Most marketing dashboards are decoration. Twelve charts, three colour schemes, a big number at the top that goes up. They are built to look like work rather than to support a decision, and you can usually tell within ten seconds of opening one.
Here is what should be on it, in the order it should appear.
The first screen: four numbers, no scrolling
If you have to scroll to find out how the month is going, the dashboard has failed. Four numbers, above the fold, comparable to last month:
- Enquiries. Forms, calls and messages, counted as one number. This is the business outcome; everything else is an input.
- Cost per enquiry. Total spend divided by that number. The single most useful figure in marketing, and the one most reports omit.
- Where they came from. Organic, paid, direct, referral, social - as a share, not a chart with fourteen slices.
- Trend against the last two months. One month is noise. Three months is a direction.
That is the whole first screen. Everything else lives underneath, for the days you want to dig.
The lead list itself
This is the part most dashboards leave out, and the part owners use most. Not a count - the actual enquiries. Name, when it came in, which channel, which page they landed on, and what they wrote.
Two reasons it matters. First, you can sanity-check the number: if the dashboard says forty-one and the list has twelve real enquiries and twenty-nine spam, you have learned something the chart was hiding. Second, reading what people actually ask for is the cheapest market research available, and it will change your copy faster than any keyword tool.
Second screen: the things you can act on
Which pages produce enquiries
Not which pages get traffic. Which pages produce enquiries. These are usually different lists, and the gap between them tells you where to spend your next hour.
Which search terms bring people who convert
Rankings for terms nobody searches are a vanity exercise. Position four for a term that brings you three jobs a month is worth more than position one for a term that brings curiosity.
Spend by channel, against results
Side by side, in the same view. Spend on its own tells you nothing; results on their own tell you nothing about efficiency.
Site health
Uptime, load speed, broken links. Not exciting, but a slow or broken site quietly undoes everything above it, and nobody notices until the numbers drop.
If your agency's report cannot be checked, it is not a report. It is reassurance.
What should not be on there
- Impressions. A number that only ever goes up and never decides anything.
- Social followers, unless you sell to your followers. Otherwise it is a scoreboard.
- Bounce rate, in isolation. A high bounce rate on a page that answers the question and shows a phone number is fine.
- Average session duration. Longer is not better. People staying longer might mean they cannot find anything.
- Any metric that has never once changed what you did next.
Three tests for the report you get today
- Can you check it? If it says forty-one leads, can you open the list of forty-one? If not, you are being asked to take it on trust.
- Does it show bad months as clearly as good ones? Reports that only ever go up are selected, not measured.
- Is it current? A PDF describing a period that ended three weeks ago cannot change anything. By the time you read it, the month is gone.
Attribution: useful, but do not over-trust it
Every dashboard assigns credit for an enquiry to a channel. It is worth understanding how rough that assignment really is.
Someone sees your Facebook ad on Monday, searches your name on Wednesday, clicks an organic result and calls on Friday. Most tools credit organic search, or direct. The ad that started it gets nothing. Nothing is broken - this is simply what last-click measurement does.
Practical consequences worth holding in mind:
- Brand searches are downstream of something. A rise in people searching your name is usually caused by awareness spend, credited elsewhere.
- Direct traffic is a bucket of unknowns - it includes untagged links, some app referrals and privacy-blocked sources.
- Turning off a channel with poor last-click numbers sometimes drops the channels that looked good. If you suspect that, test by pausing for a fortnight and watching the total, not the split.
Use attribution to spot direction, not to settle arguments to two decimal places.
Which numbers matter at which stage
Months one to three: is the plumbing right?
Enquiry count and source. Do not judge cost per enquiry yet - the sample is too small and the campaigns are still learning. What you are checking is that leads are being captured and recorded at all.
Months four to nine: what is efficient?
Cost per enquiry by channel, and which pages produce leads. This is where you find that one campaign quietly carries the account and another has been costing money for months.
Month ten onward: what is it worth?
Now connect enquiries to closed work. Cost per enquiry is a marketing number; cost per customer is a business number, and they can point in opposite directions. The channel with the cheapest leads is sometimes the one with the worst close rate.
That last step needs your sales data, which means someone has to record outcomes. It is the least glamorous part of measurement and by far the most valuable.
Set the review rhythm deliberately
- Daily - nothing. Daily checking produces reactions to noise.
- Weekly, five minutes - enquiry count and anything obviously broken. Site down, form failing, spend spiking.
- Monthly, thirty minutes - cost per enquiry by channel, page performance, one decision about where next month's budget goes.
- Quarterly, properly - closed business against channel. This is the meeting that should change strategy.
Most people invert this: they check daily and never do the quarterly review, which is the only one that decides anything.
A number you check every day and never act on is a habit, not a measurement.
Questions your dashboard should let you answer in under a minute
- How many enquiries came in this month, and is that up or down on the last two?
- What did each of them cost?
- Which page did the most recent ten land on first?
- Is anything on the site broken or slow right now?
- What is being worked on this month, and what is waiting on me?
That fifth one is not a marketing metric, and it belongs there anyway. Most friction between a business and its agency comes from not knowing what is happening, and a shared board removes it more effectively than any status call.
Building this yourself
You do not need an agency to have most of this. Analytics with conversion events configured properly, Search Console, your ad platforms and a call-tracking number will get you eighty per cent of the way. A free dashboard tool can pull them into one screen.
What you are buying when someone builds it for you is the part that goes wrong on its own: conversion tracking that is actually correct, internal traffic filtered out, spam excluded from the lead count, and someone noticing when a data source silently stops reporting. That last failure is common and almost invisible - a connection expires, a chart flatlines, and nobody questions it for a month.
Monthly PDFs are the wrong format
The monthly report exists for the agency's convenience, not yours. It arrives late, describes a period you can no longer influence, and is assembled by the same people it evaluates.
A live view fixes all three. You look when you want. The numbers are today's. And because the underlying records are one click away, nobody has to trust anybody.
Every WebrixLab client gets this as standard - the same numbers we look at, on your own domain, with the raw source behind every figure. It is not a paid add-on and there is no per-seat charge. Read more about the dashboard portal, or how we handle campaign measurement.