Forecasting Capabilities Questions

I have a report that I built that does exactly this kind of forecasting for Attendance. Our big issue is managing sales in GA sections or venues. Since we perform in an outdoor Amp with a large "lawn" section, knowing how much we can oversell before not having enough space for attendees is our biggest management problem.

I do the forecasting based on historical attendance by Price Type Group (which I found to have the smallest Standard Deviation). So, for example, I have price type groups for Comps, Benefit Comps (for donors), Discounts, Packages, Passes, Full Price, etc.. Each of these groups attend at different rates.

I also can split by two other items we track. We use Performance Type to split Classical from Pops performance and have a special type for Movie nights, which are totally different than even pops performances. Being an outdoor venue, I also track Weather, both Predicted the day prior and Observed 2 hours prior to show. These I keep logged in the Content area of a performance (this doesn't pull into Analytics for RAMP clients yet).

Typically, I compare backwards 3 seasons, though I have the ability to do one to one comparisons for a specific concert if there is a performance that is unique or I think will behave like a specific historical performance.

I give my forecast to both our Marketing Director and Box Office manager. The former want to oversell to make budgets, the latter is very concerned with not having enough space for paying customers. They kind of battle it out and figure out how much can be oversold. Because I have one aggressive and one conservative person setting the numbers, the precision does not have to be exact (and it can't be anyway). What I have not done yet is add a "worst case" scenario. I forecast against averages, which include some statistical outliers, but do not display anything like a standard deviation calculation that could be used to predict a range.

After every concert, I do go back and rerun the report against predicted vs actual attendance, just to see how accurate the forecast was. In most cases, the forecast has been pretty close.

This is our specific use case. I'm sure we could come up with more, but for making practical business decisions, this has been the most useful model we have used.

Hi David,

Would you mind sharing a bit more about this forecasting dashboard and how you set it up?

Thanks

Hi David,

Would you mind sharing a bit more about this forecasting dashboard and how you set it up?

Thanks

Ogo - this isn't a dashboard yet, due to the complicated way I'm doing the forecasting. I do it in a report so that I can do the math in the SQL procedure exactly the way I want it to go and compare it to a specific performance that I pick in the Report Parameters. I might someday be able to do it in a Dashboard when I get a little more advanced in the formula building. The problem I have in the dashboard is figuring out how to pick the performances for historical values (that would be a filter), aggregate that data and figure out the attendance percentages based on those past performance (math that might be a bit complicated in Analytics), then multiply those calculated percentages against a specific future performance. This is where I get stumped in Dashboard currently. Not sure how to pick both a group of performances for a prediction and a future performance for analysis. I'm sure there is a way to do it using some cool formulas using the performance date somehow, but I haven't figured out how. Pretty easy to do in a report - in a Dashboard it is a little harder.