RankForge — SEO Content & Organic Growth Command Center

You cannot tell whether the change you made caused the ranking you got — unless you kept a control group.

Somebody rewrites a set of titles in March. Rankings improve in May. The two get joined with a line and that line goes in the report. Nothing about it can be checked, because what those pages would have done untouched was never observed.

RankForge is built around observing it. Its flagship screen is a difference-in-differences: what the changed pages did, minus what a matched set of deliberately untouched pages did, over the same window. In the worked sample site the headline reads 3.53 places; the held-back pages moved 1.37 on their own because a core update landed mid-window; the honest answer is 2.16.

It refuses where it cannot answer. When the groups are missing, under twenty-five a side, or did not start from comparable positions and ages, the screen prints the reason instead of a number. It also declines to extrapolate a click-through curve past the positions you have actually held, to treat a page published six weeks ago as having failed to rank, and to reconcile your query export with your page export — low-volume terms are withheld for privacy, so the two never add up and any tool where they do has invented the difference.

Twenty-one screens. The lift measurement · how a page ages, with its denominator and its median on show and the curve stopped where the sample thins · why average position cannot be read the way it is read · your own click-through curve · the query ledger and the part withheld before you saw it · pages competing with your own pages · what a page costs and when it pays back · refresh against publish, honest about which half rests on evidence · internal links · indexation · links stated as association rather than cause · what the results page itself now absorbs · de-trended seasonality · branded against non-branded.

Nine calculators and five documents — a content brief, the honest way to report a change, the answer to give when somebody asks whether SEO did it, a proposal for running the experiment, and a plain state-of-the-site summary.

In the folder: the application as one offline HTML file (no account, no server, no network request of any kind), a thirteen-sheet companion workbook where every figure is a live formula and a Checks sheet reconciles the book against itself — including whether your two experiment groups actually started from the same place — a Quick Start guide, a full manual, and the licence.