Successful marketing involves optimizing features, messaging (advertising), and prices for products and services. Conjoint analysis is a powerful marketing research tool to help you get it right before you commit large amounts of money/time to manufacture or advertise the product.
Conjoint analysis also is widely used for repositioning or revamping existing products/services.
Conjoint Analysis in Marketing Research
If you wanted to learn what a group of people wanted in a new product or service, you’d naturally think just to ask them. But, if you don’t ask good questions, you’ll get essentially meaningless information.
For example, people will say they want all the best features at the lowest price. Not helpful!
People usually cannot get all the best features at the lowest price in the real world. That’s not reality. Life involves tradeoffs. That’s where conjoint analysis excels. Buyers shop for products/services that offer superior features and value compared to the competition. They trade off features and price to find an alternative that in their mind beats other options and is satisfactory enough to purchase.
Conjoint Analysis questions mimic that process.
Conjoint Analysis involves a type of realistic series of survey questions (shown below) you ask respondents to learn which features they want, how important the features are, and their price sensitivity.
With conjoint analysis, we don’t directly ask people what’s important to them or how much they are willing to pay. Rather, we show them a series of realistic choice scenarios involving product features and (often) prices and learn that information by asking for each scenario which alternative they would choose.
How Does Conjoint Analysis Work?
In a conjoint analysis questionnaire, we show respondents product/service options (known as product concepts or profiles) and ask which one they would choose.
By scientifically varying (using an orthogonal design) the features and prices that make up the product options, we can observe what features are driving choice. We also can include a “None” option, so that respondents can tell us that none of the alternatives appeal to them.
To see how conjoint analysis works, take this interactive survey example regarding foods you can buy at a baseball stadium. It’s typical to show respondents 3-5 product alternatives per choice scenario. Over a period of typically 3-6 minutes, respondents complete usually 8-15 choice scenarios.
The product alternatives vary in each scenario so that respondents have an opportunity to respond to all attributes and levels in the study. Conjoint analysis questions are usually included as part of a longer 10-20 minute market research survey.
Once we collect the data (typically we use 200 to 1000 respondents), our conjoint analysis software fits a predictive statistical model to predict each person’s choices. The usual model is the hierarchical Bayes MNL regression model that leads to a set of preference scores (called Utilities) for each attribute and level in the experiment.
Using the conjoint analysis preference scores (Utilities), the software creates an intuitive market simulator. The market simulator can be in Excel, web-based, or Windows desktop based. The market simulator lets the manager play “what-if” games to find improvements to the product formulation. We can build new product formulations at optimal prices to optimize market share, revenue, or profit. Market simulators can also operate in automated sensitivity or search mode to find optimal product and price combinations.
Applications and Examples of Conjoint Analysis in Marketing
Product/Service Design Using Conjoint Analysis
One of the first and most famous examples of conjoint analysis in marketing was the Marriott Corporation’s development of its Courtyard hotel. The consultants used conjoint analysis and worked with Marriott stakeholders, designers, and customers to develop a long list of potential features (attributes) that could be in the hotel. Next, the consultants focused on interviewing the target market segment (business travelers) and presenting different hotel concepts made up of the variety of features they were studying. Using the conjoint analysis preference (called Utilities) and importance scores, the team designed the first versions of the successful Courtyard hotel.
Speakers at the Sawtooth Software conferences have discussed other examples of well-known companies using conjoint analysis for marketing research:
- Microsoft (design of peripherals and product line decisions)
- Procter & Gamble (design, messaging, and pricing for consumer packaged goods, CPG)
- NBC Universal Parks & Resorts (theme park experience design)
- Riot Games (video game design)
- Bose (product design and product line extension)
Marketing Segmentation Using Conjoint Analysis
Finding groups of people with similar needs (market segmentation) is critical to modern marketing practice. This lets you create a product line of differentiated products that appeal to unique market segments.
Segmentation algorithms find groups of people who have similar preferences within each group, yet different preferences between groups. Common algorithms for segmentation with conjoint analysis include latent class analysis and cluster/ensemble analysis.
When you develop market segments using conjoint analysis, you can identify which respondents belong to each segment. Profiling segment membership using other customer information, including attitudes, channels, demographics, and firmographics, helps marketers target successful communication and advertising strategies.
If you leverage needs-based segmentation to guide your product optimization, the offerings in your product line will be less likely to cannibalize your own products. Again, the conjoint market simulator together with the segmentation filters guide your efforts.
Packaging and Messaging Optimization Using MaxDiff
Firms like Procter & Gamble regularly use conjoint and MaxDiff for testing packaging styles, colors, graphics, and claims/messaging. You can create graphics with transparent layers that interact with our software to construct thousands of product options dynamically.
Here’s a graphic from a recent P&G presentation at the Sawtooth Software conference showing how they use MaxDiff for messaging research. P&G also uses conjoint for similar research, depending on the situation. If 1000s of combinations of features need to be investigated, then conjoint analysis is preferred. If the combinations of packaging and messages account for no more than about 60 prototypes to be tested, then MaxDiff may be the preferred route. Using MaxDiff allows you to identify the messages and combine them into claims that enhance preference.
Pricing Research Using Conjoint Analysis
One of the most important aspects of marketing is charging the right price. Conjoint analysis provides a way to measure buyers’ price sensitivity without directly asking them how much they’re willing to pay.
Lifetime Products, for example, extensively used conjoint analysis to determine the right price for a variety of their consumer products, including tables, chairs, storage sheds, trailers, kayaks, and coolers. The strategic pricing insights were critical as they formulated proper pricing with their retail channels.
When price is one of the attributes in a conjoint analysis survey, this lets us measure how respondents react to changes in price within the realistic context of competition. And, because other attributes are also varying in the scenarios, the respondent doesn’t see it just as a pricing question.
Final Thoughts: Why Use Sawtooth Software
Conjoint analysis has become the premier tool in marketing research for product design, messaging, and pricing. Sawtooth Software is recognized worldwide as the long-standing experts with the most well-developed software for executing successful conjoint analysis studies.