How do I price an add-on SKU when my base product is underpriced?

Direct Answer

Do not raise the base price mid-cycle and bolt an add-on on top — that compounds the original underpricing error and detonates retention. Instead, build a parallel add-on SKU that carries its own anchor price (40-50% of base ACV), then bundle the two so the bundle wins through Asymmetric Dominance rather than through a discount.

If your base is $500/mo, list the add-on at $250/mo standalone and price the bundle at $650 — anchored against a $750 a la carte sum. Existing contracts never reopen, blended ARPU climbs 10-16% within twelve months, net revenue retention gains roughly twelve points, and you buy 12-18 months of clean runway to plan a properly grandfathered base-price relaunch.

The single most important precondition: run a Van Westendorp willingness-to-pay study and a gross-retention-by-ARR-band cut before you ship anything. If your base is underpriced because your ICP is wrong rather than because your pricing is wrong, a parallel SKU does not fix it — it accelerates the churn you were trying to escape.

(For the eventual base-price relaunch itself, see (q80); for how many tiers to run, see (q77).)

1. The Core Decision: Parallel SKU Versus Base Hike

When a base product is underpriced, founders instinctively reach for the wrong lever. They want to fix the number that is wrong — the base price — by changing the base price. It feels honest and direct.

It is also the most expensive mistake available to you, because the base price is the one variable that is contractually load-bearing for every customer you already have. Touching it reopens negotiations you have already won, and it does so across your entire installed base at the same moment.

1.1 Why the Instinct to Raise the Base Is a Trap

The pull toward a direct base increase is strong precisely because it feels like the most direct route to the problem. It is also the route with the worst risk-adjusted return. The failure modes are specific and compounding.

• The contract is the asset, not the price. Every signed customer represents a closed negotiation. The moment you raise the base mid-cycle, you reopen that negotiation across your entire installed base simultaneously, and you do it from a position of weakness — the customer did not ask to renegotiate, you did. You have converted a settled relationship into an open question.
• Renewal procurement gets activated early. A mid-cycle increase tells the buyer's procurement organization that price is now a live variable. That trains them to push back harder at every future renewal, even after the immediate increase is absorbed. You do not just pay the churn cost once; you pay an elevated negotiation tax on every renewal thereafter.
• You signal a scarcity of confidence. A vendor who raises prices on existing customers mid-contract is read — correctly — as a vendor who mispriced and is now scrambling to correct it on the customer's dime. That perception is sticky, and it bleeds into win rates on new logos through reference calls and public review sites.
• The math is asymmetric against you. As the worked example in section 4 demonstrates, the churn cost of a base hike on a 1,000-account base can swing net ARR by more than $1.3M against you in year one alone, before any second-order reputational cost is counted.

The Asymmetric Dominance path sidesteps every one of these failure modes because it never touches the existing contract. The customer's base price stays exactly where they signed it. You are not taking anything away or asking for more on the same terms — you are offering a genuinely new thing at a new price, and letting the customer opt in on their own timeline.

1.2 What "Parallel SKU" Actually Means

A parallel SKU is not a feature gate, and it is not an upsell tier carved out of the base. It is a separately priced, separately positioned product line item that carries its own anchor. The distinction is not pedantic — it is the entire difference between a bundle that lifts ARPU and a bundle that triggers a churn wave.

• A feature gate says: "the thing you wanted is now behind a paywall." Buyers experience this as a takeaway, and loss aversion makes them resent it far out of proportion to the dollar amount. A capability they already had, now priced, feels like theft.
• A parallel SKU says: "here is an additional capability with its own value and its own price." Buyers experience this as a choice, and choice does not trigger loss aversion. The same code shipped two different ways produces opposite NPS outcomes.
• The parallel SKU must clear a value bar. Patrick Campbell, who founded ProfitWell before its acquisition by Paddle, has repeatedly made the point in his pricing teardown work that an add-on priced below 25% of base ACV reads as a feature rather than a product. Buyers do not assign it independent value, so it cannot anchor a bundle. The add-on has to be substantial enough that a buyer would plausibly purchase it standalone.
• It needs its own buyer narrative. If you cannot write a one-paragraph standalone value proposition for the add-on without referencing the base, it is not a parallel SKU — it is a base feature wearing a price tag, and the bundle psychology in section 2 will not fire.

The test is concrete. Sit down and write the standalone landing page for the add-on. If that page cannot stand on its own — if every sentence has to lean on "and it works with your existing product" — you have not built a parallel SKU. You have built a feature, and you should not be pricing it separately.

1.3 The Twelve-to-Eighteen-Month Window

The parallel SKU is not a permanent escape from an underpriced base. It is a bridge. It buys you two things at once: time and data.

• It buys time. A base relaunch done well takes planning — grandfathering windows, customer communication, sales enablement, billing-system changes. The add-on lets you lift ARPU now while that planning happens, so you are not under panic-level pressure to reprice.
• It buys data. During the window the add-on is live, you collect real attach behavior, real willingness-to-pay signal from how buyers respond to the standalone anchor, and real expansion-ARR data segmented by customer type. That data is precisely what you need to design a defensible base relaunch with proper grandfathering.
• It de-risks the eventual reset. A base increase planned from evidence — actual WTP curves, actual attach data, actual segment retention — is a fundamentally different and safer act than a base increase planned from a spreadsheet guess.

