Traditional A/B or switchback tests use fixed prices over fixed time periods (for example, one week at €50, the next week at €55).

AI Price Testing works differently:

• Prices are tested at the same time, not sequentially
• The system learns continuously from real customer behavior
• Better-performing prices are shown more often automatically
• You don’t have to wait for a test to “end” to see results

This approach balances learning and revenue performance at the same time.

AI Price Testing uses a multi-armed bandit model combined with interval-based price rotation. Different price points — including the original price — are shown in controlled time intervals, while the system continuously learns from live performance data.

Based on real results, the model increases exposure to price points that deliver stronger outcomes in conversion, revenue, and profit. This enables continuous optimization, balancing learning and performance without relying on fixed, time-boxed experiments.

Price testing is most useful when you want to understand how customers react to price changes and find the right balance between margin, volume, and profit. Common scenarios include:

• Testing higher prices (margin-focused)
• See how much margin you can safely add
• Run short high-price experiments
• Capitalize on trending or high-demand products
• Identify your true pricing ceiling
• Testing discounts (volume-focused)
• Boost seasonal demand

• Clear slow-moving products
• Increase order volume without over-discounting
• Find the optimal discount level
• Exploring multiple price points (Price Testing AI)

• Test several price variations at once
• Optimize for margin, volume, or both
• Compare price sensitivity with small price differences
• Learn continuously which prices perform best

Yes — but only within the boundaries you define.

You control:

• Which products are tested
• Allowed price ranges (higher / lower limits)
• Margin or cost guardrails
• When tests start and stop
• At each moment you can revert the prices to their original value. Price Testing does not make uncontrolled or irreversible changes.

There is no fixed “test period”. You have the option to set a timeframe.

The test starts learning immediately and improves continuously as more data comes in. The test can be stopped at any time.

Early signals often appear within days, sometimes sooner for high-traffic products.

Confidence improves as:

• More orders are collected
• Traffic volume increases
• Price differences become clearer

The system keeps improving the longer it runs.

Yes, but results will naturally take longer.

Price Testing works best for:

• Medium- to high-traffic products
• Products with stable demand
• Categories where price sensitivity matters

For very low-traffic items, it can still run, but learning will be slower.

Price Testing is often the first step in structured price testing, helping you understand how customers respond to different price points in a controlled way. Once you build confidence and gather learnings, you can move on to more advanced revenue optimization models, such as Adaptive Pricer or other rule- and context-driven pricing approaches.

These allow you to factor in additional signals — like demand patterns, inventory, margins, or business rules — and explore more complex pricing strategies beyond isolated price tests.