DataKnobs

Website analytics

Traffic is the easiest number to celebrate and the least connected to results.

A visit by itself says nothing about whether anyone found what they came for, did anything about it, or was ever going to become a customer. The number worth watching isn't how many people showed up — it's where they stopped going further.

Opening slide of a presentation on website analytics
Title slide from the source deck. It carries its own text, so it opens full size in a new tab rather than being read at page scale.

The shape

Four stages, not one number

Every visitor passes through the same four stages on the way to becoming a customer, whether or not anyone's tracking them: they arrive, they either find what they came for or don't, some of them take an action, and a smaller number still complete the outcome the business actually needs.

A single traffic number collapses all four of those stages into one figure that can rise while every stage after it stays flat. The funnel is what keeps that from happening — it forces the question of exactly where people are stopping, rather than whether enough of them showed up in the first place.

01

Visitors

Raw arrivals, from search, social, direct links, referrals — and increasingly, AI answer engines that summarize or cite a page and send someone through. This last source deserves its own line rather than being folded into "other" or missed entirely; a visitor who arrives after an assistant has already filtered their intent often behaves differently from one who typed a search query cold.

Traffic on its own answers exactly one question: did anyone show up. It answers nothing about whether showing up meant anything.

02

Engaged

Visitors who actually found what they came for, measured by behavior rather than raw page views — time actually spent reading, how far they scrolled, whether they moved to a second page on purpose. A high bounce rate here isn't automatically bad: someone who lands on a contact page, gets the phone number, and leaves in ten seconds had a completely successful visit that still counts as a bounce.

The mistake is reading engagement metrics the same way regardless of what the page is for. A blog post and a pricing page should be judged against different expectations entirely.

03

Leads

Visitors who took a meaningful but intermediate action — filled a contact form, started a signup, added something to a cart. This is the first stage that requires the visitor to actually do something, which makes it a much stronger signal than time-on-page ever is, but it still isn't the outcome the business needs.

Tracking leads without tracking what happens after them is how a funnel can look healthy in a monthly report while the business underneath it isn't.

04

Customers

The outcome that actually matters: a purchase completed, a booking confirmed, a contract signed. Every stage before this one exists to explain how many people reach this point and, more usefully, exactly where the rest of them stopped.

This is the number every other stage should ultimately be justified against. A traffic spike that never moves this number wasn't really progress, whatever the earlier stages looked like.

Signature tool

Find the leak

Start with 1,000 visitors and set a realistic conversion rate for each stage. The bar that turns amber is losing you the most people in absolute terms — not necessarily the lowest percentage, the biggest actual number.

55%
15%
25%
Visitors
1,000
Engaged
0
Leads
0
Customers
0
Biggest leak

Getting it wrong

How teams misread their own numbers

Chasing traffic to fix conversion

More visitors through a leaky funnel produces more people falling out of the same hole, not more customers. The leak has to be found before the traffic spend pays off.

Treating bounce rate as universally bad

A bounce on a page designed to answer one question quickly is a success, not a failure. The right benchmark depends entirely on what the page was for.

Reporting leads as if they were revenue

A rising lead count with a flat customer count means the gap moved downstream, not away. Leadership dashboards that stop at leads hide exactly the number that matters most.

Missing AI-referral traffic entirely

Visitors arriving after an AI answer engine cites a page often go untracked as a distinct source, which means a growing channel's actual value never shows up anywhere.

Questions

Common questions

Why is raw traffic considered a vanity metric?

Because a visit by itself says nothing about whether the visitor found what they needed, took any action, or was ever likely to become a customer. Traffic can rise while every later stage of the funnel stays flat, which means the easiest number to celebrate is often the one least connected to results.

Is a high bounce rate always bad?

No. A visitor who lands on a page, gets what they came for, and leaves within seconds counts as a bounce but had a successful visit. Bounce rate needs to be read against what the page is actually for — a blog post and a checkout page deserve very different expectations.

How does traffic from AI answer engines fit into website analytics?

As an emerging referral source that behaves differently from search and deserves its own tracking rather than being lumped into "other" or missed entirely. Visitors arriving after an assistant has summarized or cited a page often arrive with more specific intent already formed.

What's the difference between a lead and a conversion?

A lead is a meaningful but intermediate action — a form filled, a signup started — that signals real interest without completing the outcome the business needs. A conversion is that final outcome. Tracking leads without conversions can make a funnel look healthier than the business actually is.