Synthetic Consumer Pricing Study
When Does a Price Increase Create More Value Than It Costs?
Read Time: 18–22 minutes
Published October 2026
About This Sample
This sample demonstrates how OAI may approach an uncertain business question using a synthetic business and customer population. The investigation examines how a broad pricing question can be broken into measurable parts, tested under different conditions, and refined as new information is introduced.
Establishing the Product
To keep this investigation accessible, this sample uses a bottle of water as its product.
Water is familiar, widely purchased, and simple enough to examine without requiring specialized knowledge. It is straightforward enough to examine clearly while still allowing several meaningful conditions to vary.
Temperature: Water can be refrigerated or unrefrigerated
Availability: One bottle can be immediately available while another requires additional time to obtain
Price context: The same price can appear ordinary in one setting and unusual in another
This framework can be adapted to another product or service using variables appropriate to that business.
The purpose of this sample is not to build a universal theory of bottled water. It is to show how a decision can be broken into measurable parts before testing.
Breaking Down the Pricing Question
We begin with a customer who is already interested in purchasing a bottle of water.
What determines whether the customer chooses one offer over another? The first three variables are fairly straightforward:
Price: The bottle may simply cost too much
Quality: The water may not meet the customer's minimum expectations
Convenience: The additional time or effort required to obtain it may not feel worthwhile
These become the first three variables in our model:
Price Sensitivity
Quality Sensitivity
Convenience Sensitivity
These variables are useful starting points because they respond directly to conditions a business can often influence. Before introducing more variables, we can first examine how these three interact.
Examining Price, Quality, Convenience
Price Sensitivity (A)
Price Sensitivity describes how much a customer is willing to pay before price alone eliminates the offer. Our model uses ten levels, each tied to a maximum accepted price ranging from approximately ten cents to ten dollars.
This variable was configured using real-world examples of what it can cost to acquire a bottle of water across different settings.
At the low end, bottled water purchased in bulk can cost roughly ten cents per bottle. At the high end of our reference set, bottled water sold at a major venue reached $8.50.
The experimental range was extended to $10 so the model could continue beyond the highest observed example rather than ending directly on it.
Quality Sensitivity (B)
Quality Sensitivity describes the minimum product condition a customer requires. In this model, temperature provides a clear progression because it is easy to recognize and can meaningfully change how water is experienced.
This variable was configured using research showing that temperature can affect how water is perceived and enjoyed. Rather than assigning dozens of vague quality levels, our model uses four conditions that are easy to distinguish.
Warm → Room Temperature → Cool → Refreshing
Convenience Sensitivity (C)
Convenience Sensitivity describes how much additional acquisition time a customer will tolerate. Our model uses ten levels, from immediate availability to ten minutes of additional wait or travel time.
Research suggests that additional travel, waiting, or effort can make an otherwise acceptable offer less attractive.
The ten-minute range is an experimental choice rather than a measured bottled-water threshold. It gives the model a practical window over which the effect of additional acquisition time can be observed.
Building the Core Model
Our first three variables can now be combined into a population of unique profiles. Think of each profile as:
Profile N = [A, B, C]
where:
A = Price Sensitivity
B = Quality Sensitivity
C = Convenience Sensitivity
The possible values are:
A = 1 through 10
B = 1 through 4
C = 1 through 10
If every possible combination appears once, the population contains:
10 × 4 × 10 = 400 unique profiles
This creates a neutral comparison set in which every Price, Quality, and Convenience combination appears once. Each profile receives equal weight, allowing the same offer to be tested across the full set while one condition at a time is changed. This gives us a way to begin exploring how the variables affect one another.
Beginning the Investigation
Even in this simplified model, clear relationships begin to emerge between the variables. To illustrate this, we will hold Quality at Level 4 (Refreshing) and compare Bottle Price and Acquisition Time across their full ranges.
Throughout this study, modeled purchase rate represents the model’s estimated acceptance rate under the conditions being tested. The 400-profile core establishes the Price, Quality, and Convenience response points; when Price or Convenience falls between those established points, the model interpolates between them so intermediate conditions can also be tested.
At $1.21 with 2:13 of additional acquisition time, 90% of the population passes the price condition and 80% passes the convenience condition.
