Free Experiment Analysis Service

Process Builder and Math Options are jointly offering a free design check service.  What's the catch?  Well, we are hoping to sell you something.   Read on.

If we see that your design is flawed (incapable of reliably predicting the best possible answer for the least possible cost), we'll tell you how we know that.  We will tell what is wrong and why it will not work.  That's free, as is telling you that your design passes our tests.  We'll send you the complete report in any event including all of the statistics and what they mean.

If your design is flawed, we'll also figure out exactly what you need to do to fix the design.  For an all too reasonable fee, we'll tell you that as well.

If you have a good design but still didn't get the results you were looking for, we'll take a real hard look with all of the tools, techniques, and methods at our disposal.   If we find that your experiment is fine and the answer is just not there, you haven't lost anything.  We'll tell you exactly that. 

If we do find a set of results that meet your goals, we'll even tell you part of that answer for free.   Which part?  The response values, of course.  (You didn't think we were going to give you the factor settings, did you?)  Of course, if you want the factor settings, we will be glad to part with them for a little cash.  Naturally, there is a money-back guarantee on the settings, and they will come along with an explicit explanation of how we arrived at the answer.

Now we're not quite so optimistic that either of those items will make a lot of money, but we think we might break even selling a few design repairs and analysis tricks.   Where we think we'll really make a killing is after we have done a couple for you.   We know that you are going to say: "I'm at least as smart as these guys.   Why don't I go to their training and buy their software?"  Why not indeed!  It just so happens we're in the business of selling training and software.

You have 2 Options:

Option 1:

Check the design before you run the experiment!

If you send us your experiment design before you make any runs we will check it for you for free.  We will provide you with the log condition number for your design and the variance inflation factors (VIF's) for each term in your model.  Don't worry!   We will explain in plain English what these numbers mean.  If we feel your experiment could benefit from one of our services, we will tell you about it.  For example, suppose your design has a large log condition number, indicating a poor design which could give incorrect or misleading results.  We can recommend our custom design service to help you get your work done well.  Knowing whether your experiment design is good before you collect any data can save time, money, and frustration.

Option 2:

Done with the experiment but didn't get the results you were looking for?  All may not be lost.

If you send us your experiment design along with the data you have collected we will provide you with the same information as in Option 1 plus we will tell you if we see any need for data transformation or any indication of data contamination by extraneous factors.  If we see either of these, or anything else that might cause you trouble, we will recommend a service to help you get things right.   Knowing about potential trouble before it hurts you is certainly helpful -- and knowing how to avoid the trouble is invaluable!

What We Need From You

We will need a data file in either ASCII or Excel format including factor settings and run order.  If you have already collected data, please send your response values too.  We will also need to know your goals and what model you plan to fit. 

If you need help deciding what to send us, please call (206) 525-9093 or write to info@processbuilder.com.   We'll be happy to help you.