DISTRIBUTOR SELECTION ON THE IMPACT OF DEMAND FOR COFFEE PRODUCTS: AHP – SINGLE EXPONENTIAL SMOOTHING

Abstract

The purpose of this study is to assess the performance of suppliers based on the AHP method at the highest weight level, the consistency level of supplier performance based on the smallest consistency value and predict demand with the selected value in the conversion in the kilogram model. The research methodology is quantitative integration of AHP- Single Exponential Smoothing. The data of this study is primary data covering the AHP of the questionnaire, secondary data covering the data of actual requests. This study states that the performance of suppliers includes Quality, Cost and Delivery with consistency with the smallest criteria, namely Delivery, Quality and Flexybility. Meanwhile, the demand prediction with a capacity of 1336 cups of arabica coffee was converted to a capacity of arabica coffee beans of 27 kg in April 2022. The implications of this study are expected to be carried out in determining the dumping factor is experiment with a dumping factor decision-making model that is adjusted to the needs of the TKP Coffee Shop. The suggestion of this study for researchers can then determine the estimated capacity of safety supplies and an economical ordering model.

Full text

136
Airlangga Journal of Innovation Management, Vol.3, No.2, October 2022
Vol. 3 No.2 October 2022
e-ISSN: 2722-5062 DOI：10.20473/ajim.v3i1.39655
DISTRIBUTOR SELECTION ON THE IMPACT OF DEMAND FOR COFFEE
PRODUCTS: AHP – SINGLE EXPONENTIAL SMOOTHING
Johan Alfian Pradanaa*, Rizki Puspita Dewantib, Mohamad Fauzin Abdullohc,
Andrean Pradana Hidayatd
a Master of Industrial Engineering, Institut Teknologi Adhi Tama Surabaya
b Agribusiness Vocational School, Universitas Sebelas Maret
c Industrial Engineering, Institut Teknologi Mojosari Nganjuk
d Industrial Engineering, Universitas Kadiri
*Corresponding Author: zoehuntz34@gmail.com
ABSTRACT
The purpose of this study is to assess the performance of suppliers based on the AHP method at the
highest weight level, the consistency level of supplier performance based on the smallest consistency
value and predict demand with the selected value in the conversion in the kilogram model. The research
methodology is quantitative integration of AHP- Single Exponential Smoothing. The data of this study
is primary data covering the AHP of the questionnaire, secondary data covering the data of actual
requests. This study states that the performance of suppliers includes Quality, Cost and Delivery with
consistency with the smallest criteria, namely Delivery, Quality and Flexybility. Meanwhile, the
demand prediction with a capacity of 1336 cups of arabica coffee was converted to a capacity of arabica
coffee beans of 27 kg in April 2022. The implications of this study are expected to be carried out in
determining the dumping factor is experiment with a dumping factor decision-making model that is
adjusted to the needs of the TKP Coffee Shop. The suggestion of this study for researchers can then
determine the estimated capacity of safety supplies and an economical ordering model.
Keywords: Analytical Hierarchy Process, Exponential, Distributor Selection, Coffee Shop Business.
1. Introduction
Coffee production from 2017 to 2019 was fairly rapid in Indonesia. Export activity reached 8.65
thousand tons or reached 1.17% of the total production. Dominant export activity in the arabica type
where the Asian, Australian, American and European continents are the main export commodity
continents (BPSStatistik, 2017). Connoisseurs of brewed coffee derived from arabica types almost all
predominantly consume from adolescence to adulthood (Edelmann et al., 2022). One of the Coffee
Shops, namely the TKP Coffee Shop, is a briefing place for customers who want to just enjoy Arabica
coffee and do formal agenting. The demand level for arabica coffee over a period of 36 months is
predominantly more than 8000 cups daily. Therefore, the level of importance of this research is to