Together:
90% × 80% = 72% modeled purchase rate
Across the 400 profile comparison set, this means 288 profiles accept the offer. For this example, each accepted profile is treated as one hypothetical modeled purchase.
Now assume the business purchases each bottle for $0.25. At a selling price of $1.21:
$1.21 − $0.25 = $0.96 margin per modeled purchase
Across 288 modeled purchases:
288 × $0.96 = $276.48 modeled contribution
Now consider a different configuration. The bottle price increases to $2.31, while additional acquisition time decreases from 2:13 to 1:07. At $2.31, 80% of the population passes the price condition. At 1:07, 90% passes the convenience condition.
Together:
80% × 90% = 72% modeled purchase rate
The modeled number of purchases therefore remains 288, but the margin per modeled purchase increases:
$2.31 − $0.25 = $2.06 margin per modeled purchase
Across 288 modeled purchases:
288 × $2.06 = $593.28 modeled contribution
The difference between the two configurations is:
$593.28 − $276.48 = $316.80 additional modeled contribution
This creates a more interesting question than simply asking whether a higher price loses customers. Could part of the additional margin be invested into faster service while preserving the modeled purchase rate? In other words, can we raise prices, invest in faster service, and preserve modeled purchase rate while still increasing modeled contribution?
In this synthetic comparison, the second configuration creates $316.80 more modeled contribution across the 400-profile comparison set. That difference shows how much additional contribution the improved configuration creates on the current synthetic scale. A real business would need to translate that difference into its actual purchase volume before comparing it with the cost of improving convenience.
What the Core Model Reveals
Lower price does not always recover a weak offer
Once Quality or Convenience becomes the binding condition, lowering price can produce diminishing returns. A cheaper bottle can still lose customers if the water is not desirable enough or if obtaining it requires too much additional effort. Discounting is not always the only way to increase sales quantity, and it may even be the wrong response.
Better quality can support a different price
A cool or refreshing bottle can retain quality-sensitive customers who would reject warmer water. The cost of improving the product may be lower than the value created by the additional customers it retains.
Convenience can support a higher price
A more expensive bottle can remain competitive when the cheaper alternative requires enough additional time or effort. However, this does not mean prices should simply be raised whenever service becomes faster.
Improving every variable is not automatically efficient
The model can also begin identifying where improvements stop adding enough modeled purchases to justify their cost.
At this point, the core model is still limited. Price, Quality, and Convenience do not capture every reason customers may react differently to the same offer.
What Else Could Affect This Offer?
Suppose two customers encounter the same bottle at the same price, with the same quality and convenience. Even when Price, Quality, and Convenience are acceptable to both customers, they may still respond to the offer differently.
For example:
A customer may be willing to pay $4 for water while still thinking $4 is unusually expensive if they expected the bottle to cost closer to $2.
Another customer may compare a $1 bottle with a $2 bottle and use the price difference as a clue about quality.
A higher price may create a different reaction depending on whether the increase feels justified.
When several acceptable choices exist, visible popularity may make one option more attractive than another.
These situations allow us to expand the model without replacing the first three variables. The next four variables help us explore how the customer interprets that offer once those basic conditions are considered.
Accounting for Customer Interpretation
These variables operate differently from Price, Quality, and Convenience. A business can directly change a price, improve a product, or reduce a wait time, but it cannot directly control what a customer expected to pay, what a low price communicates, whether a higher price feels justified, or which option appears more popular.
Rather than acting as simple on/off gates, these four variables adjust the modeled purchase rate produced by the first three. Together, they describe how expectations, price signals, perceived justification, and visible popularity can shape the way an offer is interpreted.
Expected Price
Expected Price describes the price a customer expected to pay before seeing the offer. This expectation may come from a nearby competitor or another price the customer already had in mind.
Expected Price differs from Price Sensitivity:
Price Sensitivity: How much is the customer willing to pay?
Expected Price: What price was the customer expecting to pay?
A customer can therefore be willing to pay $3 while still reacting negatively if they expected the same bottle to cost $2.
Price as Quality
Price as Quality asks a different question:
Why might someone buy a $2 bottle when a $1 bottle exists?
