[MUSIC] Welcome back to our course on Quantitative Customer Insight Techniques. I'm James Lenz with the University of Illinois. And today, we're going through lecture three of the module two on concept testing. And again, this whole module is about you getting an experience and doing surveying and doing analysis with a data to determine customer benefits and the valuation of the different product ideas. So again, we're talking about conjoint analysis. You remember this, there's three aspects to it. Collecting trade-off data, estimating buyer value systems, and then making choices. This is an important part of any type of analysis is for you to use the data and make your choices. It's your decisions that come out of this not the data's decisions that come out of this. An important part of this is you can also do these surveys besides face to face surveys. You can also do them online as we know. We call them PC-based interviewing, so you can do them with just sending out mailings or as well as having some sort of way of collecting the data. These typically are done with more broad type activities. I like to work what I call a taste test approach when you're first starting out. And it's a little bit more fun to collect the data and work with a smaller group of people. So let's go through this analysis that we have, this problem that we wanted to work on. And this, again is a very simple I think market question compared to a market question you might have about a completely new idea. I've worked on many new ideas, new to the world ideas. Here it's much more difficult to really phrase the question properly. But it's a good example because let's take something that we already kind of exist, but let's ask a market question about them. And in this case is can we determine the market preference in a certain neighborhood for macaron cookies? And here's six boulangeries located in an area of Paris and I talked with each of these, and found out what is the most popular macaron that they sell. All of them said it was chocolate. So we purchased a chocolate macaron in each one of these. You'll see the prices at each one of these places. And so then when we set up a conjoint analysis to determine, can I determine what is the market preference? So I thought about what evaluation factors should we consider? Well, appearance certainly is the first one to consider. Because that's how you're going to buy the cookie. When you're looking at them, you're going to look at which one, so that appearance is an important aspect. And so if I'm a boulangerie selling this cookie, I want to make sure I have a unique appearance and I'm going to money or put effort into this appearance. Texture, of course is another thing, especially for many of us. The texture of these products is important in order for whether we really like it or not. And a third aspect, of course, is taste. But again, coming to the idea of just two, trying to compare two. I said, well, let's boil this down to appearance and taste. Because the appearance sells the cookie the first time, but its taste is what's going to get you to come back and want to buy a second one. So as a baker, I'm going to have to trade these two items off. And each of these six companies must trade them off, decide what is the most profit for them and what is the best way for their business to run. I chose a ranking from one to six, because I like rankings as opposed to just a what is called a knock out scoring. Do you like this one versus that one or just grouping. Put them groups of things and then having three rounds and sort of a grouping, scoring system. I think it's ranking, because you sort of get all the product at one time. The best ones boil to the top this way. So that was the technique that I selected and these were the two parameters that I chose to work on. So the survey was created. There was blind testing that was done with groups of customers. And you saw in the video, one of those groups five people that were comparing them and scoring them live as part of that video. So the first analysis, I'm just going to use these five people. I've got a letter for their name and then F and M, whether they're female or male just to login who they were. So in the world of appearance how each one ranked, these six different boulangeries and from that is easily able to calculate an average and I can also plot the data. So there's the data shown in a table above, as well as plotted below and what's interesting to look at in here is how much spread there is in the data. In some cases like Maison Hilaire, you can see almost everybody from appearance did not like the appearance of this cookie. Whereas in the 28, Maison Landemaine, almost everyone liked the appearance of this cookie. So you can see this. But on average, you kind of get this strain. So there's clearly a difference in the customer expectations. The expert's appearance of ranking the appearance and what do they think of the appearance of these cookies. The second place is in tasting. So they ask them to rank the tasting. And you can see here again, it's not the same. It doesn't follow the appearance. The taste is a different parameter than appearance and it's scored differently, and comes up a little bit differently. In fact here, there's a little bit more spread in the data I want to say. And again, you should analyze this. I use all the data, but maybe some of these outliers, I said we could take them out and might be helpful for making a more accurate estimate of what the customer preference really is. So from here now, I'll calculate what we call a value. I've already said what the price is, what are the costs to the consumer for each one of these cookies. It's listed here. And now, I can calculate from this data a normalized value. This normalized value is based on three factors. It's the cost and then I add two factors to it. A taste survey factor and an appearance survey factor, and these are normalized to be in the form of cost. So the taste survey factor I take is the average rank that it has -3, because 3 would be the average score. And the appearance survey factor is an average rank -3, as well. So I take those that average ranking that you saw in the previous two charts, subtract 3 and then I modify that and add it in to create this normalized value. A little more descriptive way of doing this, a little more detail is a normalized value as again is this average cost of all the six cookies. And