System 04 · Data & planning

Tuesday morning is empty and Thursday evening turns people away.

Every appointment business knows this and almost none can say by how much, for which service, with which practitioner, or what it costs. The data exists, in three systems that do not talk. This layer puts it together and turns capacity into a decision with evidence behind it.

Why capacity decisions are made on instinct

Not because operators are careless. Because the numbers needed to decide sit in systems that were never designed to be read together.

The data is in three places

Appointments in the scheduler, calls in the phone system, purchases at the till, and none of them share a client identifier. Every question that crosses two systems requires someone to build a spreadsheet, so those questions get asked once a year at most.

Aggregates hide the problem

Monthly utilization at seventy percent sounds healthy and can conceal a Tuesday at thirty and a Thursday at ninety-five with people turned away. The average is the one number that guarantees you cannot see either.

Demand that never became a booking is missing entirely

Utilization measured against bookings tells you how full you were. It cannot tell you how full you could have been, because the enquiry nobody answered left no record. Half the input to a staffing decision is structurally absent.

The forecast nobody trusts

Someone produced a projection once, it was wrong, and it has been ignored since. Usually because it was presented as a single confident number instead of a range with its assumptions written down.

What the system actually does

Four steps. The first is the longest and the least interesting, and skipping it is why most analytics projects produce nothing.

1

Reconciles the sources

Brings appointments, calls, cancellations, no-shows and transactions into one view with a consistent client identity. Where two systems disagree, the disagreement is surfaced as a finding rather than silently resolved, because it usually points at an undocumented process.

2

Measures capacity honestly

Reports utilization by hour, by day, by practitioner, by room and by service, next to demand rather than in isolation. An empty hour with unanswered enquiries and an empty hour nobody wanted are different problems, and only the combined view distinguishes them.

3

Models what changes if you move something

What happens to utilization and to retention if opening hours shift, if a practitioner moves day, if a service is offered at a different time. The assumptions are visible and yours to argue with, which is the point.

4

Forecasts with the uncertainty stated

Where history supports it, forward demand by period and service with a range rather than a single figure. Where history does not support it, the system says so instead of producing a confident number someone will staff against.

Runs on the integration built for intake, retention and conversation analysis. Where those exist, the data layer is largely already in place.

What we will not do with your numbers

This is the layer where it is easiest to produce something that looks authoritative and means nothing.

No confident forecast on thin history

A projection from six months of a business that changed in month four is a guess with a decimal point. We will tell you when that is the situation.

No algorithmic pricing presented as truth

You get the demand curve, the observable elasticity and the retention effect. The pricing decision stays with you, with the assumptions in the open.

No dashboard with forty metrics

Output is built around decisions you actually make. Anything that would not change an action is left out, because the fate of most analytics is being opened twice and then never again.

No attribution we cannot defend

If utilization moved in a period when a campaign ran and a competitor closed, that goes in the report instead of a clean percentage attributed to us.

How it connects, and how it is measured

What your systems expose decides the work. Success is whether a decision got made differently.

Scheduler, phone system, point of sale

Through APIs where they exist and exports where they do not. Reconciliation quality depends on whether client identity crosses systems, which is established during scoping.

Into your BI, not beside it

Where you already have a reporting environment, output goes there. Adding a second place to look is a good way to guarantee neither gets read.

Where the data lives

Your infrastructure, in the region you specify. Your decision, recorded at contract. After handover it is your database, not a copy of it inside a vendor.

Judged by decisions, not dashboards

The measure is whether opening hours, staffing or timetabling changed and what happened afterwards, compared against the baseline recorded before the work started.

What we need from you

Access to the scheduler and phone systemAPIs or exports. Client identity crossing the two is what determines how much can be answered.
Transaction data where it existsPoint of sale or billing, if purchase behaviour is part of the question.
Enough historyTwelve months is comfortable. Less is workable for utilization and unreliable for forecasting, and we will tell you which of those you are in.
The decision you need to makeStaffing, opening hours, timetabling, pricing or capacity investment. Analysis without a decision attached produces a report nobody opens.
Someone who can act on itA person with authority to change the schedule. Findings with no owner change nothing.

Questions we get asked

We already have reports in our booking software. Why is this different?Because the reports in a booking system describe what the booking system knows, which is appointments that happened. They cannot see demand that never became an appointment, contact that went unanswered, or the relationship between the two. Utilization measured only against bookings tells you how full you were, not how full you could have been, and those are different businesses to manage.
What is the difference between utilization and demand?Utilization is the share of your capacity that was used. Demand is how many people wanted it. A Tuesday at forty percent utilization might be a Tuesday nobody wants or a Tuesday where three enquiries went unanswered at nine in the morning, and the decision you make is opposite in each case. Most businesses only measure the first and then make staffing decisions as though they had measured the second.
Can this tell us what to charge at off-peak?It gives you the demand curve, the elasticity you can observe from your own history, and what happened to retention when clients moved to quieter slots. Pricing is your decision, and anyone presenting an algorithmic price as an answer is hiding assumptions inside it. The system is meant to make the assumptions visible rather than to make the decision for you.
Our data is messy. Three systems, none of them agree.That is the normal starting position rather than an obstacle, and the reconciliation is most of the value. Where two systems disagree about the same appointment, that disagreement is itself a finding, usually about a process nobody documented. Cleaning it is unglamorous and it is the reason the first useful output takes weeks rather than days.
Will it forecast demand?Where there is enough history and the pattern is stable, yes, with the uncertainty stated rather than hidden. Where history is thin or the business changed recently, a forecast is a guess with a decimal point and we will say so. A confident number on insufficient data is worse than no number, because someone will staff against it.
Does this replace our accountant or our BI tool?No. It answers operational questions about capacity, demand and client behaviour. Financial reporting stays where it is. Where you already have a BI environment, output goes into it rather than becoming another dashboard nobody has time to check.
What is the first thing this usually finds?In most businesses, that capacity is much more unevenly used than anyone believed, and that the imbalance is by hour and by practitioner rather than by day. The second finding is usually that a meaningful share of enquiries arrives in hours nobody is staffed for, which is a different problem with a different fix.
How long does it take?A first useful output typically takes two to six weeks, dominated by access and reconciliation rather than analysis. The number that matters is how quickly you can act on it, which is why the work starts from a decision you need to make rather than from the data you happen to have.
Let's scope it

Start with the hours you cannot currently account for.

Answer five questions and the brief writes itself, or write to us directly. Either way you get an initial solution brief. A proposed architecture and a fixed quote follow a scoping call, once we have seen your systems.

Build a brief → Contact form