Demand Forecasting for Live Events: How to Know Eight Weeks Out Whether a Show Will Sell Out

Demand forecasting for live events answers one question early enough to act on it: will this show sell out, and if not, by how much will it miss? Most promoters get that answer one week before doors, when the only options left are a discount, a paper house or an empty block of seats. The sales curve usually knew eight weeks earlier. This article explains how forecasting works in ticketing, what a forecast reads, what it returns, and how it changes the decisions a promoter makes between announcement and show night.

What demand forecasting means in ticketing

In supply chain or retail, demand forecasting predicts how many units will sell in a period. In live entertainment, the unit is a seat and the period is fixed: the show happens on a date whether the seat is sold or not. A ticket that is unsold at doors is revenue that never comes back. That makes forecasting in this category less about inventory and more about time. The question is not only how many tickets will sell, but when, and whether the current pace is enough.

A demand forecast for a show typically returns three things:

Sellout probability. The likelihood, based on the current sales curve and comparable shows, that the show reaches capacity by the event date.
Projected final sales. The expected number of tickets sold at doors if nothing changes.
Budget-to-ticket ratio. How many additional tickets a given amount of marketing spend is likely to move, given where the show sits on its curve.

The third output is what makes a forecast operational rather than academic. It tells the promoter not only that a show is weak, but what it would cost to fix it and whether the money is better spent elsewhere.

What a forecast reads

A forecast is only as good as the signals behind it. For a live event, the relevant signals fall into three groups.

The show's own sales curve. The announcement spike, the presale, the first 72 hours of general on-sale and the pace since. Sales velocity in the first days after on-sale is the single strongest predictor of final sales. A show that sells 30 percent in the first week and a show that sells 8 percent in the first week rarely end in the same place.

The show's attributes. Artist or production type, genre, venue capacity, market, price level, day of the week, time of year, and how many competing events sit in the same city in the same window. Two shows with identical first-week sales can have very different outcomes if one is a Friday in a 3,000-capacity room and the other a Tuesday in a 10,000-capacity arena.

Comparable history. Thousands of shows with similar attributes and their complete curves from announcement to doors. This is what turns a single show's pace into a probability. Without comparables, a promoter can see that sales are slow. With comparables, they can see how often shows that looked like this one at week four ended up selling out, and what the ones that did had in common.

NYBA's forecasting model scores every show against thousands of comparable campaigns from twelve years of ticket demand data across 25 markets: 75 million tickets analyzed at ticket level, plus the audience, creative and channel performance of every campaign that ran alongside them. Every new event feeds back into the model and sharpens the next forecast.

The three shapes of a sales curve

Nearly every show follows one of three curve shapes. Recognizing the shape early is half the forecast.

Front-loaded. A large share of tickets sells in the first days, then the curve flattens. Typical for major artists with strong fan bases, reunion tours, limited runs. The risk here is not weak demand but wasted spend: marketing money poured into a show that was going to sell out anyway.

Steady build. Sales grow at a roughly constant rate from on-sale to doors. Typical for established mid-size acts, long-running theatre productions, family entertainment. These shows respond well to campaigns because there is a real audience still deciding.

Late surge. Slow sales for most of the cycle, then a rush in the final two to three weeks. Typical for local shows, festivals in their early years, and events with a spontaneous audience. The danger is misreading the shape: a late-surge show looks like a failing show at week six, and a failing show looks like a late-surge show that never surges.

Distinguishing the last two cases is exactly what comparables are for. A model that has seen thousands of late-surge shows knows what their week-six curve looks like, and how it differs from a show that is simply not selling.

The cost of guessing, in one show

Take a show with a capacity of 10,000. At week eight, 4,500 tickets are sold. The promoter's instinct says it will pick up. The comparables say shows with this profile at this point end at around 6,000.

If nothing changes: 6,000 sold, 4,000 seats empty. At an average price of 80 euros, that is 320,000 euros of revenue that never comes back. The show still happens, the artist still gets paid, the venue still charges rent. Only the revenue is gone.

If the forecast is on the table at week eight, the promoter has options. A campaign that moves 2,000 additional tickets at a cost that is a fraction of 160,000 euros in ticket revenue is an easy decision. At week one, the same 2,000 tickets are far more expensive to move, if they can be moved at all.

That is the whole argument for forecasting: it does not sell tickets, it buys time.

What changes once the forecast is on the table

A forecast changes four decisions a promoter makes on every show.

Where the budget goes. Without a forecast, budgets are set by habit: the same split for every tour stop, the same amount for every show. With a forecast, budget moves off the shows that will sell out anyway and into the ones that need the push. Cutting spend on a safe sellout is as valuable as adding spend to a weak show.

When the campaign starts. A steady-build show benefits from an early, sustained campaign. A late-surge show may need most of its budget in the last three weeks. A front-loaded show may need almost nothing after the first week.

Which shows get attention. On a tour with thirty stops, the promoter's attention is the scarcest resource. A forecast ranks the stops by risk and lets the team work on the five that matter instead of scanning thirty dashboards.

Whether to add a show. When the sellout probability of the first date crosses a threshold early, the forecast is the first signal to open a second date while demand is hot, rather than after the first date has sold out and the momentum is gone.

How NYBA runs forecasting inside the campaign

In NYBA OS, forecasting is not a report that lands in an inbox. It is the first step of every campaign. The platform connects to the ticketing system, scores the show against its comparables, and returns a sellout probability and a budget-to-ticket ratio before the first ad goes live. From there, campaigns run on every channel that sells tickets and keep optimizing against real sales from the ticketing system until the show is full.

When the curve moves away from the forecast in either direction, budget moves with it. A show tracking ahead gets less; a show falling behind gets more, earlier. The promoter sees both the forecast and the verified sales in one place and keeps the final say on every decision.

Frequently asked questions

How early can a show's outcome be forecast?
A first usable forecast is possible from the announcement and presale signal, and it becomes reliable within the first week of general on-sale. The first 72 hours of sales carry most of the predictive weight.

Does demand forecasting work for small venues and regional promoters?
Yes, provided the show has comparables. A 1,500-capacity club show in a regional market has thousands of comparable shows in a large enough dataset. What matters is that the model has seen shows like it, not that the show is big.

What data does a promoter need to provide?
Access to the ticketing system's sales data and the basic attributes of the show: artist or production, venue, capacity, price levels, on-sale dates. Historical sales from previous shows improve the forecast but are not required.

Can a forecast be wrong?
Yes. A forecast is a probability, not a guarantee, and unusual events (a viral moment, a cancellation elsewhere, a weather event) move demand in ways no model predicts. The value of a forecast lies in being right far more often than a gut feeling, and in being available eight weeks before the gut feeling kicks in.

What is the difference between demand forecasting and dynamic pricing?
Forecasting predicts how many tickets will sell and when. Dynamic pricing changes the price in response to demand. They can work together, but forecasting is useful regardless of whether prices ever change, because its main output is a marketing and budget decision, not a price decision.

Eight weeks of warning, not one

Every promoter has a story about a show that looked fine until it did not. The sales curve knew. Demand forecasting is the discipline of reading that curve early, against enough comparables to trust the reading, and moving budget before the show tips.

To see what a forecast looks like for your next on-sale, book a demo. One event, clean numbers, then you decide.