Have an idea?

Visit Sawtooth Software Feedback to share your ideas on how we can improve our products.

How to generate a MBC design using SSI web

Hi,

I want to make a MBC design so that there are 4 base models (with varying prices) and 15 other levels (with varying prices). Need to get feature wise pricing and top combinations. How do I generate a design? Do I need to use SSI web or any other platform.

Thanks,
Amrita
asked Aug 8, 2019 by anonymous

1 Answer

0 votes
Armita,

If I understand correctly you have 4 base models with varying prices, of which a respondent must select one, and then 15 other variables each with varying prices.  You can use our software to make a design (I usually make it as a single-concept CBC (with 19 attributes in your case), but you can also create it as a single-profile CVA experiment (though the latter will make the design with a very limited number of blocks).

Simply create a design with 19 attributes and the same number of levels as you have price points.
answered Aug 8, 2019 by Keith Chrzan Platinum Sawtooth Software, Inc. (90,575 points)
Thanks Keith for your response. I work in the same team with Amrita :)

We can create a design with 19 feature (4 base feature say brands +15 add-on features) each with varying prices.Thus each choice task would have 19 concept ( for 19 features) with one tested price level (say we are testing 5 price levels for each feature).

We program it in a way that respondent chooses any one brand + any number of features (i.e multiple choice response). Input  the data (in desired format) to be used in Sawtooth software MBC for estimation.

In that case, Can we still get feature wise pricing & top combination of features? We would like to know what are the other deliverable of MBC?

If we make a single concept CBC then we are showing only one concept (i.e one feature) +none in the task where as CVA produces either single concept or pairwise full profile designs. In both cases our aim of getting consumer response on multiple feature selection at various prices to execute MBC could still be achieved?
Just to be clear, you can create the experimental DESIGN using our software or any other experimental design software or technology you want.  But you will NOT program your survey as a CBC or as a CVA - you need to program your questions themselves as custom questions, to allow the kinds of responses you need to get for your MBC.  

In addition, MBC is analysis-only software that will run your models and produce your simulator, but it will NOT program your questionnaire.
Hi Keith,

Thanks for your response. It is very helpful.

We have an additional question with respect to the deliverable and the capabilities of MBC simulator.

Using the MBC analysis tool, can we get the following reports?

- Feature wise pricing - how much the consumers are willing to pay for an additional feature, since we are showing price for all the features separately in  the survey and people are making a choice for the base model + any additional features.

- Top combination of the products (Product comprises of any base model + any features)

- Can we still get the option of playing with the scenarios by including or excluding the feature and varying prices to see the shift in consumer's choice from one product to another?

It will be really great to know these answers for the right use of the MBC tool.

Thanks again,
Amrita
Please read the MBC manual for a complete description and examples of deliverables.  

You will be able to answer your questions about feature-wise pricing by using the simulator and seeing what happens at different price levels for a given feature.  

You will only be able to exclude features if you build product availability into your design and you estimate cross-effects (typically if I have a 3 level price variable for a feature I might add a 4th level to be "not shown" meaning I don't show that feature at all at any price in a quarter of my choice questions.

The simulator does not do optimization to help you find the top combination.  That's more of a manual process on your part, unless you write an optimization function in Solver in your Excel simulator.
...