The mistake is treating the add-on as the destination. It is the vehicle that gets you to a clean base reset with evidence instead of guesswork. The operators who fail with this playbook are almost always the ones who forgot it was a bridge and tried to live on it permanently.

2. The Theory: Asymmetric Dominance and the Decoy Effect

The reason a parallel SKU plus a bundle outperforms a raw price increase is not folk wisdom. It rests on one of the most replicated findings in behavioral pricing research, and understanding the mechanism precisely is what separates a bundle that lifts ARPU from a bundle that just gives margin away.

2.1 The Huber, Payne, and Puto Experiment

In 1982, Joel Huber, John Payne, and Christopher Puto published "Adding Asymmetrically Dominated Alternatives: Violations of Regularity and the Similarity Hypothesis" in the *Journal of Consumer Research*. The experiment is now a fixture of behavioral economics curricula, and its result is one of the load-bearing facts of pricing strategy.

Subjects were asked to choose between two options that traded off on two dimensions. In the canonical beer example, they chose between Beer A at $1.80 with a quality rating of 50, and Beer B at $2.60 with a quality rating of 70. A is cheaper; B is better.

Roughly a third of subjects chose B. This is the baseline: a genuine trade-off, and most people picked the cheaper option.

Then the researchers introduced a third option: Beer C at $1.80 with a quality rating of 40. C is dominated by B on every dimension — it costs the same as A but is lower quality — and crucially it is *not* dominated by A. C exists only to make B look obviously superior. With C on the menu, Beer B's share jumped from 33% to roughly 60%.

The decoy did not get chosen by anyone in meaningful numbers. It made the target irresistible. This violated what economists call "regularity" — the assumption that adding an option to a choice set cannot increase the share of an existing option.

The decoy effect proved that assumption false.

2.2 Why "Asymmetric" Is the Operative Word

The word "asymmetric" is doing precise technical work, and getting it wrong is the most common way operators build a decoy that does not function.

• The decoy must be dominated on every axis by the target. This is what "asymmetric" means: C is worse than B in all respects, but C is not uniformly worse than A. The dominance relationship is one-sided. If your decoy is dominated by *both* of the other options, it is just a bad option and it shifts nothing.
• The decoy makes comparison easy. Human buyers genuinely struggle to compare options that trade off — cheaper-but-worse versus pricier-but-better is real cognitive work, and the brain looks for a shortcut. A dominated decoy hands the buyer that shortcut: "B is strictly better than C, and C is in the running, so B is the smart pick." The decoy converts a hard comparison into an easy one.
• The decoy is never expected to sell. If your decoy tier sells well, it is not a decoy — it is a real option that you have mispriced. A properly constructed decoy captures single-digit share. Its job is share transfer, not revenue, and you should not be alarmed when it produces almost no direct sales.
• The decoy must be visible at the point of decision. A decoy that the buyer never sees cannot influence the buyer. This sounds obvious but it is violated constantly when marketing teams "simplify" a pricing page by removing the option that was functioning as the decoy.

2.3 The Ariely Replication

Dan Ariely revisited the effect in *Predictably Irrational* (2008) using a real-world artifact: a subscription pricing page from *The Economist*. The page offered a web-only subscription at $59, a print-only subscription at $125, and a print-and-web subscription also at $125. The print-only option is the decoy — nobody rationally buys print-only when print-and-web costs exactly the same dollar amount.

Ariely ran the experiment on 100 MIT students. With all three options present, 84% chose the $125 print-and-web bundle and 16% chose the $59 web-only option; nobody chose print-only. When he removed the decoy and offered only web-at-$59 and print-and-web-at-$125, the result inverted: 68% switched to the cheap web-only option and only 32% chose the bundle.

The decoy was worth a swing of more than 50 points toward the premium bundle. Removing it cost *The Economist*, in Ariely's modeling, a 43% drop in revenue per visitor. The decoy was not a curiosity.

It was nearly half the revenue.

2.4 Translating the Theory to Your Tier Ladder

The application to an underpriced base is direct. Your base SKU and your bundle are A and B. The bundle wins not because it is cheaper but because the *structure of the menu* makes it the obvious choice.

The a la carte sum of $750 is your structural decoy. A rational buyer who wants both products will never assemble them separately at $750 when the bundle delivers the identical capability for $650. The bundle dominates the a la carte path on price.

It simultaneously dominates the base-only tier on capability breadth, at a marginal cost the buyer perceives as small. That two-sided dominance is what drives the middle tier to win the majority of selections — and it does so without you having raised the base price by a single dollar.

2.5 The Anchoring Mechanism Underneath It

Beneath the decoy effect sits an even more fundamental mechanism: anchoring. Daniel Kahneman and Amos Tversky's work on the anchoring-and-adjustment heuristic established that human numerical judgments are pulled toward whatever number is presented first, even when that number is logically irrelevant.