then this weighting factor is a function of what I call the weight. So I give twice as much weight to taste over appearance. If I were trying to sell this, I feel like the baker or the person's going to feel like the taste is more important. And to a consumer, the taste is going to be more important than appearance for me to like that cookie or want to purchase that cookie. So then again, it becomes a little bit more complicated by adding a weighting factor. This taste factor that I mentioned before and also a sensitivity factor. I have to put a sensitivity factor in there to help scale the scoring sheets into the pricing to make that fit properly. So these are numbers that you can play with and I've got a spreadsheet that we have, that I can play with these numbers to see does this make sense or not. So you can see all this analysis a little bit subjective. Even though there's very concrete numbers, you can use this to sort of tweak it and work with the data and say, is this what it means? So I can play with the scoring a little bit to say, does this value make sense or not? And as a result of that, you get two graphs. You get the cost as it's plotted for each of these boulangeries and then you get this normalized value. And now, you can see a difference. There's a difference then between the cost and the normalized value that you've computed from your survey. And now, you can make this as a table. So each of these, I can calculate what I call added value. So if there normalized value is above their cost, then they're adding value. The value that you think you should pay, they're giving you more value than what you're paying for the product. If the normalized value is below their cost, then they're taking value. They're taking value from you as a consumer and they're not putting these expensive ingredients in. They're not doing something there to give you the real value of what they're trying to price this for. And many times, they can get other value because they have a nicer store and they make it cleaner in a nicer part of town. There's maybe other reasons they've added value besides just to the cookie itself and you should consider that. But those added values is the price difference that the survey will suggest from what the price currently is. What we ought do is now make choices. What choices could we make from this data? We've got data that's available. So which macaron will people buy the most? As I said, it's quite interesting around Paris. In many cities, how different consumer products can be attracted by one place or the other. And you have to always understand the importance of social media as well as playing a role in driving prices or driving customer behavior. In this case we were not using that we were just using the scoring of appearance and taste so I look at the data again from my value chart. Would I go for the lowest price and the fairer value? This may sound hilarious. It's actually the lowest one, but you'll see they priced it at really what the value of the marketplace says. So I think that's a very fair pricing and there are businesses trying to get a low cost provider and not give you the best tasting, but give you one that's fair for the value that you pay. The other one where most people on the survey, they came back and when I show this data to people and we analyze it. And again, throwing back our own conclusions, really the best value is this 28 Maison Landemaine. Here they're pricing it at 1.50, but they have a value of 1.74. So many people are more and more attracted to this cookie. And in fact, in the neighborhood, everybody believes this is the best cookie. And so now, we have data to illustrate why they think this is the best cookie in the neighborhood. And what I've now created for you is a spreadsheet that we can use and I want to show you an example of it. Again, in a very simple form, we'd like you to run a taste test of this in the next couple modules and help you prepare for this. But if I just look at this, if I just enter the data. Say I have five people that ran the test and each one ranked from the product one, product two, product three, product four. I call them makers here and they just scored it, everybody scored it the same. You can see my graph and my data looks very similar. The graph is very similar to the table above. Everybody just scored it exactly the same and they are scored and ranked one, two, three, four. And this would be an example for taste. If I did the same thing for appearance, I get a graph of again one, two, three, four. Everybody's score is identical. So this is just a very simple example of how the data would come out and how you can think about how this analysis is done. So from here, I can again, calculate a value or cost from the equations that I gave you. I've got a weight of tastes is two. A weight of appearance is one. Sensitivity is five. And now I can see my normalized value is a straight line as I would expect from the analysis that I've done. But now, their costs are different. Yes, because they are pricing it different. So even in this simple analysis, I can see who's adding value and who's taking value away from if they were just all being scored this way. I'm the number one cookie in town. I'm the number two cookie, the number three and number four. I'm still not pricey and I'm choosing ways to price things differently. So you'll see that you can even in these situations still find more value. As a supplier, how important is price? Customers do perceive value. They find value. And so you have to understand the pricing sensitivities and especially that's what I like about taste tests is because it's pretty clear people that buy things for the value not just the lowest price. So now, how do I find that value and this type of analysis can help me find that value. So in summary here, we want you to envision how you might structure a conjoint analysis for your product idea. What parameters are important to the design? What will drive cost? Which will drive customer preferences? Can you envision this survey? Can you start to think about how you would position a conjoint analysis around your idea, the idea that you've got already put into a spreadsheet? Can you envision this? I want you to do some thought experiments and thinking about this, and then there'll be a short quiz here. And then we'll come back and have a lecture that will help you prepare a survey for a conjoint analysis. [MUSIC]
