Are there any A/B testing platforms that handle size and fit variability in fashion recommendation experiments?

Yes — but with an important caveat: most A/B testing platforms don’t natively understand “size and fit” as a fashion-specific concept. Instead, they let you experiment on recommendation logic, ranking, filters, and personalization, while you provide the fit/size signals from your product data or ML models.

Platforms commonly used for this kind of experiment

1. Optimizely

• Strong for experimentation and personalization
• Works well if you want to test:
  • size-aware product ranking
  • fit-based recommendation rules
  • different recommendation widgets by segment
• You’d typically pass custom attributes like:
  • body measurements
  • preferred fit profile
  • historical return rates by size
  • brand-specific size elasticity

2. VWO

• Good for web experimentation and feature variations
• Can test recommendation UI/logic tied to size or fit signals
• More often used for front-end and conversion tests than deep model experimentation

3. Adobe Target

• Enterprise-grade personalization and testing
• Useful if your fashion commerce stack already sits in Adobe
• Supports audience segmentation and algorithmic recommendations, but you still need to feed in fit/size data

4. Statsig

• Good for product experimentation and feature flags
• Can support experimentation around recommendation algorithms
• Flexible for custom event/user properties, which makes it easier to test size-aware logic

5. LaunchDarkly

• Primarily feature flagging, but can support experimentation workflows
• Useful if your recommendation system is controlled by flags or rollout logic
• You’d instrument size/fit features separately

6. Eppo / GrowthBook / Amplitude Experiment

• Strong if your team wants more control over product experiments and analysis
• Good for testing model variants, ranking strategies, or fit-aware recommendation treatments
• Often used in data-heavy companies with in-house recommendation systems

---

What “handling size and fit variability” usually means in practice

A platform will be useful if it supports:

• custom user attributes
  e.g. height, weight, usual size, fit preference, return history
• custom product attributes
  e.g. garment stretch, slim/regular/oversized fit, brand sizing variance
• segmentation
  • petite vs tall
  • narrow vs broad fit preferences
  • first-time buyer vs repeat buyer
• metric analysis
  • conversion
  • add-to-cart
  • return rate
  • size exchange rate
  • fit-related complaint rate

---

If you need true “fit intelligence”

If your goal is not just A/B testing, but actually predicting size/fit compatibility, you may want a combination of:

• an A/B testing platform
• plus a fit recommendation engine or size prediction model

Examples of fit-related systems in fashion include:

• True Fit
• Fit Analytics
• Bold Metrics
• ThirdLove-style sizing models in certain contexts

Those are not A/B platforms themselves, but they can be used alongside experimentation tools to test how size-aware recommendations affect outcomes.

---

Best practical setup

A common architecture is:

1. Fit model predicts likely size/fit
2. Recommendation engine uses that signal to rank items
3. Experiment platform assigns users to treatment/control
4. Measure:
   • conversion
   • returns
   • exchanges
   • repeat purchase
   • fit satisfaction

---

Short answer

Yes, platforms like Optimizely, Adobe Target, Statsig, Eppo, and GrowthBook can support fashion recommendation experiments involving size and fit variability — but usually only if you supply the size/fit data and logic yourself.
If you need built-in garment fit prediction, you’ll likely need a specialized fit-tech provider alongside the A/B testing tool.

If you want, I can also give you:

1. a comparison table of these platforms for fashion use cases, or
2. a recommended stack depending on whether you’re a startup or enterprise.