• The standalone add-on price anchors the bundle. When a buyer sees the add-on listed at $250 standalone, that number becomes the reference point. The bundle is then judged relative to it, and the bundle looks generous because the buyer is anchored on paying $750 total.
• The a la carte sum anchors the discount. The buyer who internalizes a $750 a la carte cost perceives the $650 bundle as a $100 saving. Remove the a la carte option and there is no anchor for the saving to register against.
• Anchors must be credible to function. Kahneman's work also shows that absurd anchors are discounted. An add-on priced at an implausible number — far above any reasonable value — fails to anchor because the buyer rejects it outright. This is another reason the 40-50% band in section 3 matters: it keeps the anchor credible.

Richard Thaler's concept of transaction utility, from *Misbehaving*, adds the final piece. Buyers derive satisfaction not only from the thing they buy but from the *deal itself* — the gap between what they paid and what they perceive as the reference price. The 87% bundle rule in section 3.2 is engineered to manufacture exactly that transaction utility.

3. Mechanics: Anchor, Decoy, and Fence

Theory tells you why the structure works. Execution lives in three numbers: the standalone anchor of the add-on, the bundle discount that creates the decoy, and the commitment fence that locks the upgrade in.

3.1 The Anchor — Price the Add-On at 40-50% of Base ACV

The standalone price of the add-on is the single most important number in the entire architecture, because it is the reference point against which every buyer evaluates the bundle. Set it too low and the add-on reads as a trivial feature with no independent value, which collapses the decoy.

Set it too high and the bundle discount required to make the math work blows out your margin.

• The 40-50% band is the operating range. Carta's *State of Private SaaS* pricing data, drawn from the cap-table records of thousands of venture-backed companies, has repeatedly shown that the median add-on ACV sits near 38% of base ACV. Pushing your anchor into the 40-50% band keeps you in defensible territory while leaving room for the bundle discount.
• Below 25% is a feature flag. An add-on priced under a quarter of base ACV does not register as a product in the buyer's mind. They will not pay for it standalone, which means it cannot anchor anything, and the entire decoy structure collapses.
• Above 60% confuses the hierarchy. An add-on priced near or above base ACV makes buyers ask which product is actually the core offering. That confusion is not free — it lengthens the sales cycle as the buyer tries to reconstruct what they are actually buying.
• The anchor must survive a Van Westendorp test. The number you pick in this section is a hypothesis, nothing more. Section 6 is how you validate it against real buyer willingness to pay before a single dollar of pipeline is built on it.

3.2 The Decoy — The 87% Bundle Rule

The bundle price is set as a fixed blend of the a la carte sum. The working rule that practitioners converge on is 0.87 times the sum of standalone prices, rounded to a clean number.

The 13% discount is deliberately calibrated. It is large enough that the buyer perceives a genuine, motivating saving — the difference between $750 and $650 is concrete and easy to feel. It is small enough that you preserve the unit economics of the add-on, which in pure-software cases carries near-zero marginal cost.

The discount is funded almost entirely by the add-on's margin, not the base's.

This is the crux of the whole strategy: you are not discounting the base, you are discounting a blended package whose discount is absorbed by a high-margin add-on. That is structurally different from cutting the base price, even though the buyer experiences both as "a deal." The base ARPU is fully preserved.

The headline price the buyer sees is lower than the a la carte alternative, but the unit you most need to protect — the base — never moved.

• Round to a clean number. $652.50 becomes $650; $1,000.50 becomes $900. Clean numbers are easier to remember, easier to quote, and they avoid the impression of a price reverse-engineered from a formula.
• Round down, not up. When you round, round in the buyer's favor. The few dollars you forgo are trivial against the goodwill and the cleaner perceived saving.
• Hold the discount at 13%, not deeper. A deeper discount does not meaningfully increase attach — the decoy is doing the persuasion work — but it does erode add-on margin. Resist the instinct to "sweeten" the bundle.
• Keep the a la carte path on the page. The a la carte sum is the decoy. If it disappears from the pricing page, the bundle loses its reference point and the Ariely effect reverses.

3.3 The Fence — Annual Commitment as the Upgrade Gate

The fence is the mechanism that converts a one-time upgrade decision into a durable revenue commitment. The cleanest fence is the annual-versus-monthly delta.

• Price the monthly bundle higher than the annual bundle. A month-to-month bundle at $700 against an annual bundle at $650 creates a 7% incentive to commit annually. The buyer who wants the bundle is nudged toward the annual term without being forced.
• The 7-10% delta is the sweet spot. Industry survey data consistently shows annual take rates climbing sharply when the monthly-to-annual delta sits in the 7-10% range, and collapsing toward the high-teens percent take rate when the delta falls below 5%. Below a 5% gap, the buyer sees no reason to lock in.
• The fence does double duty. Annual commitment on the bundle both lifts cash collection and structurally reduces churn exposure on the upgraded cohort, because an annual contract cannot churn mid-term. You have converted a monthly churn risk into an annual renewal event.
• Co-term the add-on with the base. The add-on contract end date must align with the base renewal date. Mismatched terms create renewal-cycle chaos, double the legal review burden, and produce confusing partial-period invoices that generate support tickets and billing disputes.

