0
$\begingroup$

Yes I face the age-old problem of sales forecasting for a large enough company. Mostly to determine our staffing needs for various functions. So I'm only really interested in aggregate data, not individual product SKU forecasts like most businesses.

Here's the challenge. Traditional time series methods like exponential smoothing, Holt Winters, triple ETS with dampening, do not seem good enough. They're easy enough to understand and implement. However they are entirely backwards looking, of course. The assumption is the system will continue to move, and change, as it has moved, and changed, and that's that.

If you happen to see a brick wall, or rocket fuel, on the horizon, the time series model does not care, or account for this, of course.

So I'm thinking about adding some kind of regression equation, or at least maybe blend a regression equation with the time series (don't feel like anything overly complicated like an ARIMA with regressors which is likely not the right model anyway).

For instance, we're using salesforce and have indicators of quantities x % close rates, that may, in part, help forecast future growth.

Here's my issue with coming up with a model that takes into account historical data, and past seasonality (very strong time-of-year influences) with this future data (sales team picture of leads, lead stage, probability, size).

I can't just add the two together. The time series model >already accounts< for some latent growth in the model. By which I mean, each year, old business may expand in size, and new business may be landed. The model already accounts for this.

It's hard for me to "backwards separate" previous years' 'Salesforce' data -- because that data was never collected/ doesn't exist.

So how do I combine these models? Is it a factor of separating components (level a, trend b, seasonality g) --- and only using the Salesforce data as a modifier to the trend component?

But also, the Salesforce data is only new business, it doesn't account for old business growth --- as a result, should I separate those two items out in the forecast as well?

All this seems like an EXTREMELY common sales forecast procedure that many sizes of companies may have to use, yet everywhere I look online, the forecasting advice is either : a painfully simple model that doesn't even begin to touch on seasonality or any complexity.

Or the opposite: a very statistically rigorous method that either ignores future judgment (time series) or discounts time patterns (regression) ... or overfits the data.

Any help in where to begin to tackle this problem?

To be clear, I have 6 years worth of very granular data (every product, client, category sold) --- but only need an aggregate forecast.

Time series fails because our growth has been slowing year after year to a halt (0%) ... the best-fit model extrapolates this quite well ... but it can never 'pre-empt' systemic changes or future events. It's inadequate. How do I combine regressors/ future judgement into this? Is it less formal than I think?

$\endgroup$
  • $\begingroup$ It is a little unclear I don’t understand all the terminology you are using. It sounds like you need to understand more about the market dynamics in your particular case and find a way to model that explicitly. i.e. you cannot be purely data driven since historical patterns will not necessarily represent those in the future. $\endgroup$ – Zachary Blumenfeld Nov 26 '15 at 6:55
  • $\begingroup$ There is no panacea or cookbook technique to completely solve your problem. Sometimes creating a regression framework out of a supply and demand economic model or a similar utility function can be extremely informative as you can use your knowledge of the industry to construct a framework that is realistic and intuitive to others in your industry. It may be hard to forecast with this model directly, but you could pose hypotheticals, i.e. how much will sales decrease given a 2% increase in the price of mine or my competitor’s product. $\endgroup$ – Zachary Blumenfeld Nov 26 '15 at 6:57
  • $\begingroup$ You likely need data on companies similar to yours to create a panel data set (one where there are multiple units observed over multiple time periods). In combination with your ability to forecast other variables in the market and your specific industry knowledge this type of model can be EXTREMELY useful in anticipating future sales, if done correctly of course. You should research sales forecasting and economic modeling techniques in your specific sector. Also look at fixed/random effect regression and regression with panel data and see if that is something you think will be useful. $\endgroup$ – Zachary Blumenfeld Nov 26 '15 at 7:06
  • $\begingroup$ I'm not sure if the sector in particular is extremely relevant in sales forecasting. I mean, perhaps it is if it's something off the wall, like auto sales, industrial parts, electricity (consumed hourly). This is software sales -- but I've also worked in the CPG industry (food and bev) and other industries where it's just retail/ consumer goods. Sure, the goods have different seasonality and trends, but the fundamentals are the same. You have current product lines and new product lines. You have sales pipelines and historical data. Modeling techniques are needed moreso than industry research. $\endgroup$ – user45867 Nov 30 '15 at 16:28
  • $\begingroup$ There is also benefit in simple models at times, when complex models are either executed improperly or do not give greater accuracy, or perhaps overfit the data. I simply want to use a time series to add in our typical 'seasonality' from historical data, but also want a regressor that accounts for, essentially, the data in our sales pipeline and judgment from our sales team. Pure time series assume the system will stay/ change the same way it always has. I'm really trying for a basic model here. This problem is decades old so I'm surprised that there isn't much academic info on it. $\endgroup$ – user45867 Nov 30 '15 at 16:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.