Traffic is vanity, signups are sanity: content metrics that matter

There's a chart that has fooled more business owners than any scam: monthly traffic, up and to the right. It's the chart agencies put on the first slide, the one that makes a content program feel like it's working, and it can climb for a year while revenue from content stays at zero. Plenty of businesses have discovered, eventually, that they'd spent twelve months building an audience of students, job-seekers, and competitors' employees, none of whom would ever buy anything.

The problem isn't that traffic is fake. It's that metrics answer different questions, and most content reporting answers the easy question while leaving the important one untouched. The fix is a sorting exercise: every content metric you'll ever meet belongs to one of three tiers, and each tier has a legitimate job and an illegitimate one.

Tier one: vanity metrics

Pageviews, sessions, users, total traffic, social shares, follower counts, and "we rank #1 for X" trophies. What defines the tier isn't that these numbers are meaningless, it's that they measure activity, not outcomes, and they share three seductive properties: they're easy to get, they're big, and they mostly go up. A growing blog's pageviews rise almost no matter what, which makes them perfect for feeling good and useless for deciding anything.

The tell of a vanity metric is that you can't act on it. If traffic rose 30% last month, what should you do differently? You can't say, because the number contains no information about who came, why, or what happened next. A ranking trophy has the same flaw: position 1 on a query nobody commercial ever types is a decoration. Even genuinely valuable positions need context, since click-through falls off a cliff down the results page: Backlinko's study of four million search results found position 1 takes 27.6% of clicks and the top three together take over half, which is why "we rank #8" and "we get traffic" are very different claims.

The legitimate job of tier one: scale context and morale. It answers "is anyone coming at all?", which matters enormously in the first months of a content program when the honest answer is "barely yet". Its illegitimate job is being the headline that decides budgets.

Tier two: diagnostic metrics

Impressions and average position in Search Console, click-through rate, engagement rate, time on page, scroll depth, pages per session, returning visitors. These measure quality of attention rather than volume of it, and unlike tier one, they're actionable, because each one points at a specific fixable thing.

It's worth knowing what the instruments actually measure, since the definitions are narrower than the names suggest. In Search Console, an impression means your site appeared in results, whether or not anyone saw it; clicks are actual arrivals; average position tracks your topmost result. In GA4, the engagement machinery runs on a precise rule: a session counts as "engaged" if it lasted more than ten seconds, included a key event, or viewed at least two pages, and bounce rate is now simply the percentage of sessions that weren't engaged. Ten seconds is a low bar. A "70% engagement rate" means most visitors gave you at least a glance, not that they read anything.

The legitimate job of tier two is diagnosis, and it's genuinely good at it. Impressions climbing while clicks stay flat: your titles aren't earning the click, or an AI answer is intercepting it. Strong traffic with weak engagement: the page attracts the wrong visitors or disappoints the right ones. High engagement, no conversions: the content works and the next step is missing, no call to action, no path to contact. Tier two is where you look when tier three disappoints, because it tells you *where* the pipeline leaks. These are also your early-warning instruments: in the quiet first months, impressions and position move long before money does.

The illegitimate job: being dressed up as outcomes. "Engagement rose 12%" is a clue, not an achievement. Nobody banks engagement.

Tier three: money metrics

Signups, quote requests, calls, bookings, purchases, and their derivatives: conversions per post, leads per month from organic, revenue attributable to content, cost per lead versus what you're spending. This tier measures the only thing the content program exists for: strangers becoming customers.

GA4's mechanism for this is the key event, which Google defines as "an event that measures an action that's particularly important to the success of your business", its renamed version of conversions. You choose what counts, form submit, call click, signup, and mark it, and from that moment every report can be sliced by the question that matters: which pages produce these? Setting this up takes an afternoon, and it's the afternoon that separates businesses that know from businesses that hope.

What makes tier three different in kind, not just importance: its numbers are small and slow. A post might convert eight visitors a month against three thousand pageviews, and small numbers wobble, one good week doubles them. That's the honest price of measuring what matters, and it's why the tier needs patience and per-post granularity rather than daily refreshing. It's also why vanity metrics stay popular: big numbers are more comfortable than true ones.