3.4 The Comp Plan Is Part of the Mechanics

A pricing architecture that the sales team is not paid to sell is not an architecture — it is a hope. The comp plan is the fourth mechanical element, and it is the one most often forgotten.

• Add an explicit attach SPIF. Reps must be paid specifically for add-on attach, not just for total contract value. A 6%-or-greater accelerator on add-on ACV is the working benchmark. Without it, the rep optimizes for the easiest path to a number.
• Pay for mix, not just size. If comp rewards only total contract value, the rep will discount the base to land the bundle, the headline closes, and your blended ARPU does not move. The SPIF must reward the *composition* of the deal.
• Make the attach SPIF visible on the dashboard. Reps optimize what they can see. An attach metric that is buried in a quarterly report does not change behavior; one on the daily leaderboard does.

For the broader question of how compensation scales across a multi-SKU portfolio, see (q11); for the OTE structure of the reps selling the expanded portfolio, see (q14); for governing discount autonomy so reps cannot circumvent the architecture through informal bundling, see (q9516).

4. The Worked Numbers: ARPU and Three-Year LTV

Abstract structure does not move a board. The case for the parallel SKU is a number, and the number is large. The model below uses a 1,000-customer base on a $500/mo SKU — $6.0M of installed ARR — and compares the parallel-SKU path against the base-hike path head to head.

4.1 Year-One ARPU Under the Parallel-SKU Path

Launch the add-on at $250 standalone, bundle at $650, so the bundle delta a customer pays above base is $150/mo.

The median case — 34% attach — is not aggressive. Cross-vendor benchmarking from Pavilion's compensation and pricing work has put median add-on attach near a third of the eligible base within twelve months of launch. At that rate you add $612K of ARR and lift blended ARPU 10.2% without touching a single existing contract. The top-quartile case adds nearly a million dollars of ARR on the same logo base.

4.2 The Base-Hike Path, Modeled Honestly

To produce the same 10.2% ARPU lift through a base increase, you would push the $500 base to $551. Mid-cycle increases above 12% on mid-market accounts churn in the 18-22% range within six months against a 6-9% baseline. Model it at the midpoint, 20%.

The swing is $1.32M in a single year on a mid-sized base — and the base-hike path also loses 200 customer relationships, 200 sets of reference value, and 200 potential future expansion accounts. The churned logos do not just stop paying; they remove themselves from every future expansion model and they show up as detractors in the diligence calls your next 200 prospects will make.

4.3 The Three-Year LTV Bridge

Attach is not static. Pavilion's 13-24 month cohort data shows attach climbing as the add-on matures in market and as sales teams build muscle around the motion.

Over three years the parallel-SKU path compounds to roughly $2.4M of cumulative incremental ARR, while the base-hike path compounds *negatively* to between -$1.8M and -$2.1M as churned logos fail to renew and the shrunken base produces less expansion. The cumulative gap exceeds $4M.

That is the number that belongs in the board deck. For the framework on auditing whether the underlying pricing model still fits the business over time, see (q9524).

4.4 Why the NRR Line Moves Twelve Points

Net revenue retention is the metric public-market investors weight most heavily for SaaS, and the parallel SKU moves it structurally rather than cosmetically.

• Multi-tier ladders build expansion paths. Bessemer Venture Partners' *State of the Cloud* analysis has consistently shown that public SaaS companies operating three or more paid tiers post median NRR around 114%, versus roughly 102% for single-tier peers — a twelve-point gap.
• The mechanism is structural. A multi-tier ladder gives every account a built-in expansion path, so the same renewal event can produce a net upsell instead of a flat or contracting renewal. The renewal stops being a defensive moment and becomes an offensive one.
• NRR drives valuation multiple. A twelve-point NRR difference is the line between a Rule-of-40 company and a sub-scale one, and it flows directly into the revenue multiple a company commands. The parallel SKU is, in that sense, a valuation lever and not just an ARPU lever.

For the mechanics of computing gross retention versus net retention cleanly across a base with multi-year contracts and annual escalators, see (q9518).

4.5 The Margin Story

The worked numbers above are ARR numbers. The margin story is, if anything, more favorable, and it is worth stating explicitly because it is what protects the Rule-of-40 calculation.

• Pure-software add-ons carry near-zero marginal cost. Once the add-on is built, each incremental attach is almost pure gross margin. The $612K of year-one incremental ARR in the median case flows to gross profit at close to 100%.
• The base-hike path adds revenue at the cost of fixed-cost absorption. When 200 logos churn, the fixed costs of serving the base — support, infrastructure, account management — are now spread across fewer accounts, so unit economics on the survivors quietly worsen even as their price rises.
• Net: the parallel SKU improves both the numerator and the denominator of the Rule-of-40. It adds high-margin revenue and it protects the customer base over which fixed costs are amortized.

