How to Test Different Pricing Options With Real Users

Methods like the Van Westendorp Price Sensitivity Meter and conjoint analysis are genuinely useful tools. But they capture stated preferences, not actual behaviour. When a respondent answers a survey question about whether they would pay $49 per month for a software product, they are answering hypothetically, without their credit card in hand, without the friction of a checkout flow, and without the competitive context they would normally face.

That said, this is not an argument against survey methods. Van Westendorp is legitimately valuable when launching a brand new product with no pricing history. It produces four price thresholds (too cheap, cheap, expensive, too expensive) and identifies both the acceptable price range and the optimal price point. That range gives you a realistic bracket to work within before you invest in live testing. Conjoint analysis, specifically choice-based conjoint, helps you design tier structures and feature bundles by revealing which feature-price combinations your target customers value most. The problem arises when businesses treat these survey outputs as final answers rather than as inputs to live experimentation.

The most reliable approach to price sensitivity testing follows a clear three-stage sequence. Each stage builds on the last, and skipping any of them increases your risk of a costly misstep.

The most common mistake in pricing experiments is changing too many variables at once. A reliable pricing A/B test isolates one variable — whether that is the price point, billing frequency, or tier structure — and runs it exclusively on new user cohorts. Exposing existing customers to different prices risks churn and creates a perception of unfairness that can damage your brand in ways that are difficult to reverse.

If you are new to pricing experiments, start with lower-risk variations before shifting your base price outright. Testing charm pricing ($9.99 versus $10.00) or introducing a decoy tier to anchor perception toward your preferred plan carries significantly less revenue risk than a wholesale price change. Anchoring, displaying your most expensive plan first to make other options appear more reasonable, is a technique that conversion rate optimisation research consistently links to meaningful conversion lifts when paired with social proof near call-to-action buttons, though results vary by context, audience, and page design. These are sensible starting points that build your experimentation capability before you attempt more aggressive tests. The first test is always the hardest part; many teams benefit from running a focused Pricing First-Impression Test to see how initial perception shapes downstream behaviour.

For Australian SMEs with modest traffic volumes, sample size is where most pricing experiments go wrong. The relationship between your baseline conversion rate, your minimum detectable effect (MDE), and the sample size you need is not intuitive. The lower your baseline conversion rate, the more traffic you need to detect a meaningful difference. For a typical SaaS pricing page running at the industry median of around 3.8 per cent conversion, detecting a 10 per cent lift requires patience and a realistic test duration.

For businesses with limited monthly traffic, target an MDE of 10 to 15 per cent rather than the 3 to 5 per cent that large consumer platforms can realistically detect. Aim for at least 100 conversions per variation before drawing any conclusions. Use free calculators such as Evan Miller's or Statsig rather than manual calculation; they handle the statistical formula automatically once you input your baseline rate, target MDE, significance level (typically 5 per cent), and desired power (typically 80 per cent).

• Never stop a test early because an early result looks promising. Statistical significance reached on day three with low volume is almost always noise.
• Never run a pricing test during a seasonal spike. Set your target sample size before the test begins and commit to running it for at least two full billing cycles for subscription products.
• That discipline is what separates actionable results from expensive guesses.

Your website traffic is not a clean test audience. Organic visitors, retargeted users, and referral traffic all behave differently on a pricing page — they arrive with different levels of intent, different familiarity with your brand, and different price sensitivity profiles. When you run a live A/B test across your full traffic mix, you are averaging across those segments. The result is data that may be statistically valid but difficult to act on.

Recruiting real users who match your target customer profile gives you cleaner, more actionable data, particularly when you want to understand not just what users clicked but why they hesitated, what confused them, and where the pricing page lost them. This is where platforms built for matched-user research earn their place in a pricing strategy.

Dlyte connects Australian businesses with real users matched to their target market profile, allowing teams to run price sensitivity testing and pricing experiments across people who reflect their actual buyer demographics. Where analytics tools show you where users drop off, Dlyte surfaces the reasoning behind those drop-offs: hesitation, confusion, perceived value mismatches, and the specific moments where a pricing page loses a potential customer. For Australian founders and growth teams who need pricing intelligence without agency timelines or agency budgets, it is a practical way to build that research capability in-house.

Under the Australian Consumer Law, businesses cannot create a misleading impression about the price a customer will pay, regardless of what is disclosed elsewhere on the page. Drip pricing (advertising a headline price that excludes mandatory fees) and false discounts (using inflated "was/now" prices that were never genuinely held at the higher price for a reasonable period) are enforceable breaches. The ACCC has made misleading pricing a declared enforcement priority through 2026, and penalties under the Competition and Consumer Act 2010 for corporate breaches can be substantial — in some cases reaching $50 million or more per incident depending on the provision and circumstances. Confirm the specific penalties applicable to your situation with legal counsel.

So what is the best way to test different pricing options with real users? Not a single method, but a layered one where each stage informs the next. Van Westendorp defines the acceptable range and gives you guardrails before you spend a dollar on live traffic. Conjoint analysis sharpens your package design and reduces the number of live experiments you need. Live A/B tests then validate what actually converts with real purchase behaviour. And matched real-user research explains the human reasoning behind the numbers that analytics alone cannot surface.

Data tells you where the drop-off happened. Real users tell you why. That distinction compounds over time, because it informs every pricing decision that follows, not just the one you are currently testing. Australian businesses that build this discipline into their product and marketing process, rather than setting prices once and hoping for the best, develop a meaningful advantage over competitors who are still guessing.