Portfolio Analytics

Not just the result of one test —
the health of your entire program

AB-Labz aggregates every experiment in your archive into a three-level dashboard: operational efficiency, product results, and business impact. One place to answer "is our testing actually working?"

Try free

Three levels of visibility

Each level answers a different question — from process hygiene to board-level reporting.

1

Operational

Is the process healthy?

Capacity, SRM Health Rate, Sample Sufficiency — hygiene checks on how experiments are set up and run, before you even look at results.

Capacity SRM Health Rate Sample Sufficiency
2

Product

Are hypotheses working?

Win Rate and Signal Rate tell you whether the team is generating real learning — and Consistency tracks whether that improves over time.

Win Rate Signal Rate Consistency Signal by tags
3

Business

What is the program worth?

GMV PnL and CR PnL aggregate the revenue impact of shipped experiments into a running total — ready for your quarterly business review.

GMV PnL CR PnL Positive / Negative / Noise
Level 1 · Operational

Is the testing process running well?

Tracks the hygiene of how experiments are set up and run — before looking at whether they win or lose.

Capacity

Share of filled days

75.1%
Target: 100%
75%

SRM Health Rate

Clean splits ratio

89.5%
Target: 100%
34/38

Sample Sufficiency

Adequately powered

84.2%
Target: 100%
32/38
Level 2 · Product

Are our experiments actually working?

Win Rate and Signal Rate tell you whether your team's hypotheses generate real signal — not just whether tests are running.

Win Rate

Positive key metric result

23.7%
Target: 20%

Share of experiments with a significant positive result on the key metric. Benchmark: 20–30%.

9/38

Signal Rate

Any detectable effect

34.2%
Target: 60%

Any meaningful effect on any metric. A higher bar than Win Rate — measures hypothesis quality, not just outcome.

13/38
Consistency

Does your team learn from its experiments?

Consistency is Signal Rate tracked as a trend over time. When you watch how Signal Rate moves month to month, you're asking a harder question: is the team getting better at identifying quality hypotheses before they launch?

A rising line means analysts are learning from past experiments. A flat line means the team is still running on intuition alone.

Signal rate — monthly

↑ Trending up

100% 75% 50% 25% 0% Aug 25 Sep 25 Oct 25 Nov 25 Dec 25 Jan 26 Feb 26 Mar 26 Apr 26 May 26 Jun 26 Jul 26
Signal rate Trend line
Level 3 · Business

What is the AB department actually worth?

Effect is measured as normalised ARPU delta per user per month. Multiply by your MAU — and that's your number for the business review.

GMV PnL

Cumulative GMV per user · 30d

Total uplift

+13.80

2025-10 2025-11 2025-12 2026-01 2026-02 2026-03 2026-04 2026-05 2026-06

CR PnL

Cumulative CR delta · pp

Total uplift

+4.7 pp

2025-10 2025-11 2025-12 2026-01 2026-02 2026-03 2026-04 2026-05 2026-06

Positive Experiments

Positive GMV impact

+21.85
per user 8 experiments
Sticky add-to-cart button +5.82
Smart search autocomplete +4.14
Product image gallery v2 +3.27
Free shipping threshold banner +2.89
Express checkout flow +2.41
Personalised homepage feed +1.73
Review summary widget +0.94
Wishlist share feature +0.65

Negative Experiments

Negative GMV impact

−8.05
per user 2 experiments
Countdown timer on product page −5.44
Aggressive upsell modal −2.61

Noise

Inconclusive GMV signal

+2.70
per user 8 experiments
Category filter sidebar redesign +3.18
Related products carousel −0.82
Same-day delivery badge −0.47
Guest checkout option +0.93
Product comparison tool −0.34
Loyalty points display −0.22
Bundle discount banner +0.44
Newsletter popup delay ?

See your program's health in one dashboard

Portfolio Analytics is included in all Corporate plans.

Register free