5. The Operator Evidence: Who Has Run This Play

The parallel-SKU-plus-decoy architecture is not theoretical. It is the dominant pricing pattern across public SaaS, and the public filings show it working at scale.

5.1 HubSpot — The Hub Architecture

HubSpot (NYSE: HUBS) is the cleanest large-cap example of parallel SKUs done deliberately. Rather than one monolithic CRM with feature tiers, HubSpot sells Marketing Hub, Sales Hub, Service Hub, Content Hub, and Operations Hub as parallel products, each with its own standalone pricing and its own Starter, Professional, and Enterprise ladder.

The bundle — the "Customer Platform" or CRM Suite — is priced well below the sum of the individual Hubs, which makes the suite the structurally dominant choice for any customer using two or more Hubs. HubSpot's investor materials have repeatedly attributed multi-Hub adoption as a principal driver of expansion revenue, and the company has reported average subscription revenue per customer rising as multi-Hub attach climbs.

The architecture is Anchor-Decoy-Fence at scale: each Hub anchors itself, the a la carte sum decoys the suite, and annual contracts fence it.

5.2 Atlassian — Add-Ons and the Marketplace

Atlassian (NASDAQ: TEAM) runs the add-on logic in two directions at once. First, its own products — Jira, Confluence, and Jira Service Management — function as parallel SKUs that bundle into cross-product pricing. Second, the Atlassian Marketplace turns thousands of third-party apps into add-ons priced as a function of the host product's seat count, which institutionalizes the "add-on at a fraction of base" pattern across an entire ecosystem.

Atlassian's cloud migration cohorts have shown expansion driven heavily by app and edition attach rather than by base seat-price increases — exactly the parallel-SKU logic, scaled into a platform.

5.3 Salesforce — Clouds as Parallel SKUs

Salesforce (NYSE: CRM) is the canonical case and arguably the originator of the pattern at enterprise scale. Sales Cloud, Service Cloud, Marketing Cloud, Commerce Cloud, and the rest are parallel SKUs, each independently priced and tiered. Salesforce's expansion engine has historically run on selling additional Clouds into existing accounts rather than on raising the price of Sales Cloud for installed customers — precisely the parallel-SKU logic.

The company's reported net revenue retention has sat in territory consistent with a mature multi-product ladder, and its investor narrative has long centered on "Clouds per customer" as the expansion metric.

5.4 Zoom and the Add-On Reset

Zoom Video Communications (NASDAQ: ZM) demonstrates both the upside and the limit of the strategy. Zoom layered parallel SKUs — Zoom Phone, Zoom Rooms, Zoom Contact Center, and the Zoom AI Companion — on top of a meetings base, and add-on attach became a central part of its post-pandemic expansion narrative once core meetings growth flattened.

The cautionary read, developed in section 7, is that parallel SKUs lift ARPU on a healthy base but do not manufacture demand where the core product's growth has structurally stalled. The add-on is a multiplier on underlying health, not a substitute for it.

5.5 Datadog — The Extreme Case

Datadog (NASDAQ: DDOG) has taken the parallel-SKU architecture further than almost anyone, selling more than twenty distinct observability products as parallel modules — infrastructure monitoring, APM, log management, security, and the rest. Datadog's investor disclosures have routinely highlighted the percentage of customers using four, six, and eight or more products as a headline expansion metric — a direct readout of add-on attach.

The lesson for an underpriced-base operator is that there is enormous room to run: the constraint is rarely "we have too many SKUs," it is "we have not built the first parallel SKU yet."

5.6 Snowflake and Consumption Models

Snowflake (NYSE: SNOW) shows that the parallel-SKU logic survives translation into a consumption-based model. Snowflake's add-on capabilities — features layered onto the core data-warehouse consumption — expand the surface area over which a customer consumes credits. The "add-on" in a consumption model is not a fixed monthly line item but an expansion of consumable workload, yet the strategic logic is identical: grow ARPU by adding capability the customer opts into, not by raising the price of what they already buy.

The mechanics of the anchor and the fence differ; the principle does not.

6. Validate Before You Ship: The Van Westendorp WTP Test

Every number in sections 3 and 4 is a hypothesis until it is tested against real buyer willingness to pay. The Van Westendorp Price Sensitivity Meter, developed by Dutch economist Peter van Westendorp in 1976 and presented at the ESOMAR Congress, is the standard instrument, and you run it *before* you build a single quote.

6.1 The Four Questions

Survey a panel of 100 or more buyers — a deliberate mix of current customers and qualified prospects — with exactly four questions about the add-on.

• Too expensive (TE): At what price does this add-on become so expensive you would not consider buying it at all?
• Expensive but considered (E): At what price does it start to feel expensive, but you would still consider buying it?
• Bargain (B): At what price does it feel like a bargain — a genuinely good buy?
• Too cheap (TC): At what price does it feel so cheap that you would question its quality or doubt that it works?

The four questions are deceptively simple, and that is the point: they extract a price *range* from the buyer's intuition without ever asking the unanswerable direct question "what would you pay?"

6.2 Reading the Curves

Plot cumulative response curves for all four questions and read the intersections.