The one legitimate criticism of tier three is that it undercounts: content often assists a sale it doesn't get credit for, the reader who returns via a branded search weeks later. That's real, and it argues for measuring assisted paths and asking customers how they found you, not for retreating to pageviews.

Using the tiers together

The tiers aren't ranked by "which to look at", they're a sequence, and the direction you traverse them depends on the job.

For diagnosis, go downhill. Start at money: are conversions where they should be? If not, drop to tier two to find the leak: findability (impressions, position), click-earning (CTR), attention (engagement), or the missing next step. Only then touch tier one for scale context. This order means you always start from the outcome and work backward to the cause, instead of admiring causes all day.

For decisions, only tier three votes. What to write more of, which posts deserve a refresh, whether the program justifies its budget, whether an agency is earning its fee: these get decided by which posts produce customers, with tier two as supporting evidence and tier one as none.

For patience, tier two is the antidote. In months two through six, money metrics are structurally silent, and the choice is between vanity comfort and diagnostic truth. Impressions spreading across more queries and positions drifting up are the honest early signals that the compounding is underway.

The same numbers, read two ways

To feel the difference the hierarchy makes, watch two owners read identical data.

The dashboard says: traffic up 22% this quarter, driven by one post about an industry controversy; engagement rate steady; four leads from organic content, same as last quarter, all landing on two older service-adjacent posts.

Owner A reads tier-one first, which is to say, reads the headline: traffic's up, content is working, write more about controversies. Next quarter's calendar fills with commentary posts. Traffic climbs again; leads stay at four; and a year later the blog has an audience of industry insiders who will never hire anyone and a lead count that never moved, at which point content gets declared "good for brand, bad for sales", which was never true, it was mismanaged into that shape.

Owner B starts at tier three: four leads, from those two posts, again. That consistency is the finding. The controversy post gets correctly filed as reach, pleasant, not the business. The two converting posts get the follow-up questions: what queries feed them, what positions do they hold, what adjacent topics would attract the same kind of visitor? Tier two supplies the answers, one sits at position 6 on a commercial query, worth a push, and next quarter's calendar fills with siblings of what demonstrably works. Leads go to seven. Nothing about the data differed; the reading order did.

The pattern generalizes: tier-one-first reading always steers toward whatever is easiest to grow, because vanity metrics respond to volume. Tier-three-first reading steers toward whatever pays, because money metrics only respond to relevance. Over a year, the two reading orders produce different businesses.

The one-page monthly view

Everything above fits on a single page you can fill in twenty minutes:

• Leads or signups from organic content this month, and which three posts produced them. (Tier three, the headline.)
• Conversion count trend, three-month view, because monthly wobble lies. (Tier three.)
• Impressions, clicks, and average position trend from Search Console, plus any post whose CTR or engagement broke sharply from its normal, with a one-line hypothesis. (Tier two, the diagnosis.)
• Total traffic, one line, for scale. (Tier one, in its place at the bottom.)

If a report can't fit on that page, it's usually because vanity metrics are padding it. If a report leads with traffic, someone is answering the easy question.

Quick answers on the hierarchy

Isn't traffic at least correlated with revenue? Loosely and unreliably, which is the problem: the correlation holds across sites and breaks within one, where a single wrong-crowd post can carry half the traffic and none of the buyers. Correlation you can't act on per-page isn't measurement; it's astrology with a bigger sample.

My agency reports rankings and traffic. Should I push back? Ask one question: "which pages produced inquiries last quarter?" A good agency answers with a list; a report factory answers with a slide about visibility. Either answer tells you what you're paying for, and the ask itself, repeated monthly, reshapes what gets optimized.

What's a good conversion rate for content pages? Whatever your own converting pages already do, benchmarked internally: cross-site "average blog conversion" numbers mix industries, offers, and definitions so freely that they mislead more than they anchor. Your top page's rate is the standard your other pages answer to.

How long before money metrics say anything at all? Months, structurally, on the normal search timeline, which is exactly why the diagnostic tier exists and why judging a young program on tier three alone kills good programs. Judge the phase you're in with the tier built for it.

The one-line version

Sort every content number into vanity (activity), diagnostic (attention), or money (outcomes), and give each tier only its proper job: vanity for scale context, diagnostics for finding leaks and surviving the early quiet, and money metrics, alone, for every decision that costs anything. Traffic is what content does. Customers are what it's for.