is not incidental — it is the load-bearing element of the whole model. Self-serve conversion works because there is no reason to talk to anyone. The moment a prospect suspects the posted price is an opening position, they stop converting and start emailing, and the motion's cost structure collapses. So in PLG the governance question is not "how much can a rep give away" — there is no rep — it is "what prevents us from developing a negotiation surface at all." The only sanctioned price variation is the annual prepay incentive, typically fifteen to twenty percent off the monthly rate, and that is not a discount in the governance sense: it is published, universal, and structural. Delegated discount authority in a pure PLG motion should be effectively zero, and the correct owner of any exception is the founder or a pricing committee — not because founders are better negotiators, but because each exception is a decision about whether to remain PLG, and that decision does not belong to a rep.

When price reaches the buyer through a person, discounting is not an exception to the model — it is the mechanism of the model. A list price exists, but it functions as an opening position, and the realized price is the output of a negotiation spanning discount depth, term length, payment terms, scope, ramp schedule, and bundling. Here the governance job is not to prevent discounting, which is impossible and undesirable. It is to bound, channel, and audit it fast enough that deals still move. That means delegating real authority downward, because a VP who must personally approve every routine discount becomes the bottleneck on every deal in the pipeline.

The escalation ladder in sales-led exists because different levels of the organization can see different things. A rep sees the buyer, the competitive alternative, and the urgency. A first-line manager sees five to eight reps' pipelines and can tell whether a discount request reflects genuine buyer pressure or a rep who has not done the value work. A VP or RevOps lead sees segment-level pricing consistency and can spot that a proposed discount would put this customer materially below three comparable accounts. Finance and the CEO see gross margin, CAC payback, and the cost to serve. The tiers are not arbitrary rungs — each one exists because a specific piece of context first becomes available at that level.

Hybrid runs both mechanisms simultaneously on the same product, which is why it is the hardest case. The answer is not a moderate blend — a single compromise regime is too heavy for the self-serve bottom and too light for the enterprise top, mis-governing both halves of the revenue. The answer is two distinct regimes with an explicit fence between them, and the fence is defined by a trigger everyone can name: an ACV threshold, a named Enterprise tier, or a deal-shape signal like "requires a custom contract, SSO, a security review, or procurement involvement." Wherever the fence sits, discount authority attaches at the crossing and never before. A rep cannot pre-discount a self-serve account to pull it into the sales-led pipeline, because that is precisely how the published-price anchor erodes.

Real thresholds, ranges, and what the ladder looks like in practice

Specific numbers matter here, with a caveat: these are structural patterns that vary enormously by gross margin, competitive intensity, and segment. A business at ninety percent gross margin can absorb discount depth that would destroy a business at fifty-five percent. Use these as shapes to calibrate, not as constants to copy.

The margin floor comes first, and everything else is derived from it. The floor is not a discount percentage picked for symmetry — it is the discount depth at which the deal stops making economic sense given your unit economics. Work it backward: take your target gross margin, your fully loaded cost to serve, and your CAC payback target, then solve for the minimum effective price that still clears them. That number, expressed as a percentage off list, is your floor. Below it, deals require CEO or founder sign-off and are presumed value-destructive until someone argues otherwise in writing. A common failure is deriving the floor from competitive pressure instead — "our competitor discounts to forty, so forty is our floor" — which is not a floor at all, it is a competitor's pricing strategy imported into your P&L.

The authority matrix is two-dimensional, not one. Depth alone is insufficient, because fifteen percent off a $20,000 deal and fifteen percent off a $2,000,000 deal are different governance events with different dollar consequences. The matrix should cross discount depth against deal size, so a rep might hold, say, wide latitude on small deals and much narrower latitude on large ones. A workable enterprise shape looks roughly like: reps own the shallow band unilaterally; first-line managers own the next band; the VP of Sales or RevOps owns the band above that; the CRO or CEO owns everything from there down to the floor; and below the floor requires founder sign-off with written rationale. The specific percentages depend on your list-to-floor spread, but the ladder should have four or five rungs, not two and not eight. Two rungs means everything escalates to one person. Eight means nobody remembers the rules and reps route by asking a colleague.