• The Range of Acceptable Prices runs from PMC to PME. If your $250 hypothesis falls inside that range, you have a defensible launch price. If it falls outside, you are guessing, and you re-anchor before building anything.
• The OPP is your starting list price. It is the price at which the fewest buyers reject the add-on outright — the point of least resistance.
• The IPP tells you about brand power. When the IPP sits above the OPP, it usually signals a strong brand that can sustain a premium; when it sits below, it signals price sensitivity that argues for the lower end of the range.
• Pair Van Westendorp with a Gabor-Granger test. Van Westendorp gives you the acceptable range; the Gabor-Granger method, developed by economists André Gabor and Clive Granger, gives you the revenue-maximizing point *within* that range by testing purchase intent at discrete price points. Run both.

6.3 The Retention Cut You Run in Parallel

While the WTP survey is in the field, pull gross retention by ARR band. This is the diagnostic that tells you whether you have a pricing problem or an ICP problem — and it is the most important thing in this entire section.

• Cut gross retention into ARR bands. For example, sub-$10K, $10-50K, and $50K-plus annual contract value. Each band is a different economic animal.
• If the sub-$10K band churns above 15% annually, stop. That cohort cannot afford the base, and a $250 add-on will not change that — it will accelerate their exit. You have a segmentation problem, not a pricing problem, and section 7.1 is your situation.
• If retention is healthy across all bands, proceed. Healthy retention means the underpricing is genuinely a pricing problem, and the parallel SKU is the right tool for it.
• Document the baseline. The day-60 and day-90 kill criteria in section 8 compare beta-cohort retention against this control, so you need the number formally on record before launch, not reconstructed afterward.

7. Counter-Case: When the Parallel SKU Is the Wrong Move

Intellectual honesty requires a hard look at the conditions under which this entire playbook fails. The parallel SKU is a powerful tool on a healthy base. On an unhealthy base it is an accelerant, and the failure modes are specific and predictable.

7.1 When Your Base Is Underpriced Because Your ICP Is Wrong

This is the single most dangerous failure mode, and it is the reason section 6.3 exists. If your base appears underpriced but the real issue is that you are selling to a segment that fundamentally cannot pay more, then the price is not the problem — the customer is. Adding a $250 SKU to customers who already strain to afford $500 does not lift ARPU; it raises the friction of every renewal and pulls forward the churn.

The gross-retention-by-band cut is the test. If the cheap cohort is already leaking, you fix segmentation first and revisit pricing two quarters later. SMB cohort churn pressure is a well-documented condition that SaaS companies routinely disclose as a risk factor in their SEC filings — it is a common situation, not an edge case.

7.2 When the Add-On Is Just a Feature You Removed From the Base

If you build the "add-on" by carving an existing capability out of the base and re-pricing it, customers will notice within a quarter — usually faster. Sentiment on G2 and TrustRadius swings negative, the word "nickel-and-diming" appears in reviews, and the reputational damage extends CAC payback by months as prospects encounter the complaints during their diligence.

The add-on must be genuinely net-new value. If it is not, you are running an unbundling exercise dressed as a pricing strategy, and NPS will tell the truth to anyone who looks. The standalone-landing-page test from section 1.2 is your guardrail here.

7.3 When Sales Comp Does Not Change

If you ship a bundle but leave the comp plan untouched, reps will optimize for total contract value, which means they will discount the base to land the bundle. The bundle closes, the headline number looks acceptable, and your blended ARPU does not move because the discount ate the gain.

The fix is the explicit attach SPIF from section 3.4 — an accelerator on add-on ACV specifically — so the rep is paid for *mix*, not just *size*. Without it, the architecture leaks at the exact point where it should be capturing value. For handling the upstream "send me pricing on call one" conversation that often precedes a discount spiral, see (q58).

7.4 When You Stack Too Many Tiers

The decoy effect does not scale linearly with tier count. Beyond three real options, choice overload sets in — the cognitive ease that made the bundle obvious is replaced by analysis paralysis, and close rates fall. Four-tier ladders have been observed closing meaningfully worse than three-tier ladders.

The architecture is three tiers plus a structural a la carte decoy. Resist the temptation to keep adding tiers; each one past three is subtracting from your close rate, not adding to it. For the dedicated treatment of how many tiers to run, see (q77).

7.5 When Raising the Base Genuinely Is Correct

There are real conditions under which you should skip the parallel SKU and raise the base directly.

If competitor base prices sit at $1,200-plus and yours is $500, the add-on does not close that gap — it papers over it. In that case you plan the grandfathered base relaunch *before* shipping the add-on, not after, so you do not subject customers to two painful repricing events 18 months apart. For executing that base increase cleanly, see (q80).

7.6 The Real Failure Pattern

A documented failure pattern from a Series-B company illustrates how the counter-case plays out in practice. The add-on launched, and day-90 attach came in at 12% — well under the 25% threshold the team had set. At that point the kill criteria should have triggered.

