To actually measure whether an SEO change worked, you define the metric before you touch anything, isolate the change so nothing else confounds it, compare against a baseline plus a control of similar untouched pages, and wait a real re-crawl-and-settling window before judging. “Watch your rankings go up” is not measurement; it is hoping, because rankings drift for dozens of reasons that have nothing to do with your edit. The discipline that turns an edit into evidence is before-and-after with a control, and most “track your results” advice skips every hard part of it.
Start by deciding what success means and recording the starting point. Pick the specific metric the change is supposed to move, average position for the target query, clicks to the page, click-through rate, indexed status, whatever fits the goal, and capture its current value for the affected pages. Without that baseline you have nothing to subtract from later, and you will be tempted to read random fluctuation as a result. Write the number down before you change anything, along with the date, so the comparison is honest rather than reconstructed from memory.
Next, isolate and control. Change one thing at a time, because if you rewrite the title, add internal links, and update the copy in the same week, no later data can tell you which one moved the needle. Then choose a set of comparable pages you are not changing and watch them alongside the edited ones. If the whole set rises together, the cause is probably sitewide or external, not your edit; if only the changed pages move while the control holds flat, your change is the likely driver. That control is what separates a real signal from the background noise of seasonality, algorithm updates, and competitor activity.
Finally, give it time before you call it. Google has to re-crawl, re-index, and let the result settle, so judging after a day or two reads mostly noise. Set a window appropriate to how often the pages are crawled and how much traffic they get, then compare baseline to post-change against the control at the end of that window. So in practice: name the metric and record the baseline first, change one variable, pick a control set, wait the window, then compare. Decide the measurement plan before you make the change, not after you are wondering what happened.