Calibrate the ladder to deal profile, because "sales-led" is not one thing. A transactional motion — four-figure to low-five-figure ACV, cycles of days to weeks, a rep closing perhaps twenty-five or thirty deals a quarter — needs high rep-authority thresholds and wide pre-approved bands, because a rep at that volume physically cannot route each deal through a three-tier approval chain. The deal desk in a transactional motion is an exception handler that sees a small minority of deals. An enterprise motion — six and seven figures, cycles measured in quarters, a rep closing perhaps four to eight deals a year — inverts this: lower rep thresholds, narrower bands, and a deal desk engaged early and substantively on most deals. The overhead per deal is high, but each deal is large enough to justify it, and the downside of an ungoverned seven-figure discount dwarfs the cost of the review. Companies running both motions need two calibrations of the same stack, segmented by deal profile. The principle: depth of governance scales with deal size; speed of governance scales inversely with deal volume.

Discount bands are what make the routine fast. A band is a pre-approved range tied to deal characteristics — a defined range for deals under $50,000, a different one for $50,000 to $250,000, a different one above that. Inside the band, the rep moves without asking. Outside it, escalation triggers automatically in CPQ. Bands are the difference between a governance system that reps use and one they route around; without them, every discount is an approval event and the organization learns to treat approvals as theater.

The strategic-exception lane is not optional. There are deals where a deep discount is a legitimate investment rather than a discipline failure: a marquee reference logo in a segment you are entering, a competitive displacement that removes an incumbent from an account you will expand into for years, a design-partner arrangement. The lane requires senior sign-off and written rationale, but it must be *fast* — same-day or next-day. Without an explicit exception lane, every strategic deal becomes an argument, and reps learn to game the ordinary process instead of using the extraordinary one. Track exception frequency: if strategic exceptions run above roughly one in ten deals, the label has stopped meaning anything and the bands need re-cutting.

The quarterly review is what keeps the numbers honest. RevOps, finance, and sales leadership should review, every quarter: average realized discount by segment, the trend over the last four to six quarters, the distribution of deals by approval tier, exception frequency, and margin by cohort. That distribution is the most diagnostic number in the set. If eighty percent of deals sit in the rep-authority band, the bands are calibrated correctly. If eighty percent escalate, you have built an approval bureaucracy rather than a delegation structure, and you should widen rep authority rather than add reviewers.

What you give up at each level of delegation

Every point of delegated authority buys speed and costs control, and the trade is real in both directions. Founders who treat this as a purely technical exercise miss that they are choosing which failure mode they prefer.

Delegating more authority buys cycle time and rep ownership. Reps who can commit to a price in the room close faster and negotiate better, because a rep who must say "let me check with my manager" has told the buyer that the person across the table is not the decision-maker — which invites the buyer to negotiate past them. Wide authority also reduces the internal coordination tax: every escalation consumes a manager's attention, a queue slot, and usually a day or two of elapsed time. In a high-volume motion, that tax compounds into a material drag on capacity.

Delegating more authority costs pricing consistency and margin discipline. The path of least resistance in every individual negotiation is to give more, and a rep's incentive horizon is this quarter. Absent a counterweight, realized discount drifts upward monotonically — not because reps are undisciplined, but because the local optimum in every deal is a slightly deeper discount and nobody experiences the aggregate. Inconsistency is the second cost: when similar customers land on materially different prices, the variance eventually surfaces through user communities, procurement benchmarking, or a champion changing jobs, and it damages pricing power across the entire base rather than in one deal.

Compensation design is the most underused alternative to approval chains, and it works where matrices do not, because it operates continuously and at scale rather than at discrete checkpoints. The available mechanisms differ sharply by motion. In sales-led, commission on net revenue or gross margin rather than gross bookings means a rep who discounts deeply earns less, which aligns the incentive with price protection without any approval step. Discount-tiered commission rates — a higher rate on deals at or near list, a lower rate on heavily discounted deals — do the same thing more explicitly. Clawbacks on early churn stop reps from discounting a bad-fit deal across the line. The critical point: if comp pays full commission on heavily discounted deals, no authority matrix will hold the line, because the comp plan is actively fighting the governance. Design them together or you have designed neither. In PLG, comp is not a discount lever at all — product specialists and growth roles are typically paid on conversion, activation, or expansion, and a founder engineering "discount-discipline incentives" into a PLG comp plan is solving a problem the motion does not have.

Structural levers are the alternative that gets forgotten. A rep who can only move one variable — price — will move price. A rep who can trade a longer term, quarterly-to-annual prepay, a ramp schedule, a narrower initial scope, a case-study commitment, or a reference call has multiple ways to create value in a negotiation without cutting the effective rate. Pricing architecture is what makes those levers exist: a sales-led architecture should have negotiation dimensions built in deliberately, precisely so that discount depth is not the only currency in the room.