For some customers, a lower price is simply better. For others, price may signal quality. If a customer strongly associates unusually low prices with lower quality, a cheaper bottle can become less attractive while a moderately more expensive bottle becomes more attractive.
Our model uses five categories to describe how strongly price is being used as a signal of quality:
No association → Slight association → Moderate association → Strong association → Very strong association
Research supports the broader idea that price can sometimes act as a quality cue. The exact strength and location of that effect depend heavily on context, so its effect in this model is treated as an experimental assumption.
Fairness
Fairness captures how justified a price above expectation feels. The same higher price may be interpreted differently if the customer can see a reason for it. For example:
the product improved
service became faster
the business is operating in a more expensive setting
or another meaningful change occurred
These conditions influence how justified the higher price feels to the customer.
Feels unjustified → Mostly unjustified → Mixed → Mostly justified → Feels justified
A price that feels fully justified receives no additional Fairness penalty. As the same premium feels less justified, Fairness creates a larger negative adjustment.
Social Proof
Social Proof describes how visible popularity can influence a customer’s choice.
The model compares how popular the bottle appears relative to nearby alternatives:
Much less popular → Less popular → About the same → More popular → Much more popular
When the bottle appears more popular, Social Proof creates a modest positive adjustment. When it appears less popular, it creates a modest negative adjustment.
Social Proof remains a secondary influence and should not outweigh poor Price, Quality, or Convenience conditions.
How These Variables Differ From the Core Variables
Price, Quality, and Convenience form the profile-based core of the model.
Expected Price, Price as Quality, Fairness, and Social Proof describe how that offer may be interpreted relative to expectations or alternatives.
That gives us two useful layers:
Does the offer clear the customer’s direct Price, Quality, and Convenience requirements?
How do the surrounding expectations and signals change the way that offer is interpreted?
The first layer is represented by 400 profiles. The additional variables then adjust the modeled purchase rate produced by those profiles rather than becoming new person-level profile dimensions.
Testing the Additional Variables
We can test these variables the same way we tested the core model: hold the surrounding conditions steady and change one variable at a time.
Unless otherwise noted, Quality is held at Refreshing and the bottle requires one additional minute to obtain. The interpretation variables are held at their neutral settings unless one of them is being tested: Price as Quality is 'No association', Fairness is 'Feels justified', and Social Proof is 'About the same'.
Expected Price can change how the same $3 bottle is received
Suppose the bottle costs $3. Now change only Expected Price.
As Expected Price falls from $3 to $1 while the bottle remains priced at $3, modeled purchase rate falls from 67.0% to 46.8%. The farther the offered price moves above what the customer expected to pay, the less attractive the same $3 bottle becomes.
Price as Quality can change which price appears more attractive
Now compare two price relationships under the same surrounding conditions, changing only Price as Quality within each one.
For the $1 bottle, the nearby alternative costs $2.
For the $2 bottle, the nearby alternative costs $1.
As Price as Quality becomes stronger, the relationship between the two offers changes. The $1 bottle loses modeled purchase rate as its unusually low price becomes a stronger negative quality signal, while the $2 bottle gains modeled purchase rate as its higher price becomes a positive signal.
The same price difference can therefore favor the cheaper or more expensive option depending on how strongly price is interpreted as a signal of quality.
Fairness can change how the same premium is received
Suppose the bottle costs $5 and the customer expected something closer to $3. Now change only Fairness.
As Fairness moves from Feels justified toward Feels unjustified, modeled purchase rate falls from 40.0% to 32.8%, even though the bottle remains priced at $5.
The price itself does not change; the modeled response changes because the same premium feels less justified.
Social Proof can change how an otherwise identical offer is received
Suppose the bottle costs $2 and the customer expected about $2. Now change only Social Proof by varying how popular the bottle appears relative to nearby alternatives.
As Social Proof moves from Much less popular to Much more popular, modeled purchase rate rises from 64.1% to 86.7% while the offer itself remains unchanged.
Visible popularity can therefore strengthen or weaken the modeled response to an otherwise identical offer.
Returning to Our Original Question
We can now return to the issue that started this investigation: What happens when a business raises its price?
