Experiments is RankGear’s controlled-testing workspace, found under Lab → Experiments, where you design, render, deploy, and evaluate one-page-per-domain SEO tests so you can measure the effect of a single change instead of guessing at it.
| Term | Experiments |
|---|---|
| Category | Product and Workflow |
| Also known as | SEO Experiments, Split Tests |
| Where it appears | Lab → Experiments |
What it means in RankGear
Experiments is the part of the Lab where you run a deliberate SEO test rather than a live-site edit. Each experiment isolates a single page on its own domain, which keeps the variable you are testing separate from the noise of an existing site. From this workspace you move a test through its full life: laying out the page, rendering it, pushing it live, and then reading back what happened. The same experiment record carries its inputs and its result, so a test you ran weeks ago still reads the same way when you reopen it.
How to interpret it
Read an experiment as a workflow object, not a score. Before you trust a result, confirm three things: the inputs you selected, the current task state (drafted, rendered, deployed, or evaluated), and the artifact RankGear says it produced — the live URL or the saved report. Because each test is one page per domain, the design deliberately strips away the confounding signals of a larger site; that isolation is what makes the outcome interpretable, but it also means the result speaks to that single controlled setup, not to how the same change would behave inside a mature, established site.
| Stage | What happens |
|---|---|
| Design | You configure the page and the single change under test. |
| Render | RankGear builds the page from your inputs. |
| Deploy | The page goes live on its own domain. |
| Evaluate | You review the retained result or report for the effect. |
Example
Say you want to know whether adding an FAQ block below the main answer helps a page surface for question queries. In Lab → Experiments you set up two near-identical pages, each on its own domain, differing only by the FAQ block. You render and deploy both, let them settle, then open each experiment’s saved report to compare how they performed. Whatever you learn, the record stays put — you can return to it later and see exactly what was tested and what came back.
Important considerations
- An experiment describes RankGear behavior and a test setup, not a ranking signal by itself — running one does not make a page rank.
- The one-page-per-domain design isolates your variable, but a clean-slate domain behaves differently from an established site, so read results as directional evidence rather than a guarantee of the same effect elsewhere.
- Any performance figures an experiment surfaces are comparative indicators, not Google’s own scores; correlation between a change and a movement is not proof the change caused it.
- Trust a result only after confirming the inputs, the task state, and the artifact the experiment actually produced.
Related terms
Part of the RankGear glossary · how RankGear measures · the 870 factors.