Tooling follows the same logic. A PLG motion needs a billing system and a pricing page; the pricing page *is* the quoting tool, and a heavyweight CPQ platform is tooling bought for a process the company does not run. A sales-led motion needs real CPQ with the authority matrix encoded in it, automatic approval routing, and an audit trail — because a matrix that lives in a document depends on people remembering and choosing to comply, while a matrix encoded in CPQ simply prevents the quote from being sent. Hybrid needs both stacks, integrated on customer data but separated on flow, so self-serve never gets routed through the heavy path.

Where this goes wrong and how to catch it early

Five failure patterns account for most of the damage, and each has a specific early signal.

The unacknowledged slide from PLG into hybrid. A PLG company sees large accounts self-serving in, hires someone to "talk to the big accounts," and that person — doing their job — starts negotiating. The first custom deal is a great logo and feels worth it. Then the rep asks what their discount authority is, and the company discovers it has none, because PLG never needed one. Governance gets bolted on reactively: an ad-hoc Slack approval channel, a hastily drawn matrix, a rule invented per deal. The company now has a sales motion governed by improvised rules, which is worse than either pure state. The fix is naming it. The moment leadership makes a deliberate sales hire and starts doing negotiated deals, they should say out loud: we are now hybrid, the PLG side keeps its minimal governance, and the sales-led side needs its own regime built deliberately. That naming converts an unmanaged drift into a managed transition, and every other piece of hybrid governance depends on it having happened.

Contamination across the fence, running both directions. Downward: sales-led discounts become visible outside their context — a customer mentions their rate in a user community, a procurement team benchmarks it, a champion changes jobs and expects their old price — and the market learns the published price is negotiable if you push. The published price stops being *the* price and becomes an *opening* price, which a self-serve motion cannot survive. Upward: the low PLG entry rate becomes the anchor procurement negotiates the enterprise deal down from — "your page says $15 a seat, why are you quoting $40?" The defenses are structural. Make the enterprise offering a visibly different package — security controls, SSO, admin tooling, SLAs, support tier — so the price gap is justified by contents rather than by seat count. Publish an Enterprise "contact us" tier rather than an enterprise price, so procurement has no list number to anchor on while the self-serve tiers stay transparently published. Train reps explicitly to re-anchor enterprise conversations on the enterprise package rather than letting the buyer drag the frame back to the pricing page.

Fuzzy graduation triggers. The graduation point — where a self-serve account becomes a negotiated deal — is the most governance-intensive moment in a hybrid motion, and ambiguity there blurs the fence permanently. Make it a formal event with five explicit rules: accounts cross on defined triggers, not rep discretion; crossing re-tiers the account, assigns a rep, and enters it in the sales-led pipeline; discount authority attaches *at* the crossing and never before; the deal is anchored on the enterprise package and value, with the deal desk actively helping break the prior-price anchor; and the customer experiences it as an upgrade to a more capable tier rather than a notice that they will now be charged more.

Governance that does not evolve when the motion does. Motions change on a yearly timescale; governance tends to sit frozen at whatever was last deliberately designed. PLG companies add enterprise sales teams and keep governing with "we don't really discount." Sales-led companies add self-serve tiers and route them through the same heavy deal desk, strangling the new motion's velocity. In both cases the governance is not wrong for the motion it was built for — it is wrong for the motion that now exists. The discipline: make discount governance an explicit line item in every motion-change decision. When leadership approves a sales-led motion, the same decision approves the governance for it. When leadership approves a self-serve tier, the same decision defines the fence.

Measuring nothing, which is what makes every other failure invisible. The single most diagnostic question a founder can ask is: *what is our average realized discount by segment, and is it getting worse?* If nobody can answer within a day, the governance is not under-built — it is absent, regardless of what documents exist. Run the fit diagnostic honestly. Over-build looks like a deal desk rubber-stamping standard transactions, approval workflows adding days to a motion that should close in minutes, and governance headcount disproportionate to negotiated-deal volume. Under-build looks like Slack approvals, no floor, no bands, similar customers on very different prices, and margin drift that surfaces only after it is severe. Good fit looks like a PLG company whose governance fits on one page, a sales-led company where routine discounts move without asking and only genuine exceptions escalate, and a hybrid company with two visibly distinct regimes and a fence everyone can describe.

