Case Three: Kristen’s Cookie Company
Kate Johnson

INTRODUCTION
For many students on a college campus, late nights are as common as group projects and comprehensive tests. During these late nights, snacks can help keep a student focused. Kristen’s Cookie Company (KCC), hopes to capitalize on this fact by providing these students with customizable (perceived quality), high quality, freshly made cookies (aesthetic quality) with a quick delivery time (serviceability) in close proximity to the students (serviceability). Ordinarily, students can get cookies in many venues—including grocery stores, restaurants, and bakeries—that are premade for reasonable prices; but these venues may be closed late at night, or are too far from the campus to fit into a student’s busy schedule. For this reason, many students may be receptive to the convenience and fast delivery time of KCC. Additionally, the easy online ordering process will benefit the company as it continues to grow. KCC is positioned to be successful if it can continue to effectively deliver on all of its order winners.

COOKIE PRODUCTION
The production process for making the cookies is shown in the process flow diagram in Exhibit 1. Since the time to preheat the oven is omitted from this case, the total processing time for the first batch of one dozen cookies is 26 minutes.

If a mistake is made on one of the batches of cookies, it will set KCC back by a minimum of 25 minutes. When the first batch is put into the oven to bake, Kristen can clean the mixer, mix the next batch, and spoon the cookies onto the next pan before the first batch is finished baking; thus they will be ready to bake as soon as the first batch comes out of the oven. As soon as her roommate removes them from the oven to cool, she can place the next batch in the oven. While the first batch is cooling, the second batch is in the oven baking. When the first batch finishes cooling, her roommate can pack them in the box and collect the payment all before the second batch is finished baking. Since both Kristen and her roommate can perform these tasks while the cookies bake, the only additional time that will be added to the 25 minutes is the 10 minutes (plus the one minute to collect the payment) it takes to bake the cookies thereby making the total processing time 36 minutes. After the 10 minutes to bake the cookies, if the process has to be redone, then the process will start over as if it is the initial batch of the night making the next processing time 26 minutes; this gives us 36 minutes wasted in the mistake. This assumes that the mistake is made after the initial batch has been made. If, on the other hand, the mistake is made during the initial batch, then the time taken for each step will be added to the total processing time. Exhibit 2 shows the total time it will take to make a new order during the different stages of the process. Thus, the longer it takes for the mistake to be caught, the more time it will take to make a new order (for the initial batch). This affects not only that particular batch, but also pushes back all of the other orders that would have been made starting at the time the batch had to be remade.

CAPACITY AND CONSTRAINTS
Kristen’s Cookies is open for four hours each night; therefore, 240 minutes are available for production. Because two people are operating the business, the washing, mixing and spooning that is done first can be performed while the previous batch is cooking, as can the cooling, packing, and payment. Therefore, after the initial 26 minutes, the remaining batches could take only 10 minutes processing time, allowing KCC to process 22 batches of cookies in one night assuming one dozen cookies per batch (the throughput table and calculations are shown in Exhibit 3). This also assumes that Kristen and her roommate work continuously through the four hour period without taking any breaks.

KCC’s production process is greatly hindered by its bottleneck: baking. Only one tray can fit into the oven at a time and baking is the longest cycle time in the process. This constraint limits the total number of batches that Kristen and her roommate can produce in the four hour span. Even if the preparation stages performed by Kristen and the finishing stages performed by her roommate could be reduced, it would only affect the time taken to make the initial batch. Since these stages can continue while the cookies are baking, reducing the cycle time of these stages does not affect the overall time taken to make any subsequent batches of cookies. In order to reduce the total processing time, the baking process must be shortened. This could be accomplished by using multiple racks or adding an extra oven if the oven could not support multiple racks. Doing so would allow Kristen and her roommate to make more than one batch at one time. The reduction time and increased capacity of adding an extra oven will be discussed in greater detail in the “Minimizing Constraints” section.
Given the processing time and the process constraint, Kristen must determine a target capacity cushion. Typically, large capacity cushions are advised for businesses with uncertain demand. KCC is facing uncertain demand for two reasons. The first is that the company is new, so while Kristen may have predictions, she cannot be sure of the future demand. Second, the demand may change on a nightly, weekly, or monthly basis. When KCC has been in operation for several weeks, a demand pattern may emerge allowing Kristen to reduce the capacity cushion by a small amount; however if it remains uncertain, a larger capacity cushion is advised. Since KCC is a mixed order company, a larger capacity cushion is also advisable. On the other hand, since the work process is a pull method, Kristen may choose to lessen the capacity cushion as the cookies will be made to order, eliminating the potential to produce more than the demand. Thus, the capacity cushion will be primarily to allow for mistakes made on any of the orders. Therefore, the advisable capacity cushion should be 9% (or the equivalent of two batch orders).