137
Airlangga Journal of Innovation Management, Vol.3, No.2, October 2022
optimize the quality of arabica coffee, suppliers with certain criteria are needed and inventory capacity
is needed every month. The role of the supplier is very important in order to provide arabica coffee bean
products that are in demand by customers and the role of the right inventory to meet customers who
visit the TKP Coffee Shop. The role of suppliers and coffee shops has a communication role that must
be maintained to provide the right inventory capacity so that customers always visit (Alfian Pradana et
al., 2020).
The issue of this study is how important the criteria must be met by suppliers and how much
capacity the Coffee Shop provides to meet the supply of arabica coffee for customers. This research has
never been done by anyone because the research that will be carried out tries to integrate the AHP and
Single Exponential Smoothing methods. Every month arabica coffee suppliers are not optimal in
carrying out performance. This happens with several cases, namely defects in arabica coffee bean
products, high costs of transportation activities on weekends, estimated delivery times hampered due to
weather and there are delays in information services, the expectations of the owner of the TKP Coffee
Shop have not been able to be met and the information conveyed still has inappropriate communication.
From these factors, the supply of arabica coffee beans is automatically late and customers who come
will complain because the coffee ordered is not fulfilled. This incident occurred within this 6-month
period. To suppress the occurrence of current cases, it is important to set supplier criteria to reduce
delays, improve supplier performance and predict customer requesting needs for arabica coffee
consumption at the TKP Coffee Shop to suppress the occurrence of current cases.
Findings of (Desha Aguslian Bermano & Gustian, 2021; Jawak & Sinaga, 2020; Saputra &
Novita, 2021), states that quality, delivery, service and price as criteria with the highest vector
eigenvalues are shipments. It means that delivery is the most important performance benchmark. While
the statement (Ahmad et al., 2022), the most important criterion is quality because quality as a weighing
of product inventory. In contrast to the statement (Baroto & Utama, 2020), the level of price importance
as a support in the classification of getting high profits from the sale of coffee products. In contrast to
the expression (Mario et al., 2018), The selected classification of suppliers who are at the average level
is delivery. Where the role of delivery is closely related to the estimation of inventory time which will
be the fulfillment of customer needs.
AHP is more complex to discuss suppliers with criteria and sub criteria. Therefore, the subcriteria
used in this study is an advantage role from the previous findings (Desha Aguslian Bermano & Gustian,
2021). From the level of importance of the supplier criteria, it is used to achieve a large number of
predictions resulting from the Single Exponential Smoothing method to meet the needs of arabica coffee
in the future. After the criteria are met, there will be a target of meeting the needs of arabica coffee to
achieve the target according to the calculation of the Single Exponential Smoothing method. Findings
(Endra & Laurina, 2021), states that the estimated prediction is 53,262 cups daily. This became the
benchmark for the least inventory. While the findings (Amelia et al., 2019), states that the Mean
Absolute Percentage (MAPE) level has a significant impact that the smaller the MAPE the predicted
value is worth using. Research (Deina et al., 2021), estimated predictions over time will decrease if the
data history decreases. This means that the predicted estimates will be less than the actual data.
Based on the issues and findings of the predecessor, the discussion of supplier selection and prediction
of dominant demand is not one finding. Rather, it is split from each article. Therefore, this is a research
gap that is the development of AHP-Single Exponential Smoothing integration. The scope of this study
is 1 kg arabica coffee beans capable of producing 50 cups of arabica coffee, research observation data
for 36 months, AHP criteria model using respondents, namely Owner, Barista and Manager at TKP

