iSixSigma

Measuring Sales Forecasts Accurately

Six Sigma – iSixSigma Forums Operations Marketing and Sales Measuring Sales Forecasts Accurately

Viewing 4 posts - 1 through 4 (of 4 total)
  • Author
    Posts
  • #54074

    Anand Chitanand
    Participant

    Measuring the sales forecast accuracy is a big challenge. It is also a bigger challenge to convince the other side of the table, how a particular method is a better method.
    Suggest me some good methods which would be acceptable in an engineering goods company.

    0
    #193394

    Gary Cone
    Participant

    I can’t imagine that it’s really a challenge. What was predicted, what was accomplished. What else would you want to do?

    0
    #193400

    Rob Fioto
    Member

    Harvard and other universities conducted a series of surveys on sales metrics. One of their findings was that many organizations have no clue as to what they are measuring. Regression analysis on the entire scorecard processes they sampled indicated that organizations would be better served by abandoning their current metrics and just guessing at a number.

    Forecast accuracy is the first of a series of important steps. If your prediction methodology is hosed, so is your sales strategy, BD investment and financial strategy. Think of it as a quick surrogate for an MSA on your sales process. Many companies waste millions of dollars lying to themselves with useless sales forecasts.

    Harvard published a series of case studies on this topic:
    – Delusions of Success
    – What you don’t know about Decision Making

    Also, another scary number published was the cost per metric. Since most organizations rely on the roll-up of manually-collected “lagging-indicator” metric data (which is often too late to be useful), many metrics can cost an organization up to $10K per metric cycle.

    0
    #193408

    Anand Chitanand
    Participant

    Hi Rob,

    Thanks for that very good information.

    As the prediction methodology is so rude, measuring the accuracy with one data point at a time becomes ridiculous. Measuring a series of data points of forecast vs. actual (i.e. the error) and then calculating Mean absolute percentage error could be a method of measurement. But I am not sure if there are any better means / methods.

    I agree with you on the cost metrics.

    0
Viewing 4 posts - 1 through 4 (of 4 total)

You must be logged in to reply to this topic.