.png)
Frameworks, core principles and top case studies for SaaS pricing, learnt and refined over 28+ years of SaaS-monetization experience.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Maximizing Revenue with Adobe Target for SaaS Pricing Experiments
Pricing pages look deceptively easy to test. Change £49 to £59, move the annual discount, rename a package, split traffic, and wait for a winner. In SaaS, however, the revenue consequence may appear weeks or months after the page view. A price that lifts checkout conversion can lower ARPA. A package that creates more demo requests can attract weaker accounts. A discount that wins the first purchase can damage renewal economics.
Adobe Target is powerful enough to run these experiments at enterprise scale. As of March 2026, Target Standard supported A/B and multivariate testing, Auto-Allocate, rules-based and geographic targeting, and server-side optimisation; Premium added capabilities including Automated Personalization and Auto-Target. Adobe licenses both editions against Annual Page View Traffic rather than tester seats. Yet the tool does not decide what deserves to be tested, which customer should enter the experiment, or which financial measure should determine the winner.
Monetizely's position is that Adobe Target maximises SaaS pricing revenue when it is used as a controlled decision system around pricing, not as a machine for randomly changing prices. The highest-value programme tests packaging first, price presentation and rate second, and personalisation last - while judging each experiment against revenue per qualified account and downstream customer quality rather than click-through or signup conversion alone.
Adobe Target works best when pricing remains a business decision rather than a page-editing exercise
The most expensive pricing-test mistake happens before Adobe Target receives any traffic. Teams start with a visible element - the monthly price, discount badge or call-to-action - because Target makes that element easy to change. What they should start with is the commercial question.
Monetizely's 5-Step Pricing Framework keeps those decisions in order. Goals and Segmentation defines the business result and the customers whose willingness to pay matters. Packaging decides which capabilities belong together. Pricing Metric establishes what the customer pays for, such as seats, transactions or usage. Rate Setting determines the actual price and discount structure. Operationalization turns the design into something that sales, product, billing, analytics and finance can execute without ambiguity. The sequence, discussed in Monetizing Agentic AI, matters greatly for Adobe Target: experimentation can sharpen decisions inside each step, but it cannot repair a price test built on the wrong segment or the wrong unit of value.
Consider a security SaaS vendor testing £80 versus £95 per user per month. A conventional web test asks which number produces more purchases. The framework forces harder questions first: Are 50-seat technology firms and 2,000-seat banks being mixed into one test? Does the package put audit reporting in the correct tier? Should seats even be the expansion metric? Only after those issues are settled does £80 versus £95 become a useful experiment.
That distinction drives our assessment of Target itself. The software is particularly strong at the final mile of pricing experimentation, where a clear commercial hypothesis has to be delivered consistently and measured.
The scorecard below grades Adobe Target on the three parts of the framework most directly exposed by the product.
| 5-Step Pricing Framework step | Grade | Monetizely assessment |
|---|---|---|
| Packaging | A- | Standard and Premium create a clear capability ladder, with advanced personalisation and Auto-Target reserved for Premium as of March 2026. |
| Pricing Metric | B- | Annual Page View Traffic scales with deployment, but page views reflect the cost and reach of experimentation better than the economic value created by a successful experiment. |
| Operationalization | A | A/B, multivariate, API-based delivery, Auto-Allocate and integration with Adobe Analytics give enterprise teams several ways to deliver and measure controlled changes. |
Adobe therefore gets the experimentation product right more clearly than it gets the monetisation of that product right. Its own meter follows traffic, while its customer promise centres on conversion, personalisation and revenue improvement.
Adobe has built the machinery to test price at scale while keeping its own price deliberately opaque
Adobe Target has evolved from a web optimisation tool into infrastructure that can deliver experiments across a much broader customer journey. Adobe's February 2019 release history records support for its Single Page Application Visual Experience Composer, aimed at React and Angular experiences without constant developer intervention; by October 2019, a Delivery API supported experience retrieval for web, single-page applications and mobile use cases.
By 2026, the product description had formalised Standard and Premium around the same Annual Page View Traffic licence metric. Standard included A/B testing with Auto-Allocate, multivariate tests and server-side optimisation, while Premium added Automated Personalization, Auto-Target and Recommendations.
Adobe's public commercial presentation has remained much less precise. Its current UK and international pricing pages ask buyers to request customised pricing and say price is affected by product options, deployment volume and omnichannel delivery.
