marketingQED Logo
   improving marketing effectiveness
 
   
     
 
Products: modelQED® Frequently Asked Questions  

 

modelQED Overview...
What Is Marketing Mix Modelling?
modelQED Key Features...
Which Metrics Can We Analyse Using modelQED?
How modelQED Helps You Calculate Marketing ROI
Automating modelQED For Quicker Marketing Analysis?
modelQED Frequently Asked Questions...

   

 

   

Usage FAQ

  1. Q. I don’t really understand econometric analysis. Can I still use modelQED?
    A. If you are a relatively numerate person with a good understanding of marketing then you should have no significant issues understanding how modelQED works and how to get the best out of it. For those who are less sure in their abilities, we offer training and support packages to ensure that they produce the best results possible.
  2. Q. How does modelQED "automatically" find the best model?
    A. modelQED uses "Genetic algorithm" technologies to search for what it regards as the "best" model from a set of given data. The definition of the "best" model is defined by the user and is based on the available model statistics and the aims of the study.
  3. Q. How do I know when a model is complete?
    A. No modelling methodology can ever say that a model is "complete" since a model is by definition an abstraction or reality. However there are some good rules of thumb to be aware of:
    • Sense check - do the findings make sense and fit with my own understanding of how I think the world works?
    • Statistics check - do the basic statistics show that the model explains enough of the variation in the dependent variable?
    • How confident can you be in the results - the statistics will give you some guidelines to the statistical validity but sense checking results will also give you some idea of "money at risk"
  4. Q. We don’t think we have the correct data available. What data do we actually need as a minimum for the model to be of any use?
    A. The minimum data requirements for a model to be "reliable" depend on many factors. modelQED requires users to import "time series" data in order to create a model. This means that it needs to be fed systematic observations of a dependent variable (the thing which needs to be explained) and equivalent observations of explanatory variables (i.e. the activities which may explain changes in the dependent variable). As a rule of thumb, you would be advised to include a minimum of 30 observations in order to generate a reliable statistical model. More data points should increase the reliability of findings while fewer observations will mean model finding should be treated as being "directional".

 

Technical FAQ

  1. Q. We already use SAS / SPSS / Oxmetrics to run our analysis. Why is modelQED any different?
    A. modelQED is different to these packages for two key reasons. Firstly it is only designed to produce marketing mix models. You’re not paying for thousands of procedures that you will never use nor are you paying for content aimed at a different industry. Secondly modelQED is simple to learn and use - we believe that an experienced modeller will be able to pick this tool up and use it effectively in a matter of minutes. Full reports can be compiled in under an hour to presentation quality standard.
  2. Q. We are looking to run the application on a server and share resources across our team. Can we do this?
    A. modelQED cannot be used on a server type machine (as defined by the operating system type being used). A server based version of the application is in development and will be launched shortly. Please e-mail qed@marketingqed.com for further details.
  3. Q. We’re not based in the UK but we’d like to see a demo or arrange a trial. Is this possible?
    A. Absolutely - we are able to offer trials of the software to any locality and the software can be purchased via international bank transfer. We can also arrange web based demonstrations so that you can see the power of modelQED for yourselves.
  4. Q. I’m having technical issues with modelQED. How can I report this issue?
    A. To report bugs or issues with modelQED, simply email us with as much detail of the issue you are experiencing as you are able to and any supporting documentation (e.g. screenshots, project files etc...) to
    supportQED@marketingQED.com and we will endeavour to respond to you as quickly as possible.
  5. Q. We are using Windows Vista. Is modelQED compatible with this operating system?
    A. Currently modelQED will not run on Windows Vista. We are working towards a Vista compatible release which is scheduled for late 2007. Please notify us if you require an edition of modelQED for Vista.
  6. Q. What are the minimum specifications of PC I need to be able to use modelQED?
    A. You will need to have the following specification as an absolute minimum:
    • Windows XP (Professional or Home Edition)
    • Intel Pentium/Celeron family, or AMD K6/Athlon/Duron family, or compatible processor recommended
    • 256MB RAM
    • Screen Resolution 1028 x 768
    • 30MB Free Hard Disk Space
    Whilst modelQED will run under the above specification, it is quite memory and processor intensive. The speed at which modelQED can identify a model is proportional to the processing speed and amount of memory of the machine it is run on. We therefore recommend the following specification for best performance:
    • Windows XP Professional SP2
    • Single Processor running at 2Ghz or Duo Core processor or greater
    • 1GB RAM or higher (512MB is fine but program may operate more slowly) or greater
    • Screen Resolution 1280 x 1024 or higher

 

Ask A Question

If you have a question that has not been covered in the above, please feel free to drop us an email and we will endeavour to answer it for you: qed@marketingQED.com

   
 
     
 
 Need more help? E-mail: QED@marketingQED.com
Copyright © 2007  marketingQED Ltd. All Rights Reserved.