138
Airlangga Journal of Innovation Management, Vol.3, No.2, October 2022
coffee shop. The integration of these two methods as a plan answers the research objectives, namely (1)
the performance of suppliers based on the AHP method at the highest weight level (2) The level of
consistency of supplier performance is based on the smallest consistency value. (3) predicted demand
in April 2022 with the selected value in the conversion in the kilogram model.
2. Literature Review
Analytical Hierarchy Process (AHP)
AHP is a method that serves for decision-making on the condition that the respondent is more
than 1 person (Siregar et al., 2019; Tirkolaee et al., 2021). In addition, the AHP method does not require
the level of validation and reliability of the criteria used (Jawak & Sinaga, 2020; Nguyen et al., 2021;
Onainor, 2019). The role of AHP has a hierarchical design with the connection of paired comparison
matrices. Where this role can be made a correlation between the criteria and sub-criteria used (Duong
et al., 2018).
Single Exponential Smoothing
Single Exponential Smoothing has a function as a prediction of time series data in the future
(Fauzi et al., 2020). The use of Single Exponential Smoothing as a prediction with the condition of the
use of dumping factor. The dumping factor used ranges from 0.1 to 0.9(Deina et al., 2021). However,
some of the findings are different. This is in terms of the aspect of research assumptions (Duong et al.,
2018; Novanda et al., 2018; Rincón et al., 2020). Thus, the specified dumping factor is expected to
suppress the occurrence of excess coffee bean products. If there is an excess of arabica coffee bean
products, the estimated budget every month can increase (Harijanto et al., 2020; Zhu et al., 2022)
3. Method
The research design uses quantitative integration of AHP- Exponential Smoothing. AHP function
is to determine the supplier of coffee beans according to criteria and sub-criteria. Meanwhile,
Exponential Smoothing to predict the demand for coffee beans uses dumping factors 0.8, 0.5 and 0.2.
There is a research population, namely for AHP parties involved in CoffeeShop TKP. The selected
samples were Baristas with 6 years of experience (weight 0.5) as respondent 1, Owner with 6 years of
experience (weight 0.3) as respondent 2, Manager with 4 years of experience (weight 0.2) as respondent
3. As for Exponential Smoothing, it uses historical data on the demand for coffee beans for 36 months
with saturated sample types. Primary data research instruments for AHP questionnaires with criteria of
Quality (X1), Cost (X2), Flexibility (X3), Delivery (X4), Responsiveness (X5), and Reliability (X6).
While the secondary data for Exponential Smoothing is reputable demand history and scientific literacy
data on the topics of demand management and decision-making.
Table 1.
Operational Constructs
Constructs
Definition
Items and sub-Items
Measurement
Scale
Supplier
The party in
charge of
providing
product supply
for users
1. Quality (broken coffee beans, reddish-
brown coffee bean color, and smell
like medicine) (Ahmad et al., 2022),
2. Cost (negotiation, wholesale and
include expedition) (Barolo & Utama,
2020)
AHP Method
Nominal

139
Airlangga Journal of Innovation Management, Vol.3, No.2, October 2022
3. Flexibility (timing suitability,
purposefulness, and quality adaptation)
4. Delivery (inventory, information, time)
(Desha Aguslian Bermano & Gustian,
2021),
5. Responsiveness (policy, transparency,
and terms of service) (Mario et al.,
2018),
6. Reliability (fulfilled expectations,
trusted and guaranteed service)
Forecasting
Product
management
function of user
demand and
guaranteeing
reliable delivery
estimates
Exponential smoothing (Fauzi et al., 2020)
Dumping
factor 0.8; 0.5
and 0.2
Nominal
The research data analysis technique uses AHP within the following stages:
AHP uses a decision model with the following stages (Handayani et al., 2018):
1. Defining problems and goals
2. Drawing up a hierarchy of problems taken from criteria and subcriteria.
3. Creating a comparison matrix in pairs with weighting so that the level of importance of alternatives
to the criteria is clearly stated.
Table 2.
AHP Scale
Importance level
Information
9
Absolutely more influential
7
Very more influential
5
More influential
3
A little more influential
1
Equally influential
2, 4, 6, 8
Values in between
Source: (de Felice et al., 2015)
Decision makers by assessing the degree of importance of the elements in the following matrix model:
  󰇟        󰇠 …1)
4. Calculate the geometric average to get the stump results from several respondents. The geometric
mean formula is:
󰇛   󰇜 

  
 …2)
Information:
󰇛󰇜 = geometric mean
= the value that each respondent gives in the comparison
 = respondent weights
5. Determining priorities by arranging problem elements at the hierarchy level to determine the
normalized value:

140
Airlangga Journal of Innovation Management, Vol.3, No.2, October 2022

  for j=1,2, … n …3)
Information:
= the number of elements in the j-th column
Elements – elements on the matrix divided by , to obtain normalization. Then search for vector
weights with average using the formula:





 for i = 1,2, …n …4)
If the correlational comparison is complete, then the eigen vector is searched by the formula:
      …5)
Information
A = pairwise comparison matrix
= the largest eigen vector of A
6. Calculate the consistency ratio by providing a numeric value with a consistency index using the
formula:
 
 …6)
Information:
N = matrix size or the number of items compared
7. If CI is worth 0, it is declared a consistent matrix. Inconsistent limits use the consistency ratio (CR)
or the comparison of the consistency index to the random index (RI) with the formula:
 
 …7)
8. Ranking with the highest weight.
Exponential Smoothing
The steps of exponential smoothing is done as follows
a. Determine the alpha value used for forecasting with the following formula:
  
 …8)
Information:
 = Information
n = data on the multiplicity of periods
b. Calculating errors in forecasting using (MAD) Mean Absolute Deviation, as a measurement of the
average error in guessing using the formula (Indrasari, 2020) :
 