The evolution is easier to see in one exhibit.
| Date | Adobe Target structure | What changed for a SaaS pricing team | Primary source |
|---|---|---|---|
| February 2019 | Single Page Application Visual Experience Composer support | Pricing experiences in modern React or Angular journeys became easier to change without rebuilding the whole application workflow | Adobe release history |
| October 2019 | Delivery API | Experiment delivery extended more naturally into server-side and application experiences rather than remaining a browser-page exercise | Adobe release history |
| November 2025 | Analytics for Target supported Analytics metrics and segments as the reporting source | Experiment exposure could be analysed against a broader analytics dataset instead of relying on Target's native reporting alone | Adobe Experience League |
| March 2026 | Standard and Premium licensed on Annual Page View Traffic | Traffic volume remained the commercial meter while advanced personalisation stayed packaged in Premium | Adobe product description |
| August 2026 | Custom quote based on product choices, volume and channel scope | Buyers can understand the drivers of spend but still cannot calculate a public entry price before talking to sales | Adobe pricing page |
The message is striking: Adobe has steadily removed friction from running sophisticated experiments, yet it still puts material friction in the purchase journey for Target itself. For an enterprise product, sales-led pricing is understandable; in our view, the absence of even public volume bands makes budget planning harder than the product's underlying meter requires.
Scale makes that point economically important. Adobe reported Digital Experience revenue of $5.86 billion for fiscal 2025, up 9% year on year, with Digital Experience subscription revenue of $5.41 billion, up 11%. The company exited fiscal 2025 with total ARR of $25.20 billion. Target sits inside a very large enterprise software franchise, not an early product that needs pricing opacity while it discovers its market.
Revenue per qualified account prevents a conversion win from becoming a pricing loss
The central problem in SaaS experimentation is the lag between a pricing-page action and economic value.
Adobe itself tells Target users to establish the experiment goal before configuring an activity, offering revenue per visitor as one example. Its May 2026 documentation also warns that Revenue per Visit carries more variance than a simple conversion measure and generally takes 20% to 30% longer to reach the same confidence for the same measured lift.
For SaaS, even RPV is often too early. A £30 self-serve tool can reasonably attach revenue to the checkout session. A £60,000 enterprise SaaS deal cannot. The pricing page may create a demo, qualification may take a week, procurement six weeks, and the first meaningful retention signal months longer.
Adobe Analytics for Target, or A4T, makes a better design possible. As of May 2026, Adobe allowed Analytics to become the reporting source for a Target activity, making Analytics metrics, including custom or calculated measures, available for experiment analysis. Adobe also passes activity and experience information server-to-server into Analytics.
For a SaaS pricing programme, we would therefore build a measurement chain rather than declare a winner from one conversion.
| Pricing experiment | Primary question | Adobe Target setup | Decision metric | Guardrail |
|---|---|---|---|---|
| Package A versus Package B | Which bundle captures more willingness to pay? | A/B test by eligible customer segment | Expected first-year ARR per qualified account | Activation and product adoption |
| £79 versus £89 entry rate | How much demand is lost for each increase in realised price? | Controlled A/B test on new prospects | Gross profit per eligible account | Paid conversion and refund/cancellation rate |
| Monthly versus annual-first presentation | Does commitment framing increase contract value? | A/B experience with consistent underlying offers | Expected contract value per visitor | Overall purchase or demo rate |
| 10% versus 15% annual discount | Does extra discount buy enough incremental commitment? | A/B test with identical packaging | ARR less discount cost per qualified account | Annual-plan share |
| Good-Better-Best package emphasis | Which tier should receive visual prominence? | A/B or multivariate test | New ARR by package mix | Entry-tier conversion |
| Segment-specific page | Does a distinct value story improve monetisation for a known segment? | Experience Targeting after a segment hypothesis is validated | ARR or pipeline value per eligible account | Cross-segment consistency |
A click metric can tell the growth team whether a button works. Pricing needs a financial metric that tells the company whether the customer economics work.
Adobe Target can itself report conversion, revenue and engagement measures, and its May 2026 documentation allows estimated conversion value to be attached to non-revenue goals. Target's estimated revenue lift is based on the difference in Revenue per Visit between the winning and control experiences multiplied by activity traffic.
We would not put that projected lift straight into the board forecast. For a subscription product, realised revenue also depends on cancellations, upgrades, expansion and renewal. Target should determine experiment exposure; the company's billing, CRM and finance data should determine whether the commercial outcome endured.
Auto-Allocate can reduce the cost of learning but cannot rescue a weak pricing hypothesis
Traditional A/B testing contains a hidden expense. Once one treatment begins to outperform, the company may continue sending half of new prospects to the weaker experience while it waits for enough evidence.