Instead, sunk cost — a rebuilt comp plan and six weeks of engineering effort — pulled the team forward. They told themselves attach would climb. By month nine, attach was still stuck at 14%, and the team had burned nine months of focus on the wrong problem.

The actual problem, all along, was that the base SKU sat below the SMB willingness-to-pay floor. They eventually ran the base relaunch anyway, fourteen months late, and the delay cost them a CRO. The lesson is not "the playbook does not work." It is "the kill criteria in section 8 exist precisely so that sunk cost does not get a vote."

8. Kill Criteria: Define Failure Before You Launch

The discipline that separates a controlled experiment from a slow-motion disaster is a kill-criteria table written and signed off *before* launch. Pre-commitment is the only reliable defense against sunk-cost reasoning, because once money and engineering hours are spent, the team's judgment is compromised by definition.

8.1 The Kill-Criteria Table

8.2 How to Use the Table

• The CRO signs it before launch. Sign-off must happen while judgment is still uncompromised by sunk cost — that is the entire point of pre-commitment. A kill-criteria table written after the money is spent is theater.
• Each row maps to a specific corrective action. Not a vague "investigate," but a pre-decided action. If attach misses, you pause. If retention drags, you pull. The decision is made in advance, when it can be made clearly.
• The thresholds are gates, not targets. Missing a threshold is not a prompt for a debate or a round of rationalization — it triggers the pre-agreed action. The whole value of the table evaporates if it becomes negotiable.
• Day 90 is the decision point. Roll to 100% only if cohort attach clears 25% and gross retention holds within 200 basis points of the control. Anything else means pull, regroup, and run the grandfathered base relaunch instead.

8.3 Why Each Threshold Is Set Where It Is

• Attach at 25% by day 90 is the floor below which the add-on is not pulling its weight against the median 34% twelve-month trajectory; a beta running at 25% at day 90 is plausibly on track for that median, while one below it is not.
• Retention within 200 bps of control isolates the add-on's effect from normal cohort variance; a drag wider than 200 bps is signal, not noise, and it means the add-on is actively harming the relationship.
• Sales cycle within 10% catches the choice-overload failure early — if the new tier structure is lengthening cycles, the menu is too complex and needs simplification before rollout.
• Discount depth at or below 15% catches the comp-leak failure — depth beyond 15% means reps are discounting the base to land the bundle, and the SPIF needs fixing.

For the broader cadence of revisiting whether the pricing model itself still fits the business, see (q9524).

9. The Sequenced Action Plan

The playbook compresses into roughly six weeks of build, followed by a beta period with hard decision gates.

9.1 Days 1-7: Diagnose

• Pull the last six quarters of expansion ARR. If under 15% of net-new ARR comes from existing accounts, stop — you have an attach problem, not a pricing problem, and a new SKU will not fix a motion gap.
• Pull gross retention by ARR band. Confirm the sub-$10K cohort churns under 15% annually. If it does not, this is an ICP repair project, not a pricing project, and section 7.1 applies.
• Document the retention baseline. Record the control number that the kill criteria in section 8 will measure against.

9.2 Weeks 2-3: Test Willingness to Pay

• Field the Van Westendorp survey to 100-plus current and prospective buyers.
• Compute the OPP, the IPP, and the Range of Acceptable Prices from the cumulative curves.
• Confirm the $250 hypothesis sits inside the RAP. If it does not, re-anchor before proceeding — do not build pipeline on an untested price.

9.3 Week 4: Price and Plan

• Set the add-on at 40-50% of base ACV, validated by the WTP study, not by intuition.
• Set the bundle at 87% of the a la carte sum, rounded down to a clean number.
• Write the comp plan with an attach SPIF of at least 6% on add-on ACV, and put the attach metric on the daily dashboard.
• Get CRO sign-off on the kill-criteria table while judgment is still clean.

9.4 Weeks 5-6: Ship the Beta

• Roll the SKU to 20% of the installed base as a controlled beta cohort, with the remaining 80% serving as the retention control.
• Co-term every add-on contract with its base renewal date to avoid renewal-cycle chaos and double legal review.
• Instrument attach, discount depth, sales cycle, and bundle ACV so the day-60 and day-90 readings are clean and comparable.

9.5 Day 60 and Day 90: Decide

• Measure against the kill-criteria table without renegotiating the thresholds.
• Roll to 100% only if day-90 attach exceeds 25% and gross retention holds within 200 bps of the control cohort.
• Otherwise pull, regroup, and pivot to a grandfathered base relaunch rather than letting sunk cost extend a failing experiment.

9.6 Month 6: Plan the Next Move

• Compare year-one attach trajectory against the ~34% median. If on track, design the year-two add-on — the second decoy variable that extends the ladder.
• If below median, pivot to the base-relaunch path before sunk cost compounds into a fourteen-month delay.
• Re-run the WTP study annually. Willingness to pay drifts with the market, with competitor moves, and with your own product investment; a two-year-old WTP curve is a stale anchor.

10. Frequently Missed Nuances

10.1 The Add-On Needs Its Own Roadmap

A parallel SKU that does not get its own product investment decays into a feature gate within a year. Buyers can tell the difference between a product that is improving and a static paywall, and the difference shows up in renewal behavior. Budget add-on roadmap capacity from day one, or the willingness to pay you measured in section 6 will erode out from under you.