 󰇛
󰆓󰇜
 …9)
Information:
Tt = data requests in the period t
Y’t = forecast value in the period t
n = data on the multiplicity of periods
c. Calculating the error squarely (Mean Square Error) using the following formula:
  󰇛
󰆓󰇜
 …10)
󰇛 󰆒󰇜 = (demand data on period t – forecast value of period t)2
N = data on the multiplicity of periods

141
Airlangga Journal of Innovation Management, Vol.3, No.2, October 2022
d. Calculate relative forecasting errors, using MAPE (Mean Absolute Percentage Error) assuming a
percentage of the error value according to actual data that is too high or too low using the formula
(Sungkawa & Megasari, 2011) :
   







 …11)
Information:
Yt = actual value in the period t


 = forecast value in the period t
 = data on the multiplicity of periods
e. Perform forecasting calculations using SES (Single Exponential Smoothing), due to unstable data
patterns using formulas (Indrasari, 2020) :
  󰇛  󰇜 …12)
Information :
Xt = data requests in the period t
α = smoothing contanta (0,1 < x< 0,9)
Ft+1 = forecasting for the period t+1
4. Result and Discussion
Based on the processing of research data for the AHP – Exponential Smoothing method, the
following discussion results were obtained.:
Table 3.
Normalized Eigen Vektor Criteria AHP
Quality
Cost
Flexibility
Delivery
Responsiveness
Reliability
Total
Eigen
vector
Quality
0.240
0.306
0.251
0.161
0.220
0.252
1.429
0.238
Cost
0.148
0.188
0.325
0.198
0.205
0.126
1.191
0.198
Flexibility
0.138
0.083
0.144
0.284
0.155
0.181
0.985
0.164
Delivery
0.240
0.153
0.082
0.161
0.222
0.157
1.015
0.169
Responsiveness
0.138
0.116
0.117
0.091
0.126
0.181
0.769
0.128
Reliability
0.098
0.153
0.082
0.105
0.072
0.103
0.612
0.102
Source: Data Process Result (2022)
Weighting of the criteria with the highest score on the criteria of Quality rank 1, Cost rank 2
and Delivery rank 3. Each with a vector eigenvalue of 0.238; 0,198; and 0.169. The inconsistency used
was 0.005 with an estimated accuracy of 0.01.

142
Airlangga Journal of Innovation Management, Vol.3, No.2, October 2022
Figure 1. Normalized Eigen Vector Criteria
Eigen vector weighting as a normalization with the highest role on Quality, Cost and Delivery criteria.
Table 4.
Weighting of AHP Criteria and Sub-Criteria
Criteria
Weight
Rank
Sub Criteria
Partial
Weights
Global Weights
Quality
0.2381
1
Cracked coffee beans
0.4233
0.1008
The color of the coffee beans is
reddish brown
0.3055
0.0727
Smells like medicine
0.2712
0.0646
Cost
0.1984
2
Negotiated pricing
0.3484
0.0691
Wholesale prices
0.4181
0.0830
Price includes expedition costs
0.2334
0.0463
Flexybility
0.1641
4
Bookings are delivered on time
0.3524
0.0578
Bookings are delivered as intended
0.3953
0.0649
Quality adaptation
0.2523
0.0414
Delivery
0.1692
3
Ordering supplies
0.4632
0.0784
clear booking information
0.2880
0.0487
Estimated delivery time
0.2488
0.0421
Responsivness
0.1282
5
Thoughtful clarity
0.3551
0.0455
Transparent and accountable
information
0.3701
0.0474
Terms of service are met
0.2748
0.0352
Reliability
0.1020
6
Fulfilled expectations
0.6432
0.0656
Trusted
0.5543
0.0565
Guarantee the best service
0.4484
0.0457
Source: Data Process Result (2022)
The determination of Quality with rank 1 has a weight value of 0.2381 with the dominant sub-
criterion being broken coffee beans. The dominant role of broken coffee beans is stated as the failure
of suppliers who are declared unfit to be expeditioned to the TKP coffee shop. Therefore, the importance
of the condition of coffee beans is a strict consideration. Cost determination with rank 2 has a rank of 2
with a weight of 0.1984 with the dominant sub-criterion being wholesale prices. The role of wholesale
prices is very influential for resale at TKP coffee shop. Therefore, suppliers provide wholesale prices
as an opportunity for TKP coffee shops to increase the role of high profits from time to time. Delivery

