Most business dashboards fail for the same reason: they show everything that can be measured instead of the few things that change what you do on Monday morning. A screen with twenty-four tiles is not insight — it is noise with a colour scheme.
After building operational dashboards for schools, churches, retailers and service firms, we keep arriving at the same short list. Five numbers. If a dashboard shows these clearly, and nothing else competes for attention, the people running the business start making faster and better decisions almost immediately.
1. New leads this week, against last week
Not total leads. Not leads all-time. The comparison is the whole point: a single number has no meaning until it sits next to the number before it.
Track where they came from, too — WhatsApp, referral, website form, walk-in. Within a month you will know which channel deserves your time and which one you have been funding out of habit. Most businesses we work with discover that one channel produces the majority of good leads and receives the minority of their attention.
The trap here is vanity. If you count every enquiry, including the people asking for something you do not sell, the number goes up and nothing improves. Count only leads that match what you actually offer.
2. Conversion rate from lead to paying customer
This is the number that tells you whether your problem is marketing or sales. If leads are plentiful and conversion is poor, more advertising will only produce more disappointment. If conversion is strong and leads are scarce, marketing is the constraint.
Measure it over a rolling window — the last thirty or ninety days — rather than all-time. A business that converted well two years ago and poorly this quarter looks fine on a lifetime average, which is precisely how decline stays invisible.
Break it down by source once you have enough volume. Referrals almost always convert several times better than cold traffic, and that ratio should shape where the next hour of effort goes.
3. Average time to first response
Speed of first reply predicts conversion more reliably than almost anything else you can control. A lead contacted within an hour behaves very differently from one contacted three days later, and the difference is not small.
This number is unusual because it is entirely within your control. You cannot decide how many people find you, but you can decide how quickly someone answers. When a client dashboard starts displaying response time publicly to the team, that figure tends to improve within a week — not because of any new tool, but because it is finally visible.
Set a target, put it on the screen next to the actual value, and let the gap do the work.
4. Revenue against committed cost
Not revenue alone. Revenue next to the fixed obligations that arrive whether or not the month goes well: salaries, rent, subscriptions, loan repayments.
A business owner looking at revenue alone feels good in a strong month and anxious in a weak one, with no clear line between the two states. Revenue against committed cost draws that line. Above it, the month is safe. Below it, you know exactly how far below and how many days remain to close the gap.
This single framing changes behaviour more than any profit-and-loss statement, because it is legible in under two seconds.
5. Work in progress that is stuck
Every business has items sitting still: an invoice unpaid past terms, a project awaiting client feedback, an order held for a missing detail, a lead nobody has touched in two weeks.
Individually each one is small. Collectively they are usually the largest recoverable sum in the business, and they are invisible on a standard report because nothing about them is a failure yet. They are simply waiting.
A dashboard tile that counts items with no activity for more than a set number of days — and lets you click straight through to the list — is often the highest-return element on the whole screen. It converts vague unease into a specific, workable list.
What to leave off
Everything else, at least at first.
Page views, follower counts, total customers since inception, average order value across all time, and any metric that only ever goes up. These belong in a report you read monthly, not on a screen you glance at daily. A number that cannot decrease cannot warn you about anything.
Also leave off metrics nobody owns. If a figure turns red and no specific person is responsible for acting on it, it will be ignored, and once one tile is safely ignorable the whole dashboard loses authority.
Making it real
The technical work is the easy part. The discipline is in the constraint: five tiles, each with a comparison, each with an owner, each clickable through to the underlying records.
Start by writing the five numbers on paper and asking, for each one, "if this changes, what would we do differently?" Any number without a clear answer does not belong on the dashboard. Then check whether your systems can actually produce the number — often they cannot yet, and fixing that data-capture gap is more valuable than the dashboard itself.
Build the smallest version that works, use it for a month, then adjust. Dashboards that grow slowly out of real decisions survive. Dashboards designed in one sitting to look impressive are abandoned within weeks.
If you want help identifying the right five numbers for your operation and wiring them to live data, that is exactly the kind of work we do at ControlPoint Advisory.