In today's highly competitive marketplace it is not surprising that an increasing number of college administrators are looking for better data to inform their decisions about tuition pricing and financial aid allocation. In particular, they want answers to such questions as:
How much can we raise tuition without negatively impacting enrollment?
How does our tuition impact share or preference?
How do prospective students and parents perceive the institution's current brand value?
How much financial aid do we need to spend to meet our enrollment goals?
Are we targeting our aid effectively and efficiently?
Two methodologies--tuition pricing studies and predictive modeling--will provide clarity and enormous insight into those questions.
Before we proceed with a discussion of these two approaches, it is important to clarify terms. Based on conversations with clients and colleagues, we know that tuition pricing studies and predictive modeling (also known as econometric modeling) are often, and erroneously, used interchangeably. This has caused a great deal of confusion, and even frustration, as administrators sometimes undertake one study and think they are doing the other. Both methodologies measure price elasticity--the degree to which demand changes as a function of price. If your price elasticity is low, a large change in price will lead to a small change in demand. Tuition pricing studies rely on surveying prospective students, and parents, regarding their preferences to understand how demand would be impacted by a change in sticker price. Predictive modeling assesses price elasticity through analyzing the historical behavior of admitted students to understand how price discounts (financial aid) as well as various student characteristics (biographic, economic, demographic, academic, etc.) interact to impact matriculation decisions. The results of this analysis are typically used to inform financial aid awarding strategies (often called financial aid leveraging).
With upward pressure from salaries, technology, health benefits, and energy, the desire to raise tuition, often significantly, is strong. Seldom, however, are these decisions made with any real sense of what the market--prospective students and families--will pay. There may be some concern about how the increase might affect where the institution ranks vis-?-vis its competitors but, other than that, there is almost no consideration of external data. A pricing study is designed to address this omission.
Discussions about tuition tend to stress one element of the decision-making process (price) and don't acknowledge the other side of the equation (benefits). In making a decision, students juxtapose price and benefits to determine value. If they perceive that the benefits outweigh the cost, they act. Fortunately, it is possible to drill down on the benefits or attributes that students value most, and are willing to pay for, through the use of a tuition pricing study.
A tuition pricing study involves a blind survey of prospective students and their parents. Since response rates are better, and the sample can be controlled more carefully, most studies are conducted via telephone.
The actual tool we use is called choice-based conjoint analysis. This technique identifies the factors that a person CONsiders JOINTly (conjoint) when making a decision. Conjoint analysis allows us to address the fact that students and others rarely make a college choice decision solely on price. In fact, they almost never choose the lowest-priced college without considering other attributes and benefits, such as brand, location, reputation, etc. At the same time, students generally will not choose a college or university solely on reputation without considering price and other attributes.
Each study tests a minimum of two attributes--the name of the college and its competitors-and varying tuition levels. Other attributes, of course, can be tested.
To begin, you must identify the set of schools you wish to "test" as part of the study. In most cases these are your overlap institutions. In some cases you might include an aspirational school or two. Once this set is finalized, a range of tuition prices is created that encompasses the tuition points of all the institutions included in the study, including price points you wish to test. The key to a successful conjoint study is to test the college's name and perceived value, as well as the name and perceived value of major competitors, at different price points. By asking students and parents to react to each of these simulated decisions, we test different variables, or scenarios, against one another.
As part of the test, the college names and tuition prices are paired with each other in different combinations to develop a hypothetical product. Respondents are then given two options, each with two attributes (a name and a price) and asked to choose the institution that which they believe offers the best value. A series of these questions is shown below:
1. Which of the following schools would you prefer to attend?
__ Old Siwash for $24,000 per year
__ Akron State for $14,000 per year
2. Which of the following schools would you prefer to attend?
__ Vandenberg University for $18,000 per year
__ Akron State for $10,000 per year
3. Which of the following schools would you prefer to attend?
__ University of Des Moines for $14,000 per year
__ Vandenberg University for $18,000 per year
Typically, some 18 to 20 randomly-paired packets of attributes are tested.
One of the most important things you learn through a price study is how much prospective students and parents value your brand and how they compare this value with that of your competitors. These data can be presented in a number of ways. First, let's look at a sample chart that highlights perceived brand value and actual cost to attend.
The solid line, point A, indicates the percentage of respondents who would choose Old Siwash at each price point seen on the x-axis compared to Old Siwash's competitors at their current price point. For example, if Old Siwash cost only $7,500 per year, point B, and the competitors were priced at their current tuition amounts, some 66 percent of respondents would prefer Old Siwash. However, if Old Siwash was priced at $24,000, point C, and its competitors were still at their current price points, only 23 percent would still choose Old Siwash while 32 percent would choose Competitor 1 at their price of $13,000, point D; Competitors 2 and 3 would receive 18 percent of the respondents each; and 10 percent of respondents would choose Competitor 4 at their price of $21,500.
These results give Old Siwash a lot to consider. One recommendation may be that Old Siwash should be conservative with its tuition increases until it can increase its brand awareness and value. Brand value, of course, is increased by demonstrating that the university is stronger than its competitors in those attributes of most importance to prospective students. An increase in brand value will move the ultimate threshold beyond its current point as well as lead to less elasticity among prospective students and parents.