FULFILLING LARGE BATCH ORDERS
The table in Exhibit 4 depicts the time required to fill orders of one, two, or three dozen cookies. Due to the cycle time of the bottleneck, this is the length of the entire process after the initial batch of 26 minutes. Because of this bottleneck, reducing the time of Kristen’s preparation by creating a larger batch has no effect on the total processing time. Likewise, reducing the finishing stages of cooling, packing, and paying will also have no effect. The only element in the process that would directly affect the time to make multiple dozen orders would be to reduce the baking time.

DISCOUNTS
Assuming KCC is not operating at full capacity, Kristen may have to begin offering discounts. Because the company can only produce one batch of cookies at a time, it must be careful not to offer discounts that will encourage customers to order dozens of cookies at a time. The scheduling for an influx in orders at the same time will cause a big problem in scheduling. Currently, the company promises that orders will be ready for pickup within an hour. This can become very difficult to do when there are more than four orders that come in at the same time. For example, if five different customers order one dozen cookies at opening of the production window, the first order will be completed in 26 minutes and be available within in the promised 60 minute window. The second order will be completed in 36 minutes, the third in 46 minutes and the fourth in 56 minutes (assuming no mistakes) which still remains within the allotted time frame. However, the fifth order will take 66 minutes, which falls outside the promised one hour guarantee.

For this reason, Kristen should offer specials related to the time period ordered rather than trying to increase the total number of batches ordered per customer. For example, customers who order their cookies 24 hours in advance will receive a 5-10% discount. The advantage of this discount is that it improves the scheduling of the orders rather than encourages customers to order more batches of cookies at one time (the scheduling will be discussed in greater detail in the “Scheduling” section). Since the variable cost amounts will change depending on the number of added ingredients, a dollar amount discount is not advised. A discount percentage will better correlate with the variable costs. It should also be a low percentage as the total costs will not change significantly with the preorders (because the advantage is in the scheduling of the process), yet, the discount should also be significant to encourage customers to preorder. Therefore, a 5-10% discount is advised, depending on the price KCC sells the cookies and the labor costs associated with making the cookies.

MINIMIZING CONSTRAINTS
Adding another oven would benefit Kristen’s Cookies. Since this is the bottleneck, an additional oven will increase the number of batches the company can make in one night. The process flow chart of adding another oven is shown in Exhibit 5. The initial process will take 26 minutes just as when baking with one oven. However, since the second batch can begin baking while the first batch is in the oven, only five minutes will be necessary for each additional batch. The baking time for each oven will still be ten minutes, but with two ovens, the cycle time for baking is reduced to five minutes. The table for two ovens is presented in Exhibit 6. Due to the reduction in baking cycle time, the mixing/spooning process becomes the new constraint, with a cycle time of eight minutes. This means that only 7.5 batches can be produced in an hour—excluding the first hour with the first batch—meaning that 28 batches are now available with two ovens. While this increases the total number of batches available by six, renting the extra oven may not be optimal. If the cost to rent the oven exceeds the revenue of the six additional batches, then adding an extra oven at this point is not advised. If, however, the revenue exceeds the cost of rental, KCC may want to rent the oven to increase capacity. Additionally, with the new constraint of Kristen’s tasks, hiring another person might be advisable (again depending on the additional cost and revenue). But it stands to reason that as capacity increases, so does the workforce.

Finally, with this new constraint, encouraging multiple dozen orders would benefit KCC as time could be reduced by six minutes if multiple dozens are ordered. Since cleaning and mixing takes six minutes, if it only needs to be performed once for three batches, this will reduce the constraint, making baking and cooling the new constraints. In this case, the total number of batches produced each night would increase. Discounting to encourage multiple dozen orders would be beneficial if another oven is added to the process.

THE WORKFORCE
Currently, Kristen and her roommate work together to make the cookies, however Kristen is considering handling the business by herself. With the process as is, Kristen needs her roommate to perform the finishing tasks while the cookies bake, as Kristen is busy with the preparation stages. If she were to fire her roommate, the cycle time would increase to 18 minutes (10+5+2+1) with only 3.3 batches possible per hour (60/18). This would reduce the number of batches in four hours to 12 batches ([240-8=232]/18 minutes cycle time). Thus, firing her roommate would reduce KCC’s productivity by more than half. This could be detrimental to KCC. The revenue she would see would likely not cover the cost of losing an additional worker.

Kristen must also consider other factors associated with operating alone. First, if demand begins to grow quickly, Kristen will have to determine whether she would like to expand the business or turn down customers. The typical response would be to expand in order to meet increased demand. If she did so by increasing capacity, KCC would benefit by having an additional person to handle one of the ovens. Otherwise, she may barely decrease processing time, making the additional oven more costly than beneficial. Second, Kristen must predict how this will affect her business and her personal relationship with her roommate. If she fires her roommate and her roommate goes into another venture using the oven, they will have to split that time, reducing the amount of batches that Kristen will be able to make for customers. For these reasons, firing her roommate is not recommended.

