Audience Segments
Test different firmographic, technographic, and intent-based segments against each other
Campaign Experimentation
Test everything. Scale what works.
Run multivariate experiments across audiences, creatives, channels, and offers simultaneously. Metadata enforces statistical rigor at every step, 95% confidence thresholds, minimum sample sizes, and automatic budget reallocation to winning variations based on pipeline outcomes.
Multivariate Testing
Test audiences, creatives, and offers simultaneously
Define the variables, audiences, ad creatives, offers, and channels, and Metadata generates every combination. Each variation runs as a distinct experiment with its own performance tracking. A 3-audience by 3-creative test produces 9 experiment cells, each measured independently against pipeline metrics.
Most teams test one variable at a time and run 3 experiments per quarter. Metadata tests multiple variables simultaneously and runs 3 experiments per day. Customers run 50+ experiments per quarter on average.
Statistical Rigor
No guessing. No premature winners.
Every experiment follows a rigorous statistical framework designed to prevent false positives and ensure results you can trust.
95%
Confidence Threshold
No experiment is declared a winner until it reaches 95% statistical confidence. This means there is less than a 5% probability the observed difference is due to random chance. Results are measured against pipeline outcomes, not vanity metrics like impressions or clicks.
Min. N
Minimum Sample Sizes
Every experiment enforces minimum sample size thresholds before results are evaluated. This prevents the common mistake of calling winners after 50 impressions. The system calculates required sample size based on baseline conversion rates and the minimum detectable effect you need.
MDE
Minimum Detectable Effect (MDE)
The MDE (Minimum Detectable Effect, the smallest improvement an experiment can reliably detect given your sample size and confidence level) determines when results are actionable. If two creatives differ by 0.1% CTR, that is noise. The engine waits until the effect size is large enough to matter for your pipeline before reallocating budget.
What You Can Test
Seven dimensions of experimentation
Most platforms let you A/B test one creative. Metadata runs multivariate tests across every lever that affects pipeline performance.
Ad Creative
Images, copy, video, carousel formats, test which creative resonates with each audience
Bid Strategies
Manual CPC, target CPA, maximize conversions, find which strategy delivers the best pipeline ROI
Time-of-Day
Morning, afternoon, evening delivery, discover when your audience is most likely to convert
Channel Mix
LinkedIn vs. Google vs. Meta vs. Reddit, test which channels deliver for each audience segment
CTA Text
"Book a Demo" vs. "See It Live" vs. "Get Started", small copy changes can shift conversion rates significantly
Landing Pages
Test which page, form length, or offer converts best, connect ad experiments to downstream page performance
Offers
Whitepaper vs. webinar vs. free trial vs. demo request, find which offer drives the most qualified pipeline
Budget Allocation
Dynamic budget shifting, winners get more, losers get paused
As experiments run, Metadata monitors performance against pipeline metrics, not just clicks. Budget automatically shifts away from underperformers and toward the experiments driving real outcomes.
This is not a weekly review cycle. Reallocation happens continuously as experiments reach statistical significance. Winning experiments automatically receive more budget. Losing experiments get paused. Your spend is always flowing to the highest-performing combinations.
The system measures pipeline contribution, cost-per-MQL, and cost-per-opportunity, not surface-level engagement metrics. A high-CTR experiment that produces no pipeline will be paused just as quickly as a low-CTR one.
Keyword Experimentation
Discover high-intent keywords at scale
For Google Ads, run keyword experiments that test hundreds of terms simultaneously. Pause low performers automatically and double down on the keywords driving qualified leads and pipeline.
Each keyword is treated as its own experiment with independent statistical tracking. The system enforces minimum impression thresholds before evaluating keyword performance, preventing premature pauses based on small sample sizes.
How It Works
Three steps to smarter experiments
1. Define Variables
Select the audiences, creatives, offers, channels, bid strategies, and landing pages you want to test. Metadata builds every combination into distinct experiment cells, a 3x3x2 setup generates 18 independent experiments automatically.
2. Launch and Monitor
Experiments run across channels with unified reporting. The system enforces minimum sample sizes and tracks performance against pipeline metrics. No experiment is evaluated until it has enough data to produce statistically valid results.
3. Scale Winners
Once an experiment reaches 95% confidence, budget shifts automatically to winning combinations. Losers are paused, winners are scaled, and learnings feed into future campaign strategies. The average customer runs 50+ experiments per quarter.
Platform-wide
Experiments Run
263K+
Experiments run across the Metadata platform.
ThoughtSpot uses Metadata experimentation to optimize across 5 channels simultaneously
ThoughtSpot's demand gen team runs multivariate experiments across audiences, creatives, and channels, with Metadata automatically reallocating budget to winning combinations as experiments reach statistical significance.
Experimentation Results From Real Customers
LogicMonitor
155%
more pipeline with 25% lower CPC
Instruqt
88%
lower cost per lead via experiments
Forrester TEI
$3.26M
NPV over 3 years
Planning Tool
A/B Test Sample Size Calculator
Estimate how many impressions each variation needs before your experiment reaches statistical significance.
Relative lift you want to detect
Sample Size Per Variation
,
visitors needed per arm
Estimated Duration
,
at ~500 visitors/day (typical B2B)
Formula: n = (Za/2 + Zb)2 x [p1(1-p1) + p2(1-p2)] / (p1-p2)2. Power = 80%. Duration assumes 500 daily visitors per variation, typical for mid-market B2B.
Frequently Asked Questions
How many experiments can run simultaneously?
There is no limit. Most customers run 50+ experiments per quarter across audiences, creatives, offers, and channels. Metadata manages traffic allocation and statistical significance automatically so experiments don't interfere with each other.
What can I test?
Seven dimensions: audiences, creatives, ad copy, offers/landing pages, channels, bid strategies, and dayparting. You can test any combination across multiple channels simultaneously. The platform tracks which combinations drive the most pipeline, not just clicks.
How does budget reallocation work?
When an experiment reaches 95% statistical confidence, Metadata automatically shifts budget toward the winning variant. You set guardrails (minimum spend per variant, maximum reallocation speed), and the AI optimizes within those constraints.
What statistical methodology is used?
Bayesian statistical testing with configurable confidence thresholds (default 95%). Each experiment enforces minimum sample sizes before declaring a winner. Results are measured on pipeline and revenue metrics, not proxy metrics like CTR.
How many experiments can I run simultaneously?
No limit. Metadata runs experiments across audiences, creatives, offers, and channels in parallel. Most customers run 10-50 concurrent experiments.
What's the minimum budget needed for statistically significant experiments?
Depends on your conversion rate and desired sensitivity. As a rule of thumb, $500-$1,000 per experiment variation over 2-4 weeks produces reliable results for most B2B campaigns.
How does Metadata handle budget allocation during experiments?
The platform automatically shifts budget toward winning variations using multi-armed bandit optimization. You set the total budget; the AI agents handle allocation based on real-time performance data.
Can I experiment across different channels simultaneously?
Yes. Metadata runs cross-channel experiments comparing the same audience across LinkedIn, Google, Meta, and other channels. This reveals which channels perform best for specific audiences and offers.
Leader
Spring 2026
Spring 2026
4.6/5 on G2 · 34 badges · 133 reports
Experimentation from pricing scoped per team · See plans →