Rather than begin with one arbitrary starting price, we test several points across the range established earlier:
$1.00 → $1.50 → $2.00 → $3.00 → $5.00 → $8.50
These prices give us several starting points across the experimental range, while $8.50 represents the highest observed real-world price used to establish our Price Sensitivity range. These prices are experimental reference points. They are not assumed to be equally common or typical within any one market.
We then test several price increases:
0% → 5% → 10% → 15% → 20% → 30% → 40%
Within each starting-price test, the surrounding conditions are held constant:
Quality: Refreshing
Convenience: One additional minute
Expected Price: The original starting price
Price as Quality: Moderate association
Fairness: Mostly justified
Social Proof: About the same
The customer therefore begins by expecting the original price. We then raise the selling price while leaving the product, convenience, and other interpretation settings unchanged. This creates a controlled way to observe how the modeled purchase rate changes as the selling price rises above the customer’s original expectation.
Because the Price Sensitivity range ends at $10, any increase that would move the selling price above that point is not tested.
Experiment 1: How Does the Modeled Purchase Rate Change as Price Rises?
We begin by examining the most direct consequence of a price increase: How much modeled purchase rate is lost as the selling price rises above its starting point?
Each row begins at a different current price. We then increase that price while keeping the surrounding conditions established above unchanged.
The table below shows the resulting selling price and modeled purchase rate for each tested increase.
The effect of the same percentage increase depends heavily on where the price begins.
At $1.00, a 40% increase raises the price to $1.40, while modeled purchase rate falls only from 83.6% to 79.9%.
At $5.00, the same 40% increase raises the price to $7.00, and modeled purchase rate falls from 50.5% to 30.8%.
The pattern becomes even more pronounced near the upper end of the Price Sensitivity range. Starting at $8.50, even a 15% increase reduces modeled purchase rate from 21.5% to 10.8%.
The percentage increase alone does not determine the response. The starting price also matters because the same relative increase can move an offer into very different parts of the modeled Price Sensitivity range.
Increases that would move the selling price above $10 are shown as N/A because they fall outside the range established for this model.
Experiment 2: When Does Additional Margin Compensate for a Lower Purchase Rate?
A lower modeled purchase rate does not necessarily mean a price increase is worse for the business. We also need to know whether the additional margin earned on the remaining purchases compensates for those that are lost.
Earlier, we assumed the business purchases each bottle for $0.25. We keep that cost unchanged here.
To make the contribution comparisons easier to read, we express the results across 100 hypothetical purchase opportunities.
Using the $3.00 starting price as an example:
$3.00 − $0.25 = $2.75 margin per modeled purchase
Across 100 purchase opportunities, a 67.0% modeled purchase rate produces:
67.0 × $2.75 = $184.25 modeled contribution
Now increase the price by 20%, from $3.00 to $3.60.
At the new price:
$3.60 − $0.25 = $3.35 margin per modeled purchase
Across the same 100 purchase opportunities, modeled purchase rate falls to 61.3%:
61.3 × $3.35 = about $205.36 modeled contribution
The modeled purchase rate fell, but contribution increased by approximately:
$205.36 − $184.25 = $21.11 additional modeled contribution per 100 purchase opportunities
The price increase loses some modeled purchases, but the additional margin earned on the remaining purchases more than compensates for that loss in this scenario.
We can repeat the same calculation across each starting price tested in Experiment 1.
The table below shows modeled contribution per 100 hypothetical purchase opportunities. Bottle cost remains $0.25 throughout.
A higher price can therefore reduce modeled purchase rate while still increasing modeled contribution. At the lower starting prices, additional margin continues to compensate for the declining purchase rate across the tested increases.
At $3.00, modeled contribution rises through the 30% increase, then begins to decline at 40%.
At $5.00, the tradeoff becomes much tighter. Modeled contribution rises slightly at 5%, remains close to the starting level through 20%, and then declines more sharply.
At $8.50, even the first 5% increase reduces modeled contribution under the current assumptions.
A price increase can lose modeled purchases and still improve contribution, but only while the additional margin earned on the remaining purchases is large enough to offset what was lost.
Experiment 3: How Does Unit Cost Change the Contribution Tradeoff?
Experiment 2 showed that a higher price can increase contribution even when modeled purchase rate falls. But whether that tradeoff is worthwhile also depends on the cost of the product.
