Developers of innovative new technologies or services face a series of trade-offs in their design phase. They can enhance the market appeal of the product or service by incorporating richer functionality and superior quality, but doing so increases the development and production costs as well as the required price to break even. The trade-off is further complicated by the presence of multiple level offerings of the same brand, since these can often cannibalize each other. For example, if a telecommunications provider introduces a “basic” and “deluxe” version of a new service, the choices made in the pricing and features for each version will affect the overall market success and profitability of the line as a whole.
Fortunately, market research methodologies exist that can help to optimize design before production and market launch occur. Just as designers must make trade-offs in their specifications, these survey methodologies require buyers to make trade-offs as a way of quantifying the impact of price and design on choice. Termed “conjoint” or “choice” analysis, these methods involve an interesting data collection approach that presents consumers with product choices and asks them to behave as they would in a real purchase situation. Ultimately, the information is used to create a simulation tool, based on a rigorous mathematical model, to help managers test the impact of various features and price configurations on the demand and profitability of their products or services.
Implementing a Choice Study
The first step in designing research for the optimization of a product or service is to carefully identify the parameters of the optimization problem. This requires a close dialogue between a research professional and the manager responsible for design. As the chief methodologist at Rockbridge, I often get involved in these conversations with clients. Ultimately, we try to agree on the following with our client:
- What are the underlying “attributes” that determine design? Attributes are decisions that vary in the design decision
- Within each “attribute,” which “levels” need to be evaluated? An attribute must have at least two levels. For example, it may include the absence or existence of an added-value feature, such as a cell phone including a built in camera or not. Or, it may include a series of options, such as a disposable battery or rechargeable battery, four screen sizes on a TV, or three warranty options.
- What is the proposed pricing structure, and what is the range of prices under consideration? For example, some products may have two elements, a one-time (such as installation or equipment purchase) and recurring price (such as a monthly subscription or annual licensing fee).
- What is the decision context for the new products? For example, if a line of products are being offered simultaneously, it is critical to evaluate all of them at once in order to make the optimal decision. In one study, we tested a “discount,” “standard” and “premium option” for a portable internet service. The demand for the “discount” option increased if its price were lower, but it was important to know if the demand came from new sales or from cannibalizing the “standard” option.
After obtaining input on the decision and marketing context from managers, the next step is to design a “choice task” for presentation to consumers in a survey. A choice task consists of a set of two or more product alternatives, each with different features and prices. The example below suggests what this might look like for a television product study. A consumer would be shown this set of alternatives and asked to choose which one he or she would purchase if shopping for the product. It is critical to include a choice of “none” since it is possible that nothing in the set will be acceptable to the buyer.
Example of a Choice Task
| Option: | Brand X | Brand Y | Brand Z: Basic | Brand Z: Premium | None of the above |
| Features: | 42” screen | 50” screen | 27” screen | 42” screen | |
DVD Player built in | DVD Player built in | No DVD Player | DVD Player built in | ||
Not HDTV Ready | HDTV Ready | Not HDTV Ready | HDTV Ready | ||
Surround Sound | Standard Audio | Standard Audio | Surround Sound | ||
12 mo. warranty | 90 day warranty | 90 day warranty | 90 day warranty | ||
| Price: | $2,999 | $3,499 | $1,499 | $2,799 | |
| Choose One: | ( ) | ( ) | ( ) | ( ) | ( ) |
In this kind of study, consumers are presented with more than one “choice set,” typically, anywhere from 4 to 16 different scenarios. As the consumer evaluates the choice sets, the features and prices vary, forcing a continual evaluation of trade-offs. For example, the size of the TV may be larger (50” instead of 42”), but the price may be more, forcing the consumer to consider whether it is really worth trading up.
Conducting a choice study is efficient when using web-based interviewing. A web survey can be designed to present product information, including pictures, tables and pop-up descriptions. The web allows survey respondents to evaluate the information in each choice set carefully before clicking their desired option. Rockbridge has access to online consumer panels that allow screening for particular demographic groups, and we can also access panels of organizational decision-makers for b2b products and services. The sample size for a choice study is usually at least 150, and typically ranges from 600 to 1,500 individuals depending on the complexity of the design and the number of sub-groups being studied.
Turning Information into Action
The responses to a choice study are analyzed with sophisticated multivariate statistical methods, ultimately producing a mathematical model that predicts consumer choice. The model provides predictions of share based on the product features and pricing introduced. The output for management decision-makers is much simpler, consisting of a simulation tool that is powered by a spreadsheet.
The inputs a manager makes consist of price and feature assumptions. These must be within the range captured in the survey. For example, if three types of technology were tested in a study, a manager could enter either one of them as an assumption (but could not work with one not included in the design). If prices from $799 to $1,999 were tested in the study, it would not be possible to test $2,399 because it is outside the range. However, a predictive model could capture any price within the range tested so that a fine-tuned price, such as $833, could be tested.
We typically work with our clients to create a business model that captures all parameters that might affect a market decision. For instance, the output for a model can also include a revenue projection (price X share). But it can also incorporate assumptions about costs for individual features so that the model can output estimates of profit as well as share and revenue.
The table below illustrates how a product manager introducing a broadband internet package might decide on whether to offer a “fast” speed (2.0 mbps) or a “super fast” 4.5 mbps speed, assuming that the cost difference per month is $1.50. Four scenarios are examined: the baseline product with the “fast speed” at a subscription price of $10.00/month, and an upgrade to a “super fast” speed with prices of $10.00, $11.00, $12.00 and $13.00. The choice model shows that – price equal – the share increases from 25% to 35% with the higher quality. Raising the price to cover the higher cost results in lower share, and the optimal scenario is to offer the upgrade at a price of $11.00. It is interesting that the price increase of $1.00 does not cover the additional $1.50 in unit production; yet, the impact on share is so substantial that profits actually increase despite the higher cost.
MARKET SIMULATION FOR A BROADBAND INTERNET PACKAGE | |||||
| INPUTS: | |||||
| Speed (Mbps) | 2.0 mbps | 4.5 mbps (all other features same) | |||
| Market Price: | $10.00 | $10.00 | $11.00 | $12.00 | $13.00 |
| OUTPUTS: | |||||
| Market Share (from choice model) | 25% | 35% | 30% | 24% | 18% |
| Revenue | $375 M | $525 M | $495 M | $432 M | $351 M |
| Profit (variable production cost is $4/month for 2.0 speed, $5.50 for 4.5 speed) | $225 M | $236 M | $248 M | $234 M | $203 |
A Worthy Investment
An optimization study based on a choice design can be a sound investment of resources. The research will cost tens of thousands to conduct but can save millions by helping to establish optimal production and pricing standards before launch.