A seasonal demand index answers one narrow question: in this period, is demand above or below the average for the year, and by how much. Disagreements about seasonality are usually disagreements about which average the numbers were measured against.
What the index actually measures
The calculation compares booking volume in a period with the average booking volume per period across a base window. The output is a ratio, usually scaled so that one hundred means exactly average. A period at one hundred and ten carries more demand than the average period; one at ninety carries less.
That is all it says. The ratio is blind to revenue, to capacity, and to whether the busy period was busy with full-price work or with promotions. If the question is whether to hire in July, the index is a starting point rather than an answer.
The three inputs
Every version of this calculation needs the same three things, and each one causes a different kind of error when it is sloppy.
The first is a count of bookings per period, taken from the booking system rather than from memory or from a payment export. Bookings and payments diverge when a client prepays, when a package is redeemed, and when a no-show is charged.
The second is the period definition, written down before the arithmetic starts. Months are the usual choice. Weeks suit a strong day-of-week pattern but distort around holiday closures.
The third is the base window: the span of history the average is drawn from. This is a decision rather than a fact, and it belongs in the notes.
Building it in order
- Export bookings for the full base window with a date on every row.
- Assign each booking to a period using the date it was booked for, not the date it was created.
- Count bookings per period and write the counts in a single column.
- Divide the total by the number of periods to get the average period.
- Divide each period's count by that average, then multiply by one hundred for readability.
- Write the counts beside every index value so a reader can check the division.
Counting first and dividing last keeps the intermediate numbers visible. That is the difference between a figure somebody can reproduce and a figure that has to be taken on trust.
Choosing a base period
A trailing twelve months is easy to explain and sensitive to whatever happened in those months. A three-year base smooths a single strange season but blurs a real shift in the business, such as a move to a larger space or a change in the service menu.
Whichever is chosen, it must be stated in the same place as the result. The same index against a quiet base and against a strong base are different claims wearing one number.
Exclusions that bend the curve
Three kinds of distortion are worth correcting for, and all three should be recorded rather than silently applied.
Closures come first. A week the salon was shut is not a week of low demand, and leaving it in drags the average down and inflates every other period.
One-off events come second: a renovation, a promotion that ran once, a competitor that closed and sent a wave of callers your way. Correcting for these is judgement, so note what was adjusted and why.
Service mix changes come third. A shift toward longer appointments reduces the number of bookings while leaving the calendar just as full, which makes demand look softer than it is. When the mix changes, count appointments and hours booked side by side.
Reading it without over-reading it
Compare direction across periods, not decimals. A move of a point or two is noise in a small business, and reporting to a decimal place implies a precision the counts do not support.
Read three years beside each other where possible. A peak that appears in the same period each year is seasonality. A peak that appears once is an event, and building a staffing plan around it is how a business ends up overstaffed for a season that never returns.
Finally, put the index next to capacity. A month at one hundred and ten is only a problem if the calendar was already close to full. The same index against a half-empty book is an opportunity.
When the answer points to coverage rather than marketing, the options are narrow: voicemail records the call without booking it, an answering service puts a person on during bought hours, and Elevoi answers around the clock, books appointments, captures caller details, and escalates urgent calls to a person. Before the first run, work through the seasonal demand patterns checklist so the counts are clean.