To test this, we hold the starting price at $3.00 and vary only the unit cost. Everything else remains unchanged.
The table below shows modeled contribution per 100 hypothetical purchase opportunities.
The pattern changes as unit cost rises. At $0.10 and $0.25 per bottle, modeled contribution peaks at the 30% price increase before declining slightly at 40%.
At a $0.50 unit cost, the same turning point remains, but the decline is much smaller: contribution moves from $192.44 at 30% to $192.03 at 40%.
At a $1.00 unit cost, modeled contribution continues rising across the entire tested range, reaching $166.08 at the 40% increase.
When the starting margin is already large, losing too many modeled purchases can eventually outweigh the benefit of charging more. When the starting margin is thinner, a price increase can continue adding contribution across a larger part of the tested range.
The contribution tradeoff therefore depends partly on the cost structure behind the product.
Experiment 4: How Can Two Variables Pull the Same Price in Opposite Directions?
So far, we have tested the interpretation variables one at a time. Now we can test what happens when two of them influence the same offer at once.
Suppose the business currently charges $3.00 and considers a 20% increase to $3.60. To examine that interaction, we hold the surrounding conditions steady:
Quality: Refreshing
Convenience: One additional minute
Fairness: Mostly justified
Social Proof: About the same
We then vary two things:
Expected Price
Price as Quality
The table below shows modeled purchase rate at the same $3.60 selling price under different combinations of Expected Price and Price as Quality.
If the customer expected to pay $2.00, the $3.60 price sits well above expectation, which lowers modeled purchase rate. As Price as Quality becomes stronger, however, that higher price also becomes a more positive quality signal, raising modeled purchase rate from 45.2% to 55.3%.
If the customer expected to pay $4.00, the pattern reverses. The $3.60 price now looks relatively inexpensive, but a stronger Price as Quality association can make that lower price signal lower quality. Modeled purchase rate falls from 62.0% to 56.5%.
When the customer expected something closer to $3.00, changes in Price as Quality have much less effect because the selling price is already close to expectation.
The same $3.60 price can therefore become more attractive, less attractive, or change very little depending on what the customer expected and how strongly they use price as a signal of quality.
Paths for Further Investigation
The experiments in this sample use a controlled synthetic population so that individual relationships can be examined clearly. A real engagement could extend the analysis using information specific to the business and market being studied. Depending on the question, this could include:
actual product costs and operating expenses
historical prices and sales
transaction or conversion data
competitor and market pricing
customer expectations or survey evidence
These inputs could replace assumptions, refine experimental conditions, or adjust the modeled population where the available evidence supports doing so. Additional interactions could also be tested as new questions emerge.
Limitations
This sample uses a synthetic population designed to compare scenarios and expose relationships. It is not intended to forecast realized customer behavior.
Several model ranges and adjustment functions are experimental constructions informed by research rather than directly estimated behavioral parameters.
The contribution calculations also simplify the business environment. Most pricing experiments assume a $0.25 unit cost and express contribution across 100 hypothetical purchase opportunities. A real business would substitute its actual costs, traffic, conversion, inventory, and operating constraints.
The Price Sensitivity range is bounded at approximately $0.10 to $10.00. Prices outside that range are not extrapolated.
These limitations identify which parts of the analysis come from observed evidence and which remain modeling assumptions that could be replaced with better business-specific information.
Conclusion
This investigation began with a simple question:
When does a price increase create more value than it costs?
The experiments show why the answer cannot be determined from the price increase alone. The effect changes with the starting price, the number of modeled purchases lost, the margin earned on the purchases that remain, the cost of the product, and how customers interpret the price relative to their expectations.
Under some conditions, a higher price reduces modeled purchase rate while increasing modeled contribution. Under others, the lost purchases outweigh the additional margin.
The same price can also produce different modeled responses when customer expectations or price-quality signals change.
Rather than producing one universal answer, the model provides a way to identify which conditions determine the tradeoff and test how the answer changes when those conditions change.
Sources and Research Foundation
The sources below inform the behavioral mechanisms, experimental ranges, and assumptions used in this study. They do not directly estimate every synthetic threshold, multiplier, or calibration used in the model. The resulting outputs are experimental comparisons rather than forecasts of observed customer behavior.