Adobe designed Auto-Allocate to reduce that loss. As documented in May 2026, it shifts more traffic towards better-performing experiences while preserving a portion for continued exploration. Adobe describes 80% of traffic as being served through Auto-Allocate and 20% randomly, with returning visitors remaining on their original experience. The algorithm begins after each experience has received at least 1,000 visitors and 50 conversions, and Adobe says users who wait for its winner indication retain a 5% false-positive rate.
Research supports the commercial logic behind adaptive allocation. A peer-reviewed Marketing Science study published in 2019 modelled multi-armed-bandit pricing and found, in a calibrated simulation based on an existing pricing field experiment, a 43% profit improvement during the test period and 4% annually relative to its comparison approach.
Yet a bandit answers a narrower question than many pricing teams assume: given the objective we supplied, which treatment should receive more traffic? Feed it paid conversion and it will optimise paid conversion. That does not prove that the treatment maximises ARR, gross profit or lifetime value.
Adobe's own traffic requirements expose another limit. For Auto-Target, its 2026 guidance gives a rule of thumb of 1,000 visits and at least 50 conversions per day and per experience for a conversion goal; Revenue per Visit requires still more data, including at least 1,000 conversions per experience. Many enterprise SaaS companies simply do not generate that volume on their pricing pages.
The operating choice should therefore follow traffic economics.
| SaaS situation | Best Target mode | Monetizely's position |
|---|---|---|
| High-traffic self-serve SaaS with thousands of purchases | A/B followed by Auto-Allocate | Use adaptive allocation once the financial goal and guardrails are established |
| High-traffic product with meaningful customer segments | A/B first, then Auto-Target | Prove the average pricing effect before allowing personalisation to choose experiences |
| Mid-market SaaS with hundreds of monthly demos | Conventional A/B by account or stable visitor cohort | Protect statistical clarity; optimise on qualified pipeline rather than clicks |
| Enterprise SaaS with dozens of monthly opportunities | Target for message/package experiments, not live price randomisation | Combine market research, sales evidence and controlled cohort changes; web traffic alone is too thin |
| Existing contracted customers | Stable contractual treatment | Do not let a web experiment silently change an agreed entitlement or renewal price |
The underlying point is simple. Adobe Target is strongest where the SaaS business has enough traffic for controlled learning and enough downstream data to distinguish a cheap conversion from a valuable customer.
Pricing experiments also require more restraint than ordinary headline tests. In January 2025, the US Federal Trade Commission reported that intermediaries could use data such as location, demographics and browsing behaviour to enable individualised prices. Its investigation focused on possible effects on privacy, competition and consumer protection. Monetizely's position is therefore to segment pricing experiments by defensible business variables - plan eligibility, company size, geography, acquisition motion or clearly defined cohorts - rather than quietly infer the maximum price an individual appears willing to pay.
Experimentation vendors show that buyers increasingly expect spend to scale with actual use
Adobe is not operating in a vacuum. Experimentation software vendors increasingly expose both their packaging and their meters, giving buyers clearer ways to estimate what adoption will cost.
As of 13 August 2026, Statsig's free Developer tier included 2 million monthly events, while Pro cost $150 per month with 5 million events included and $0.05 per additional 1,000 events; its Enterprise contracts could be event- or experiment-based. LaunchDarkly's current Foundation pricing showed separate infrastructure-related meters, while its archived April 2025 page explicitly priced experimentation at $3 per 1,000 experimentation MAU on Foundation. VWO's current pricing page presented tiered experimentation capabilities and explicitly directed buyers to discuss MAUs, events and limits during plan selection.
Putting those models beside Adobe highlights where the market has moved.
| B2B experimentation vendor | Published structure and date | Meter visible to buyer | Commercial lesson |
|---|---|---|---|
| Adobe Target | Standard/Premium; custom quote; March-August 2026 | Annual Page View Traffic is defined legally, but public rate bands are not shown | Strong enterprise architecture, weak pre-sales price visibility |
| Statsig | Free, $150/month Pro, custom Enterprise; accessed 13 August 2026 | Metered events; enterprise can use event- or experiment-based contracts | A buyer can estimate the cost of adoption before procurement |
| LaunchDarkly | Foundation plus custom enterprise plans; accessed 13 August 2026 | Service connections, client-side MAU and other usage; April 2025 archive separately exposed experimentation MAU | Pricing evolves as the underlying product architecture expands |
| VWO | Growth, Pro and Enterprise capability tiers; accessed 13 August 2026 | Plan selection explicitly considers MAUs, events and limits | Packaging and usage scope are discussed together |
No competitor model proves that Adobe should copy it. They do show that enterprise experimentation no longer requires complete opacity around what makes the bill grow.
