The four blog numbers worth reporting, and the ones to drop

Most content reports are designed to be survived rather than read. They lead with a number that went up, contain no number that went down, and answer no question anyone actually had. The reason is usually that the honest numbers move slowly and the report is monthly.

Here is a shorter set that holds up, plus the ones we would stop reporting entirely.

Drop these first

Pageviews as a headline. Traffic is an input, not an outcome, and it is trivially inflatable by publishing more of anything. A blog whose pageviews doubled while its qualified leads stayed flat has got worse, not better, because it now costs more to produce the same result.

Bounce rate. For a blog it is close to meaningless. Someone who arrives, gets the answer, and leaves satisfied is recorded identically to someone who arrives and recoils. If anything, a high bounce rate on a post answering a specific question is a sign it worked.

Average position. Averaged across hundreds of queries, it moves for reasons that have nothing to do with you — a competitor exits, an algorithm update reshuffles a category, seasonal query mix changes. It is a number that always has an explanation and never has a cause.

Time on page, unless you have a specific reason. It mostly measures reading speed and how long a tab sat open.

Number one: posts published against posts planned

The least glamorous metric and the most predictive. It is the only one that is fully within your control, and it is the one that fails first when a programme is about to die.

If you planned four and shipped four, the system works. If you planned four and shipped two, nothing else in the report matters, because you are measuring an activity that is already stopping. We would rather see this number at the top of a report than any traffic figure.

Number two: impressions on non-brand queries

From Search Console, filtered to exclude your own company name. This measures whether the archive is being surfaced for the subjects you chose to cover, rather than by people who already knew you existed.

Impressions rather than clicks, deliberately. Clicks depend on how your title reads in a result, which is a separate problem with a separate fix. Impressions tell you whether you are in the conversation at all.

Expect this to do nothing for about four months, then move. Anyone promising otherwise is either lucky or selling.

Number three: pages that are the entry point to a session

Which posts are the first thing a visitor sees, and how many distinct posts appear in that list. One post carrying everything is fragile — a single ranking change removes your entire channel. Twenty posts each carrying something is a durable asset.

Watching the count of posts that bring in anything at all is a better health signal than watching the total, and it tells you whether your subject list is broad enough or whether you got one lucky hit and stopped.

Number four: assisted conversions, read with suspicion

Whether blog visitors eventually become customers. This is the number the business actually wants and the number that is hardest to trust.

Report it, but report it honestly: attribution windows are arbitrary, self-reported "how did you hear about us" contradicts analytics regularly, and the people your blog influenced most may never have clicked from it. Treat it as directional evidence rather than proof, and be suspicious of any report where this line is precise.

The most reliable version is the unglamorous one. Ask new customers, in a form field, in their own words. It is a small sample and it is worth more than the dashboard.

Reading them together

The four numbers answer four separate questions in sequence: are we still publishing, are we being surfaced, is the archive broad enough, and is any of it converting. A failure at any level makes the levels below it uninterpretable.

That sequence is also the right order to fix things in. There is no point tuning conversion on a blog that stopped publishing in March, and there is no point widening a subject list that nothing is surfacing for yet.