Related questions

Should a founder personally approve discounts in an early sales-led company?

Early on, yes — but as a temporary state with an exit plan. Founder approval works at low deal volume and builds the pattern library that later becomes the matrix. Once it starts delaying deals or the founder is approving reflexively, delegate.

How does discount authority differ between new business and renewals?

Renewal and expansion discounting should sit under separate authority, usually tighter, because uplift protection is a different economic question from new-logo acquisition. Blanket new-business authority applied to renewals is how price increases quietly get negotiated away at scale.

Does a low margin floor mean sales leadership needs less authority?

The opposite. A tight list-to-floor spread means less room exists overall, so each point matters more and the bands should be narrower — but reps still need enough unilateral latitude to close routine deals without escalating, or velocity dies.

When does a PLG company actually need a deal desk?

When negotiated deals become a recurring share of new ARR rather than occasional exceptions, and when reps start asking what their authority is. Before that, a deal desk is machinery looking for work — and it will find work, by creating negotiations.

FAQ

Can a pure PLG company delegate any discount authority at all?

It can, but it usually should not, and the reason is structural rather than philosophical. Every one-off custom price is a small crack in the published-price model, and the decision to make an exception is really a decision about whether to remain PLG. That decision belongs to the founder or a pricing committee, not to an individual closing a specific deal. The one sanctioned price variation — the annual prepay incentive — is published, universal, and requires no delegation because it is not negotiated.

What is the right number of tiers in a discount-authority matrix?

Four or five rungs is the practical range for a sales-led motion: rep, first-line manager, VP or RevOps, CRO or CEO, and founder sign-off below the floor. Fewer than four means everything escalates to one person and that person becomes the bottleneck. More than five means nobody remembers the rules, and reps start routing approvals by asking whoever answers fastest rather than by following the structure.

How do you keep a deal desk from becoming the sales prevention department?

Calibrate the bands so the routine discount never reaches the desk. If most deals escalate, the problem is band calibration rather than rep behavior. Give the desk a service mandate as well as a control mandate — deal structuring, economics modeling, and negotiation advice — so reps come to it for help rather than only for permission. And keep the strategic-exception lane genuinely fast, because a slow exception lane teaches reps to game the ordinary process.

Should the discount matrix be the same for transactional and enterprise segments?

No. A rep closing thirty deals a quarter and a rep closing six a year need opposite calibrations. The transactional segment needs high rep thresholds and wide bands so velocity survives; the enterprise segment needs lower thresholds, narrower bands, and early deal-desk involvement because each deal is large enough to justify the overhead. Companies running both need two calibrations of the same stack, explicitly segmented.

What is the first thing to build if a sales-led company has no governance today?

Measurement, then the floor, then the matrix. Before designing anything, produce the number: average realized discount by segment and its trend over the last four quarters. That number tells you where the floor and bands should sit and creates the urgency to build them. Building a matrix before you know your actual discount distribution means picking thresholds from intuition, and they will be wrong in ways nobody can detect.

Can compensation design substitute for an approval matrix entirely?

Not entirely, but it does more work than most founders expect. Comp operates continuously on every deal, while approvals operate at discrete checkpoints, so margin-based commission and discount-tiered rates shape behavior at a scale no matrix reaches. The matrix remains necessary for the deep end — deals approaching the floor, strategic exceptions, and anything requiring cross-functional judgment. The failure to avoid is designing them separately, because a comp plan paying full commission on deep discounts will quietly defeat any matrix.

Sources

• Harvard Business Review — Pricing and revenue management coverage
• McKinsey & Company — Growth, Marketing & Sales insights
• Bain & Company — Pricing consulting insights
• OpenView Partners — Product-led growth resources
• a16z — Enterprise and SaaS go-to-market writing
• SaaStr — SaaS sales and pricing archives
• Salesforce — CPQ (Revenue Cloud) product documentation
• Gartner — Sales practice research

Related on PULSE

• How to design a two-dimensional discount-authority matrix by deal size and depth
• When a founder should hire a dedicated deal desk versus a fractional RevOps owner
• Structuring sales compensation so the comp plan reinforces discount discipline
• Setting a margin floor from unit economics instead of competitive pressure
• Fencing PLG published pricing from enterprise negotiated pricing in a hybrid motion
• Graduation triggers: moving a self-serve account into the sales-led pipeline