Beta
SEO test results analyzer: did the change work?
Paste daily clicks for your control and variant pages and get a Positive, Negative, or Inconclusive verdict.
credit cost
1
Overview
What it does
The test results analyzer takes daily clicks for your control and variant page groups, covering a period before the change and the test period after it. It tells you whether the change worked.
What you paste. A CSV or TSV with three columns: date, group (control or variant), and clicks. Impressions are optional. Add the test start date, the test type, and optionally a note on what changed.
Checks before any credit is spent. Both groups must be present, both periods must be whole weeks of at least 14 days, and the pre-period must be at least as long as the test period. Missing days are reported back to you, not filled in.
What the tool calculates. It compares how the ratio of variant to control clicks moved between the pre-period and the test period, and puts a 95% range around that estimate. The range is adjusted for short samples and for day-to-day correlation in the data. The verdict follows from the range:
Positive. The whole range is above zero.
Negative. The whole range is below zero.
Inconclusive. The range includes zero.
When the range only just clears zero, the result says so.
The model then explains the result: the verdict, what the numbers show, caveats specific to your data, and one next step. It cannot change the verdict or calculate its own numbers.
Planning a test? Start with the SEO split test planner to check your pages can support one.
The sample below finishes the Northtrail example from Title bet and the split test planner.
Method
How it works
1. Paste daily clicks
Date, group, and clicks for control and variant, covering the pre-period and the test period.
2. Add the test details
The test start date, the test type, and optionally what changed.
3. Get the verdict
Positive, Negative, or Inconclusive, with the estimated effect, its range, and a next step.
Example
Sample output
Input
Data: northtrail-sample-test-data-2wk.csv (56 rows, 2026-05-18 to 2026-06-14)
Test start date: 2026-06-01
Test type: Title / meta
What changed: Added "Waterproof & Lightweight"-style attributes to category page titles on the variant group.
Confidence level: 95%
14 days before the change and 14 after, which matches the two-week plan the split test planner produced.
Result
Value | Result |
Verdict | Positive |
Effect | 6.39% |
Range (95%) | 1.66% to 11.35% |
Clicks a day | 150.9 |
Over the test period | 2113 clicks |
Control, before → after | 2280.6 → 2313.5 |
Variant, before → after | 2252.3 → 2435.6 |
Observed daily swing | 3.6% |
Periods | 14 days before, 14 days after |
The verdict
This test reads positive: adding the attribute-style text to category titles produced an effect of +6.39%, with a 95% range of +1.66% to +11.35%, over the two-week test window.
What the numbers show
Control moved from a daily mean of 2280.6 clicks in the pre-period to 2313.5 in the test period, while variant moved from 2252.3 to 2435.6 over the same stretch. Netting out the control's movement, the variant's gain works out to +6.39%, or about 150.9 clicks a day and 2,113 clicks total across the 14-day test window. The interval runs from +1.66% to +11.35% and stays entirely above zero, so the test did separate the variant from the control rather than just catching noise. The pre-period ratio between the two groups only swung about 3.6% day to day, which is a relatively calm baseline.
Caveats
Short window: both pre and test periods run 14 days, so this covers two specific two-week stretches, not a longer trend.
No flagged days: no missing days, unpaired days, zero-click days, or outlier days were logged, so there's nothing here undermining the read.
Comparison only: this measures variant against control across one window — anything else that changed on these category pages during that time is baked into the number.
Next step
Roll the title/meta change out. Before treating this as settled, confirm nothing else shifted on the variant pages during the test window (merchandising, internal linking, seasonality specific to those categories) that could ride along with the title change.
Northtrail Outfitters is a fictional retailer. All numbers are sample data.
FAQ
Common questions
How much does the SEO test results analyzer cost?
More tools