Measurement and reporting

Most SEO that gets abandoned is abandoned in the window between the work landing and the traffic arriving. Measuring the right things is how you survive that window without guessing, and measuring the wrong things is how people talk themselves out of something that was working.

The metrics move in a fixed order

This is the single most useful idea in the module. Effects appear in sequence, and knowing the sequence tells you whether you are early or wrong.

1. Impressions rise first. You appear in more searches. Nobody has clicked yet. Google is testing you.
2. Average position improves. Page four becomes page two. Still very little traffic, but the trajectory is the signal.
3. Clicks follow. Once you cross into the top handful of positions, the impressions you already built start converting.
4. Conversions last. And only if the traffic is the right traffic, which is a module 3 question.

So if you are three months in with impressions climbing and clicks flat, that is the expected picture, not a failure. And if impressions are flat after three months of publishing, that is a real problem worth investigating now rather than at month nine.

The mechanism behind that lag is the same one I wrote about in the compounding effect, and measurement is how you watch it happen instead of hoping.

Separate branded from non-branded or the report lies

Searches for your company name are not SEO performance. They track awareness, which your other marketing drives. Leaving them in the same number as everything else produces two specific failures.

A brand campaign makes your SEO look like it is working when non-branded terms are flat. And a decline in the terms you are actually working on gets hidden by branded volume, so you find out late.

Search Console lets you filter queries containing your brand name. Do it once, save both views, and report them separately forever.

Have us set the reporting up once, properly

Seasonality is the thing that fools everyone

Almost every business has a seasonal shape and almost every self-made SEO report compares this month against last month, which is the comparison most contaminated by it.

Compare against the same month last year as well. If you do not have last year’s data, say so in the report rather than drawing a conclusion the data cannot support. A retailer celebrating a November rise, or panicking about a July drop, is usually just describing their industry.

What not to measure

Third-party authority scores. Domain authority metrics are invented by tool vendors, not by Google. Useful as a rough comparator between sites, meaningless as a KPI to move, and watching one go from 22 to 24 tells you nothing about your business.

Total keyword count. Ranking for 1,400 terms instead of 1,100 sounds like progress and is usually noise from long-tail variants of pages you already had.

Average position across the whole site. An average that mixes your money pages with your privacy policy is a number about nothing.

Bounce rate, as a ranking proxy. Someone finding their answer in the first screen and leaving is a success. Treating it as a failure leads you to pad pages, which is the opposite of module 6.

What I would do in your position

One page. Five numbers: impressions, average position, non-branded clicks, indexation rate, and conversions from organic. Same day every month, same 28-day window, plus one sentence on what you shipped.

Then compare against the payback month you wrote down in module 1, not against zero and not against last month. That comparison is the entire reason you wrote the number down, and it is what turns a flat month from a crisis into an expected data point.