A One-Page Dashboard for a One-Person Business

Build one page that shows the few numbers you steer by, each with its comparison and its action threshold. For most small businesses that is: sessions, orders or enquiries, conversion rate, the funnel step rates, the device split of the key rate, revenue or revenue per session where mix matters, and one quality guard such as refunds, unsubscribes or unqualified enquiries. Every number carries last period beside it, because a number without a comparison is a fact without a meaning. If a number has not changed a decision in a quarter, remove it, however impressive it looks. The dashboard's job is not to describe the business; it is to interrupt you when something deserves attention.

For each candidate number, complete the sentence: when this moves by this much, I do this. Step rate falls notably: I watch recordings on that step this week. Source volume collapses: I check the campaign and the tracking before anything else. Mobile rate diverges from desktop: the next fix slot goes to the mobile step. Numbers that cannot finish the sentence, total page views, follower counts, a wall of product metrics, are reports, not dashboard material, and they belong in the monthly review in the five reports or nowhere. This test is ruthless and small on purpose; a dashboard of eight numbers with actions beats a console of forty with none.

Lay the page out in the order of a check-up: traffic, funnel steps in sequence, outcome, guard. Show each number with its previous-period twin and a simple direction, avoiding charts that need interpretation at a glance; the deep charts live one click away in the funnel view from funnel reports. Include the period plainly, whole weeks or the last full month, and never a rolling window that changes shape daily and invites nervous checking. Resist live data entirely. A one-person business does not need a real-time pulse; it needs a weekly rhythm, and real-time numbers mainly manufacture urgency about noise.

The dashboard can be a saved analytics view, a free reporting canvas connected to your analytics, or a humble spreadsheet updated at the weekly sitting. The medium matters less than the stability: same numbers, same definitions, same day each week. Automate the pull where the tool does it reliably, but keep a manual glance at one source number against the platform's own count, because automated pipelines break silently and a dashboard confidently displaying stale data is worse than a spreadsheet honestly updated by hand. Date the page visibly. If the date is old, the dashboard has failed at its first job, which is to be current enough to trust.

Small numbers swing, so set attention thresholds from your own history: a step rate outside its recent normal range, a source down by an amount a quiet week cannot explain, a guard metric moving twice in a row. Note known seasonal shapes beside the numbers, sale weeks, school holidays, your trade's quiet month, so the dashboard reader, future you, does not launch an investigation into August being August. Where the business is young and history thin, thresholds start wide and tighten as baselines accumulate. A dashboard that alarms monthly about nothing trains its owner to ignore it, which is the one outcome worse than having no dashboard at all.

The dashboard opens the weekly thirty minutes described in from data to decision. Glance in layout order. Anything inside its normal range gets no time. Anything outside earns the next layer: the funnel report for location, recordings for appearance, survey ranking for motive. End by writing the week's one decision on the dashboard itself or its companion log, with the number that triggered it. That written line converts the dashboard from a display into a steering record, and across months those lines become the fastest possible review of what the business noticed, decided and learned, which is also the raw material of the testing archive on the neighbouring bench.

Review the page quarterly with the same strike test: which numbers triggered an action, which were merely read, which were not even that? Promote a report number onto the dashboard only when a recurring decision needs it weekly, and demote anything that has been quiet and unactioned for two quarters. Keep a note of removals, because a number that disappears can make a later comparison confusing. The dashboard should shrink as readily as it grows. A page that only grows becomes the overwhelming analytics it was built to replace, one well-intentioned addition at a time, and its owner quietly returns to steering by feel while the console glows unread.