Price-Range Anchors
Sam’s Club — Member’s Mark Purified Water, 16.9 fl oz, 40 pack.
Walmart — Great Value Purified Drinking Water, 16.9 fl oz, 40 count.
Axios — Why a water bottle is $3 at Atlanta’s World Cup stadium but $8.50 in Dallas.
These references were used to establish real-world examples of bottled-water prices across substantially different purchasing settings. The model extends the upper experimental range to $10 rather than treating the highest observed price as a hard behavioral threshold.
Water Quality and Beverage Temperature
Carneiro, R. C. V., et al. (2021). Check-if-apply approach for consumers and utilities to communicate about drinking water aesthetics quality. Science of the Total Environment.
Boulze, D., Montastruc, P., & Cabanac, M. (1983). Water intake, pleasure and water temperature in humans. Physiology & Behavior.
Burdon, C. A., Johnson, N. A., Chapman, P. G., & O’Connor, H. T. (2012). Influence of beverage temperature on palatability and fluid ingestion during endurance exercise: a systematic review. International Journal of Sport Nutrition and Exercise Metabolism.
These sources support treating temperature as a meaningful product condition capable of affecting the experience and palatability of drinking water. They do not establish the four synthetic Quality Sensitivity levels used in the model.
10.1016/0031-9384(83)90044-6. PubMed
Convenience and Acquisition Effort
Etale, A., Jobin, M., & Siegrist, M. (2018). Tap versus bottled water consumption: The influence of social norms, affect and image on consumer choice. Appetite.
Pramono, A., & Oppewal, H. (2021). Where to refuel: Modeling on-the-way choice of convenience outlet. Journal of Retailing and Consumer Services.
Berry, L. L., Seiders, K., & Grewal, D. (2002). Understanding Service Convenience. Journal of Marketing, 66(3), 1–17.
Together, these sources support the broader role of convenience, time, effort, and access in customer choice. The model’s ten-minute acquisition-time range remains an experimental construction rather than an empirically estimated bottled-water threshold.
Berry, Seiders, and Grewal specifically develop service convenience around the time and effort consumers expend in acquiring and using a service. DOI
Expected Price and Reference-Price Effects
Putler, D. S. (1992). Incorporating Reference Price Effects into a Theory of Consumer Choice. Marketing Science, 11(3), 287–309.
This research supports distinguishing the price a customer is willing to pay from the price the customer expected to encounter. The study uses that distinction conceptually when Expected Price adjusts the modeled response to an offer.
Putler’s work formally incorporates reference-price effects into consumer choice and empirically examines how deviations from a reference price can affect behavior. PubsOnline
Price as a Quality Cue
Palma, D., et al. (2016). Modelling choice when price is a cue for quality: a case study with Chinese consumers. Journal of Choice Modelling.
This research supports the broader mechanism that price can sometimes communicate information about perceived quality rather than functioning only as a monetary cost.
The strength, direction, and shape of that relationship depend on context. The Price as Quality levels used in this sample therefore remain synthetic experimental settings rather than estimated population parameters.
Price Fairness
Kahneman, D., Knetsch, J. L., & Thaler, R. H. (1986). Fairness as a Constraint on Profit Seeking: Entitlements in the Market. American Economic Review.
Campbell, M. C. (1999). Perceptions of Price Unfairness: Antecedents and Consequences. Journal of Marketing Research.
Tarrahi, F., Eisend, M., & Dost, F. (2016). A meta-analysis of price change fairness perceptions. International Journal of Research in Marketing.
These sources support treating perceived justification and fairness as relevant to how customers respond to prices and price changes. The model does not assume that any one fairness condition corresponds to a measured percentage change in customer behavior. Fairness is used as an experimental interpretation variable.
Social Influence and Visible Popularity
Bearden, W. O., Netemeyer, R. G., & Teel, J. E. (1989). Measurement of Consumer Susceptibility to Interpersonal Influence. Journal of Consumer Research.
Tucker, C. E., & Zhang, J. (2011). How Does Popularity Information Affect Choices? A Field Experiment. Management Science.
These sources support the broader idea that social information and visible popularity can influence customer choice.
Social Proof remains a secondary modeled influence in this study. Its experimental settings do not represent a measured bottled-water population distribution.