A tuition price study is not suitable for every college or university. For example, if your institution is a commodity buy--in which most students choose you because of either cost or location--a price study will be less effective.
Second, the institution should have a clear sense of its competitor set. If you cannot identify your primary competitors for prospective students, it will be difficult to create meaningful and realistic attribute packages, or scenarios.
Third, price studies are most useful if the competitors you are evaluating have distinctive brand perceptions, or unique selling propositions, and a range of tuition price points. For example, it would be difficult to do a conjoint study to differentiate two top liberal arts colleges with virtually identical missions and price points. However, a price study would be clearly useful in helping you differentiate the brand value of a regional public from the brand value of a regional private.
Next, a good candidate for a price study must be willing to let data guide important decisions related to tuition. Rather than shooting from the hip and saying tuition will increase eight percent next year, these institutions want to know how this increase will affect enrollment and determine, economically, if an eight-percent tuition increase will yield more dollars for the school.
Candidates for a price study must be adroit at brand management. Your brand, or reputation, can have a significant impact on choice. Colleges and universities with strong brand often take advantage of that equity by increasing tuition at a faster rate. Your ability to develop and convey a valued brand is essential.
Finally, there is the notion of differentiation. Some institutions prefer the comfort of the herd and do not wish to differentiate themselves from their competitors. A price study can give you viable insights for meaningful differentiation. For example, one study we completed indicated that students were much more interested in the fact that the school had a conservatory than the fact that it had great faculty. From the students' perspectives, all schools have great faculty and the conservatory was an important point of differentiation.
Through predictive modeling, institutions can better understand which factors impact students' enrollment (or re-enrollment) decisions, including the role played by the financial need of the student/family and price discounts (i.e., financial aid). Once a model is developed, alternative financial aid awarding strategies can be simulated to predict the impact of new approaches on class size, quality, net tuition revenue, and other class characteristics important to the institution.
In predictive modeling, the probability of enrollment for each student is calculated as a function of individual student characteristics appropriate for the institution. The price sensitivity of students is, therefore, measured by analyzing how past admitted students with similar characteristics (e.g., financial need, quality, intended majors, distance from home, etc.) responded to different price discounts (e.g. grants).
In general, when grants offered to an admitted student are increased, two things happen:
1. The probability the student will matriculate (demand) increases
2. The amount of net tuition revenue that will be received from the student declines (Net Tuition Revenue = Tuition - Grant)
Depending on the magnitude of the change in the probability of enrollment, increasing grant may either raise or lower expected net tuition revenue. [Expected tuition revenue = (probability of enrolling) x (net tuition charges).] Groups of students are considered to be price elastic if reducing their price through providing a larger discount would increase probability of enrollment enough that the total net tuition revenue generated by the group would increase, too.
Because multiple years of data are used, and the analysis is based on actual historical behavior, the predictions of future behavior, given a reasonably similar applicant pool, are typically quite reliable.
Although the primary goal of predictive modeling is to optimally target financial aid resources, the analysis does shed light on the issue of "sticker price" as well. For example, if the vast majority of admits are determined to be price elastic, as defined above, higher-than-average sticker price increases would not be recommended. On the other hand, if demand for a particular academic program has been increasing, and the students interested in that program are predominately price inelastic, an increase in sticker price, an additional fee, or less financial aid for students in that major would be viable. Of course, such analytical results would need to be supplemented by an analysis of the sticker prices, discount rates, and "prestige profiles" of competitors. Beyond that, it would be important to understand how the institution markets its value proposition as well as its affordability.
An institution that should consider a predictive study would likely have one or more of the following characteristics. First, it might be a tuition dependent institution that is at capacity and wants to improve the mix of students matriculating (e.g., quality profile, racial and/or geographic diversity, academic program mix, average net tuition revenue). Because growth is not an option for this institution, and because maintaining or ideally increasing average net tuition revenue is critical to meeting financial goals (e.g., faculty salary increases), understanding the "cost" to the institution of changing the mix is critical before action is taken. There is no room for surprises here.
Second, it might be a struggling institution that has to maximize net tuition revenue just to survive, but has no sense of which populations are overfunded (i.e., multiple awards stacked for eligible individuals) or which populations (i.e., out-of-state) would react very positively and enroll at a greater rate with modest increases in grants.
Third, it might be an institution whose leadership (governance, administration, or both) aspires to lofty goals (e.g., significant increase in racial/geographic diversity; significant reduction in the discount rate; dramatic increase in the quality profile of the entering class; all of the above) and direct awarding and admissions policies to change without understanding the complex, dynamic, and often reactive relationships that exist between these variables.
In all of these cases, the higher the stakes, the greater the need for analysis (predictive modeling) in order to strategically deploy financial aid resources (leverage the institutional asset) to meet or exceed goals and to avoid severe, unintended consequences.
Both methodologies add to an institution's understanding of the price sensitivity of its market. Those institutions most interested in testing the strength of their value proposition in the market, and understanding the likely impact of significant changes in sticker price, should consider a tuition price study.
Those institutions most interested in effectively deploying financial aid (differential pricing) based on a variety of student characteristics should consider predictive modeling.
Regardless of the methodology selected, using data to inform strategic decisions about price (sticker or discounted) is critical given that such decisions can significantly impact the financial lifeblood of most colleges and universities.