The Problem With The Default Dashboard
Open most analytics accounts and the first screen shows users, sessions, bounce rate and time on page. None of those numbers answers the only question the business actually has, which is whether the website is producing customers.
The metrics are not wrong. They are inputs presented as outcomes. Traffic doubling is meaningless if the new traffic is the wrong audience. A three-minute average time on page can indicate deep engagement or a confused visitor hunting for a phone number. Without a defined outcome, every one of these figures can be read in either direction, which is why they generate discussion but rarely decisions.
Define The Outcome First
Before touching a dashboard, write down what a valuable action is on your site. It is usually one of a short list:
- A completed enquiry form from someone in your actual market
- A booked call
- A purchase, and its value
- A quote request with enough detail to price
- A trial or account signup that later becomes paid
Note the qualifiers. "A completed enquiry form" is not the same as "a completed enquiry form from someone in your actual market". Studios and agencies routinely celebrate a rise in form fills that turns out to be recruitment spam and out-of-scope requests. If your tracking cannot distinguish those, it will reward the wrong changes.
The Metrics Worth A Dashboard
| Metric | Why it earns its place |
|---|---|
| Qualified enquiries per month | The actual output of the site, filtered for relevance |
| Conversion rate by traffic source | Shows which channels bring buyers rather than visitors |
| Cost per qualified lead by channel | The number that decides where budget goes next month |
| Enquiry to customer rate | Distinguishes a traffic problem from a sales or fit problem |
| Revenue per session, for e-commerce | Combines conversion and basket value into one honest figure |
| Landing page conversion rate | Identifies which pages carry the business and which merely exist |
| Branded versus non-branded search | Separates demand you created from demand you captured |
| Assisted conversions | Protects the content that starts journeys from being cut for not closing them |
Why Attribution Stopped Working
Even with the right metrics, the picture is partial, and it is worth understanding why rather than assuming the data is broken.
- Consent refusals remove a share of European visitors from analytics entirely
- Ad and tracker blockers remove more, disproportionately among technical audiences
- Cross-device journeys — researched on a phone, purchased on a laptop — break session-based models
- Dark social — links shared in WhatsApp, Slack and email arrive with no referrer and land in direct traffic
- AI search assistants increasingly answer questions and pass along a visitor with little or no attribution trail
- Long consideration cycles exceed attribution windows entirely in considered B2B purchases
The practical consequence is that "direct" traffic is not a channel. It is the bucket where unattributable visits land, and in many businesses it is the largest and least understood segment in the account.
Building Measurement That Survives This
- Define and instrument conversions properly. Fire on genuine completion, not on button click, or you will count abandoned submissions as wins.
- Capture source into your CRM at the moment of enquiry, so it persists past the analytics window and can be joined to revenue.
- Ask how people found you — a single open field on the enquiry form, or a question in the first call. Imperfect and consistently more informative than the direct bucket.
- Tag every campaign with consistent UTM parameters, including email and social posts you place yourself.
- Use server-side or first-party measurement where paid media spend justifies the engineering.
- Report on a longer cadence. Weekly conversion numbers on a low-volume site are noise; monthly and quarterly trends are signal.
- Reconcile to money at least quarterly — take actual closed revenue and work backwards to source.
The Diagnostic That Matters
When results disappoint, the useful question is where in the sequence the loss occurs, because the fix differs completely at each stage.
- Low traffic, decent conversion rate — a visibility problem. Work on search, content and distribution.
- Good traffic, poor conversion rate — a website problem: unclear proposition, weak calls to action, friction in the form, or slow pages.
- Good conversion rate, poor lead quality — a targeting problem. You are attracting the wrong people, often through content that ranks for the wrong intent.
- Good leads, poor close rate — not a website problem at all. It is pricing, positioning or sales follow-up.
Most businesses assume the first and are actually experiencing the second or third. Buying more traffic to fix a conversion problem is the most expensive mistake in the list, and it is the most common. We work through the specifics of that scenario in traffic without enquiries, and what to fix in order.
A Reporting Setup Worth Having
One page. Monthly. Six numbers, each with the previous month and the same month last year alongside it:
- Qualified enquiries or orders
- Conversion rate, overall and for your top three sources
- Cost per qualified lead where you are spending on media
- Revenue attributable to the site
- Top five landing pages by conversions, not by sessions
- Non-branded organic impressions, as a leading indicator of future demand
If a metric on that page has never once changed a decision, remove it. A short report that gets read beats a comprehensive one that does not.
We set up measurement as part of every build, because a site nobody can evaluate cannot be improved with any confidence. If your analytics are running but you have never trusted what they tell you, we can review the setup and rebuild it around outcomes you care about.