Practitioners make a related point: the tool earns its keep only when experimentation becomes easier to run and easier to connect to business decisions. On their vendors' official pricing pages, operators describe that value in unusually direct terms.
The quotes point in the same direction as the product structures. Buyers want faster learning, credible results and a cost they can connect to the amount of experimentation they actually perform.
Adobe Target's next pricing reset should make its own value easier to buy and easier to expand
What does Adobe Target get right? Packaging is coherent. Standard covers serious experimentation rather than functioning as a crippled entry product, while Premium reserves machine-learning personalisation and Recommendations for customers with deeper needs. As of March 2026, both packages also share the same basic traffic meter, which keeps the commercial logic understandable once the contract is visible.
Operationalization is stronger still. A SaaS company can use a visual workflow for web experiments, deliver server-side experiences through APIs, use Auto-Allocate, and connect Target activity data with Adobe Analytics through A4T. Those capabilities allow one experimentation programme to reach marketing pages, onboarding flows and application experiences.
What does Adobe Target get wrong? Its pricing meter stops one step short of value, while its public pricing page stops several steps short of budget clarity. Annual Page View Traffic is measurable and correlated with Adobe's delivery workload, but a billion page views with inconclusive tests are worth less to the customer than 100 million page views producing a material revenue improvement. At the same time, Adobe's August 2026 pricing page discloses the drivers of price without publishing the rate bands needed to estimate spend.
Monetizely would not replace page-view pricing with outcome fees. Adobe cannot control whether a customer's test idea is intelligent, whether its product is competitive, or whether its organisation rolls out the winner. Charging a share of measured revenue lift would create endless arguments over attribution.
Adobe Target's next pricing reset should keep Annual Page View Traffic as the primary meter but turn it into a transparent annual-commit model: publish entry bands for Standard and Premium, include a stated traffic allowance, disclose marginal overage bands, and let high-volume enterprise customers negotiate committed-volume discounts. The architecture preserves a meter Adobe can verify while making the customer's three-year cost much easier to forecast.
For Adobe, such a reset would align the sales experience more closely with the product philosophy. Target exists to replace guesswork with evidence. Its own customer should not have to guess what the first meaningful deployment will cost.
The largest revenue gain comes from changing what the company learns, not from running more tests
Suppose a self-serve SaaS business receives 500,000 eligible pricing-page visitors a year. Its control converts 2.5% at £600 of first-year contract value. That produces £7.5 million of first-year booked value before the effects of churn, refunds and expansion.
Now compare three experiment outcomes.
| Pricing treatment | Conversion | First-year value per customer | Implied first-year booked value | Change versus control |
|---|---|---|---|---|
| Control | 2.50% | £600 | £7.50m | - |
| Lower-price winner on conversion | 2.90% | £500 | £7.25m | -£0.25m |
| Higher-price treatment | 2.30% | £700 | £8.05m | +£0.55m |
| Better package mix | 2.55% | £680 | £8.67m | +£1.17m |
The "winning" conversion rate in the second row destroys £250,000 of booked value. A package change with only a small conversion improvement creates more than £1 million. Conversion optimisation and pricing optimisation are not the same job.
Real experiments can produce substantial economic effects. A peer-reviewed Management Science field experiment published in 2017 tested an online pricing strategy for five weeks and reported an 11% revenue increase while preserving a retailer-specified margin floor. Another 2024 Management Science pricing meta-experiment at Airbnb found that individual-level randomisation could materially bias measured treatment effects in an interconnected marketplace, with at least 20% of the estimated effect in the studied platform-fee experiment attributable to interference bias and removed through cluster randomisation.
The lesson for SaaS is not to copy retail pricing algorithms. It is to treat experiment design as part of pricing strategy. The customer unit, financial goal and downstream observation window matter as much as the button that launches the activity.
For operators building an Adobe Target pricing programme, five decisions should now move to the top of the agenda:
Give pricing experiments an executive revenue target. Ask the programme to improve realised ARR, gross profit or another board-level economic measure over a defined period rather than maximising the number of statistically significant website wins.
Create a permanent pricing control group. Preserve a clean cohort that does not receive the stream of package, discount and message changes. Over time, that group gives leadership a better view of whether the entire programme is creating incremental value rather than simply moving conversions between experiments.