SCHEDULING
Kristen promises a 60 minute delivery time. This could pose a problem to scheduling if numerous orders come in at one time or if any one of the orders has a rush delivery. Since KCC can only produce up to 6 batches per hour, receiving five orders at the beginning of the night would prevent KCC from meeting the promised 60 minute delivery time for at least one of them. Additionally, if another order is received ten minutes later, that order would also not be ready within the promised allotted time. As it is now, a first-come, first-served policy must be followed to try and maintain delivery within 60 minutes; however, as demand increases, a 60 minute delivery time is not sustainable. The 60 minute promise will hinder the company more than help it if KCC cannot follow through.

A more pragmatic solution to scheduling orders is to ask students to place them in advance. This would give KCC the opportunity to schedule similar orders together, making larger batches of the same cookie if multiple orders are placed for one type of cookie. KCC could then give a specific time to the customer of when the batch of cookies will be ready, still baked fresh. This eases the strain on the scheduling process as well as ensures customers’ cookies are baked fresh within the promised time. The use of discounts (as discussed earlier) will help encourage advanced orders. Changing the ordering and scheduling process will allow KCC to eliminate the risky 60 minute promise.

CONCLUSION
KCC allows customers to have freshly baked cookies while they study late at night. Kristen’s mission is strong, but she must focus on her strategy and long-term goals before making any decisions. Keeping her roommate as a coworker will allow KCC to expand in the future as well as ensure that the current processing system remains capable of producing more batches within the nightly four hour limit. She must also address the scheduling issues so as to allow punctual delivery times, as well as to better organize batches with similar ingredients. Kristen must fully understand the company’s capabilities and constraints in order to make sound decisions for the future of KCC. If demand increases past current capabilities, she will have to choose between renting another oven versus renting a new space to accommodate the rise in demand. If she eliminates or minimizes the constraint of the oven, she must then work to minimize the new constraint. Kristen must align her short-term decisions with her long-term goals to make KCC successful.

Exhibit 1: Process Flow Diagram with One Oven

Exhibit 2: Extra Time for Mistakes in Baking
Stage in Which the Mistake is Caught Time Used up to Mistake Time to Restart/Redo the Process (not including payment) Total Time Taken
After Mixing 6 minutes 26 minutes 32 minutes
After Spooning 8 minutes 26 minutes 34 minutes
After Baking 18 minutes 26 minutes 44 minutes
After Cooling 23 minutes 26 minutes 49 minutes
After Boxing 25 minutes 26 minutes 51 minutes

Exhibit 3: Number of Batches Possible in Four Hours with One Oven
One Oven Mix/Spoon Bake Cool Pack and Pay Total in Four
Hours
Throughput
Time 6+2=8
Minutes 10
Minutes 5
Minutes 2+1=3
Minutes 224**
Minutes
Hourly
Capacity 7.5
Batches 6*
Batches 12
Batches 20
Batches 22***
Batches
Cycle
Time 8
Minutes 10
Minutes
5
Minutes 3
Minutes 10
Minutes per
Batch****
*Since baking is the constraint with the longest cycle time, it limits the number of batches that can be made in one hour. It can only produce 6 batches per hour (excluding the first hour with the initial batch of 26 minutes); 60/10=6 batches possible in one hour.
**4 hours * 60 minutes = 240 minutes available per night 240 minutes – 16 minutes (for the extra task times of the initial batch) = 224 minutes for baking
***224 minutes/ 10 minutes per bake = 22.4 batches possible in four hours (round down to 22 batches)
****After the initial batch of 26 minutes

Exhibit 4: Total Processing Time for Baking One, Two, and Three Dozen Cookies Kristen’s Time Roommates Time Total Time
One Dozen 8 minutes 18 minutes 26 minutes
Two Dozen 10 minutes 10 minutes 36 minutes*
Three Dozen 12 minutes 10 minutes 46 minutes*
*The additional time that Kristen takes does not add to the total time because she can spoon the next dozen cookies onto the next cookie sheet as the previous dozen is baking. The roommate’s time after the initial batch is reduced to ten minutes because she can perform the finishing tasks while the previous batch is baking. Thus, the only task that matters after the initial batch is baking.
Exhibit 5: Process Flow Diagram with Two Ovens

Exhibit 6: Number of Batches Possible in Four Hours with Two Ovens
Two Ovens Mix/Spoon Bake Cool Pack and Pay Total in Four
Hours
Throughput
Time 6+2=8
Minutes 10 Minutes x 2 ovens 5
Minutes 2+1=3
Minutes 227**
Minutes
Hourly
Capacity 7.5*
Batches 12
Batches 12
Batches 20
Batches 28***
Batches
Cycle
Time 8
Minutes 5
Minutes
5
Minutes 3
Minutes 8
Minutes per
Batch****
*Due to the extra oven, Mixing/Spooning becomes the new constraint (since Kristen is doing both); 60/8=7.5 batches available in one hour.
**4 hours * 60 minutes = 240 minutes available per night 240 minutes – 13 minutes = 227 minutes
***227 minutes/ 8 minutes = 28.38 batches possible in four hours (round down to 28 batches)
****After the initial batch of 26 minutes.