Building a marketing dashboard the owner actually reads

Before opening Google Analytics, write down the three or four questions the owner actually asks. In most businesses they are some version of:

• Are we getting more of the right leads or customers than last month, and than this time last year?
• What is it costing us to get each one?
• Which channels are earning their money and which are not?
• Is there anything I need to decide this month?

Every metric on the top-level dashboard should answer one of those. If a number does not help answer them, it goes to the second layer or nowhere. Page views, social followers and email open rates are useful diagnostics but almost never answer an owner's question on their own.

The exact set depends on the business model, but a good dashboard reads like a funnel from left to right.

Six rows, each compared against the previous period and the same period last year. For a service business, replace orders with qualified leads and add a lead-to-customer rate from the CRM. For a retailer, add revenue by channel. For a SaaS or subscription company, add trials, paid conversions and churn.

The right tool is the one that gets maintained. In rough order of complexity:

Google Analytics custom dashboards. Free, built in, up to 12 widgets, shareable by email on a schedule. Excellent for traffic and conversion data, useless for anything outside Analytics such as ad spend or CRM revenue. Good first step; most small businesses should start here and discover its limits before paying for anything.

A spreadsheet. Excel or Google Spreadsheets, with one tab per month and a summary tab that charts the trend. It handles any source because a person types the numbers in, which is both the strength and the weakness. The Google Analytics add-on for Google Spreadsheets, released last month, can now pull Analytics data automatically, which removes the most tedious copying. For a business with one or two paid channels and a simple CRM, a spreadsheet maintained by one named person on a fixed day beats most software.

Hosted dashboard tools. Cyfe (free tier, around $19 a month paid), Geckoboard (from around $20 to $150 a month depending on dashboards), Klipfolio (from about $20 a month) and Ducksboard connect directly to Analytics, AdWords, Facebook, MailChimp, Salesforce and dozens of other services and display live widgets. They make sense when you have four or more data sources and someone is spending two or three hours a week assembling numbers by hand. The trade-off is that connectors break, custom metrics need a developer, and a live screen encourages daily peeking.

Marketing platforms with built-in reporting. HubSpot, Marketo and Pardot report on the funnel they manage, which is excellent if the whole funnel runs through them and misleading if it does not. Most mid-sized businesses have revenue data the marketing platform never sees.

Business intelligence tools. Tableau and QlikView are powerful and priced for it, with licenses in the low thousands of dollars per user. Overkill for most marketing teams.

A practical decision rule: if the dashboard can be built in a spreadsheet and maintained in under an hour a week, build it there. Move to a hosted tool when that hour becomes three. Move to BI when a finance or data person owns the data model.

The dashboard is a document, not a screen. The thing that makes owners read it is the ritual around it.

1. Pick a fixed day. Monthly on the third business day works for most businesses, because ad platforms and e-commerce reports have settled and finance has not yet closed the books. Weekly on Monday morning for the marketing team's own working version.
2. Send it, do not just publish it. An email with the dashboard attached or linked, every time, from the same person.
3. Write the note. Three short paragraphs: what changed and why, what we are doing about it, what we need a decision on. This is the part that gets read; the chart supports it. A dashboard without commentary asks the owner to do the analyst's job.
4. Annotate the chart. Mark the month the new site launched, the week the AdWords budget doubled, the day the newsletter went out.
5. Review the dashboard itself quarterly. Remove metrics nobody has mentioned in three months. Add one only if it answers a question the owner has started asking.

A dashboard is only trusted once. If the owner catches a number that does not match the bank account, every future number is doubted. The common causes:

Analytics revenue versus actual revenue. E-commerce tracking misses refunds, phone orders and some payment-gateway returns. Report the Analytics figure as "tracked online revenue" and reconcile it against the cart or accounting system monthly, showing both.

Channel attribution. Analytics uses last-click by default, which gives paid search and email credit that organic and social partly earned. Note that on the dashboard rather than switching models every month.

Untagged campaigns. Email and social links without UTM parameters show up as direct or referral traffic and make those channels look better than they are. Set a tagging convention and check it monthly.

Date range drift. Comparing a 31-day month to a 28-day month, or this week to a holiday week, generates alarm about nothing. Compare like with like and say so.

Spend from invoices, not platforms. Agency fees, tools, and creative costs belong in cost per acquisition. A dashboard that counts only media spend understates cost by a third or more.

Building for the marketing team and sending it to the owner. Twenty metrics on bounce rate and social reach answer the team's questions, not the owner's. Build two layers.

Showing numbers with no comparison. 4,200 sessions means nothing. 4,200 sessions, up 12 percent on last month and 30 percent on last year, means something.

Automating before standardizing. Connecting six data sources to a live dashboard before agreeing on definitions produces six versions of "a lead."

No owner. If maintaining the dashboard is everyone's job, it will be nobody's within a quarter.

Reporting is part of how we run growth work, not an add-on at the end. In discover we interview the owner and the team about the decisions they make and the questions they cannot currently answer, audit what is tracked today, and find the gaps between analytics, ad platforms and the systems that hold real revenue. In plan we define the five to eight top-level metrics with written definitions, choose the tool level that fits the data and the team, and set the cadence. In build our technology team fixes the tracking (goals, e-commerce, UTM conventions, CRM fields), connects the sources or builds the spreadsheet, and produces the first report with its written note. In improve we run the monthly review with the client, prune and adjust, and use what the dashboard shows to steer the brand, web and performance work.

Running brand, growth, web and technology together matters here because the dashboard is only as good as the tracking in the website, the campaign tagging and the systems integration behind it. When one team owns all three, the numbers reconcile.

A representative engagement was a multi-location services business where the owner received four separate reports a month, from the web agency, the paid search vendor, the email tool and the CRM, and none of them agreed. We mapped the questions the owner was actually asking, standardized what counted as a lead across every source, fixed campaign tagging and CRM capture, and consolidated everything into a single monthly report with a one-page written note. The report took the team under an hour to produce, and within two quarters it had reshaped how budget was allocated between channels. You can read more about that kind of work in our transformation case study.