Move the unit of randomisation up to the account when several users influence one purchase. A CFO and an administrator from the same prospect should not land in conflicting price experiences because their browser cookies differ.
Fund experimentation according to the revenue at risk. A £20 million self-serve funnel deserves stronger instrumentation and more testing capacity than a low-traffic marketing page, even when the latter is easier to modify.
Make every winning price test face a second decision. Before rollout, force finance and product leadership to ask whether the observed win remains attractive after discount cost, customer mix, retention risk and sales behaviour are included.
Adobe Target can make pricing experimentation faster. Revenue rises when the organisation becomes more disciplined about what it asks Target to learn.
Assumptions
The SaaS scenario models 500,000 eligible annual visitors and treats first-year booked value as conversion rate multiplied by eligible traffic and average first-year customer value. It excludes churn, expansion, payment failure, taxes, implementation revenue and experiment-platform cost so that the effect of conversion and customer value can be seen separately. Figures are modelling ranges, not Adobe or market benchmarks.
Footnotes
Monetizing Agentic AI, Monetizely, 2026. https://www.amazon.com/Monetizing-Agentic-AI-Handbook-Transformation/dp/B0H7Z13VKJ/
Adobe, “Adobe Target - Product Description,” updated 26 March 2026. https://helpx.adobe.com/in/legal/product-descriptions/adobe-target.html
Adobe Experience League, “What is Analytics for Target (A4T)?”, updated 2025-2026. https://experienceleague.adobe.com/en/docs/target/using/integrate/a4t/a4t
Adobe Experience League, “Adobe Target Release Notes / Release History.” https://experienceleague.adobe.com/en/docs/target/using/release-notes/release-notes
Adobe, “Adobe Target Pricing & Packages,” accessed 13 August 2026. https://business.adobe.com/au/products/target/pricing.html
Adobe Inc., fiscal 2025 Form 10-K and 2026 annual materials. https://www.sec.gov/Archives/edgar/data/796343/000079634326000003/adbe-20251128.htm
Adobe Experience League, “How Long Should I Run an A/B Test?”, updated 12 May 2026. https://experienceleague.adobe.com/en/docs/target/using/activities/abtest/sample-size-determination
Adobe Experience League, “Success Metrics,” updated 12 May 2026. https://experienceleague.adobe.com/en/docs/target/using/activities/success-metrics/success-metrics
Adobe Experience League, “Estimating Lift in Revenue,” updated 12 May 2026. https://experienceleague.adobe.com/en/docs/target/using/administer/reporting/estimating-lift-in-revenue
Adobe Experience League, “Auto-Allocate Overview,” updated 12 May 2026. https://experienceleague.adobe.com/en/docs/target/using/activities/auto-allocate/automated-traffic-allocation
Kanishka Misra, Eric M. Schwartz and Jacob Abernethy, “Dynamic Online Pricing with Incomplete Information Using Multiarmed Bandit Experiments,” Marketing Science 38, no. 2, 2019. https://pubsonline.informs.org/doi/10.1287/mksc.2018.1129
Adobe Experience League, “Auto-Target,” updated 2026. https://experienceleague.adobe.com/en/docs/target/using/activities/auto-target/auto-target-to-optimize
U.S. Federal Trade Commission, “FTC Surveillance Pricing Study Indicates Wide Range of Personal Data Used to Set Individualized Consumer Prices,” 17 January 2025. https://www.ftc.gov/news-events/news/press-releases/2025/01/ftc-surveillance-pricing-study-indicates-wide-range-personal-data-used-set-individualized-consumer
Statsig, “Pricing,” accessed 13 August 2026. https://www.statsig.com/pricing
LaunchDarkly, “Pricing,” current and April 2025 archived pricing page, accessed 13 August 2026. https://launchdarkly.com/pricing/ and https://launchdarkly.com/pricing/archive-apr2025/
VWO, “Pricing & Plans,” accessed 13 August 2026. https://vwo.com/pricing/
Marshall Fisher, Santiago Gallino and Jun Li, “Competition-Based Dynamic Pricing in Online Retailing: A Methodology Validated with Field Experiments,” Management Science 64, no. 6, 2018; first published online 27 June 2017. https://pubsonline.informs.org/doi/10.1287/mnsc.2017.2753
Management Science, “Reducing Interference Bias in Online Marketplace Experiments Using Cluster Randomization: Evidence from a Pricing Meta-experiment on Airbnb,” published online 5 April 2024. https://pubsonline.informs.org/doi/10.1287/mnsc.2020.01157