10.2 Grandfathering Is a Trust Instrument

When you eventually do relaunch the base, grandfathering existing customers at their current price for a defined window — typically twelve months — is not generosity, it is a deliberate trust mechanism. It converts a potentially adversarial price event into a non-event for your installed base, preserves the reference value those customers provide in diligence calls, and buys you the goodwill to make the next change.

10.3 The Decoy Must Stay Visible

The a la carte option is the decoy that makes the bundle win. If a well-meaning marketing team "simplifies" the pricing page by removing the a la carte path, the bundle loses its structural advantage and the Ariely effect reverses. The decoy has to remain on the page, even though almost nobody chooses it.

Visibility is not optional; it is the mechanism.

10.4 Co-Terming Is a Billing Discipline, Not an Afterthought

Mismatched contract end dates between the base and the add-on create a quiet, compounding operational tax: partial-period invoices, duplicated legal review, confused renewal forecasting, and support tickets. Co-terming from day one is cheap. Retrofitting it across a base of mismatched contracts is expensive.

Decide it before the first add-on contract is signed.

Sources and Further Reading

1. Huber, J., Payne, J. W., & Puto, C. (1982). "Adding Asymmetrically Dominated Alternatives: Violations of Regularity and the Similarity Hypothesis." *Journal of Consumer Research.*
2. Ariely, Dan. *Predictably Irrational* (2008) — the Economist subscription decoy replication.
3. Van Westendorp, Peter. "NSS Price Sensitivity Meter" (1976), ESOMAR Congress proceedings.
4. Gabor, André, & Granger, Clive. "Price Sensitivity of the Consumer" — the Gabor-Granger method.
5. Bessemer Venture Partners, *State of the Cloud* — multi-tier net revenue retention benchmarking.
6. Bessemer Venture Partners, *BVP Nasdaq Emerging Cloud Index* — the public cloud cohort.
7. Carta, *State of Private SaaS* — add-on ACV as a percentage of base ACV.
8. Pavilion, *Compensation & Pricing Benchmark Report* — add-on attach rates and SPIF medians.
9. The Bridge Group, *SaaS Sales Compensation & Performance Report* — mid-cycle price-hike churn data.
10. RepVue, AE cohort dataset — multi-tier quota attainment differentials.
11. Levels.fyi, SaaS AE compensation data — OTE spread for reps with multi-tier products.
12. Gong, Revenue Intelligence research — three-tier versus two-tier close rates and cycle length.
13. SaaStr, Annual Survey — annual versus monthly take-rate by discount delta.
14. HubSpot, Inc. (NYSE: HUBS) — investor relations materials and 10-K filings on multi-Hub adoption.
15. HubSpot, Inc. DEF 14A proxy statement — SMB cohort churn risk disclosure.
16. Atlassian Corporation (NASDAQ: TEAM) — shareholder letters on cloud and Marketplace attach.
17. Salesforce, Inc. (NYSE: CRM) — annual report disclosures on multi-Cloud expansion.
18. Zoom Video Communications (NASDAQ: ZM) — investor materials on add-on attach.
19. Datadog, Inc. (NASDAQ: DDOG) — disclosures on customers using four or more products.
20. Snowflake Inc. (NYSE: SNOW) — consumption and workload-expansion disclosures.
21. Monday.com Ltd. (NASDAQ: MNDY) — 20-F filings on SMB unit economics and cross-product adoption.
22. Patrick Campbell / ProfitWell (acquired by Paddle) — pricing teardown research on add-on value thresholds.
23. Paddle, *Price Intelligently* — willingness-to-pay segmentation methodology.
24. Madhavan Ramanujam, *Monetizing Innovation* (Simon-Kucher) — willingness-to-pay-first product design.
25. Simon-Kucher & Partners — Global Pricing Study on bundling and tier architecture.
26. Thomas Nagle & John Hogan, *The Strategy and Tactics of Pricing* — anchoring and reference pricing.
27. Kahneman, Daniel, & Tversky, Amos — the anchoring-and-adjustment heuristic and prospect theory.
28. Richard Thaler, *Misbehaving* — mental accounting and transaction utility.
29. OpenView Partners, *SaaS Benchmarks Report* — expansion-ARR and NRR benchmarks.
30. KeyBanc Capital Markets, *SaaS Survey* — pricing model and retention benchmarks.
31. Bain & Company — Net Promoter Score methodology and pricing-perception research.
32. G2 and TrustRadius — buyer sentiment data on unbundling and pricing-change reactions.
33. U.S. Securities and Exchange Commission, EDGAR — primary-source 10-K, 20-F, and DEF 14A filings for every public company cited.
34. Tversky, Amos, & Kahneman, Daniel (1981). "The Framing of Decisions and the Psychology of Choice." *Science* — framing effects in choice.

TAGS: pricing-tiers,add-on-strategy,arpu-growth,bundle-psychology,margin-management