143
Airlangga Journal of Innovation Management, Vol.3, No.2, October 2022
determination with rank 3 has a weight value of 0.1692 with the dominant sub-criterion being order
inventory. Ordering inventory is the spearhead which plays an important role in meeting the supply of
coffee beans with quality except for broken coffee beans. Therefore, TKP coffee shops that have
customers every day an average of 100 visitors need to consider the estimated capacity of coffee bean
supplies.
Flexybility determination that dominates the sub-criteria is order sent according to the destination.
Orders sent according to the purpose are one of the missions carried out by suppliers to provide hope
when needed related to the supply of coffee beans at TKP coffee shop. Furthermore, the dominating
Responsiveness is the sub-criteria of transparent and accountable information and the dominating
Realibility is the sub-criteria of Meeting expectations.
Table 5.
AHP Consistency Test
Sigma VB
Lambda max
CI
CR<0,1
Decision
Rank
Quality
9.030
3.010
0.005
0.010
Consistent
5
Cost
9.103
3.034
0.017
0.033
Consistent
3
Flexybility
9.040
3.013
0.007
0.013
Consistent
4
Delivery
9.040
3.005
0.003
0.005
Consistent
6
Responsivness
9.207
3.069
0.035
0.066
Consistent
1
Reliability
9.161
3.054
0.027
0.052
Consistent
2
Source: Data Process Result (2022)
Determination of the consistency test value as the most important parameter. Where the criteria
that are declared consistent are worthy of use as parameters for determining the criteria for suppliers
who are able to provide the best performance for TKP coffee shops. Based on the consistency test of
the Criteria of Quality, Cost, Flexybility, Delivery, Responsivness, Reliability, each has a Consistensy
Ratio ≤ 0.1. The assessment of the highest consistency ratio criteria was Responsivness of 0.066, the
second order was Reliability of 0.052 and the third order was Cost of 0.033
Table 6.
Single Exponential Smoothing Kriteria Dumping Factor
Periods
Dema
nd
alpa=0,2
alpha=0,5
alpha=0,8
Foreca
st
MA
D
MSE
MAP
E
Foreca
st
MA
D
MSE
MAP
E
Foreca
st
MA
D
MSE
MAP
E
April
2019
8568
#N/A
#N/A
#N/A
#N/A
#N/A
#N/A
#N/A
#N/A
#N/A
#N/
A
#N/A
#N/A
May
8904
8,904
-
-
-
8,904
-
-
-
8,904
-
-
-
June
9296
8,982
314
98,345
0.034
9,100
196
38,41
6
0.021
9,218
78
6,147
0.008
July
8162
8,818
656
430,756
0.080
8,631
469
219,9
61
0.057
8,373
211
44,57
2
0.026
August
7840
8,623
783
612,550
0.100
8,236
396
156,4
20
0.050
7,947
107
11,36
9
0.014
Septemb
er
7836
8,465
629
396,050
0.080
8,036
200
39,90
0
0.025
7,858
22
490
0.003
October
8321
8,436
115
13,331
0.014
8,178
143
20,34
2
0.017
8,228
93
8,570
0.011
Novemb
er
8135
8,376
241
58,162
0.030
8,157
22
470
0.003
8,154
19
349
0.002
Decemb
er
7963
8,294
331
109,253
0.042
8,060
97
9,379
0.012
8,001
38
1,454
0.005

144
Airlangga Journal of Innovation Management, Vol.3, No.2, October 2022
January
8142
8,263
121
14,696
0.015
8,101
41
1,687
0.005
8,114
28
794
0.003
Februar
y
8019
8,214
195
38,174
0.024
8,060
41
1,678
0.005
8,038
19
360
0.002
March
8034
8,178
144
20,824
0.018
8,047
13
168
0.002
8,035
1
1
0.000
April
2020
8108
8,164
56
3,163
0.007
8,077
31
931
0.004
8,093
15
214
0.002
May
8082
8,148
66
4,329
0.008
8,080
2
5
0.000
8,084
2
5
0.000
June
8036
8,125
89
7,999
0.011
8,058
22
478
0.003
8,046
10
93
0.001
July
8180
8,136
44
1,905
0.005
8,119
61
3,729
0.007
8,153
27
722
0.003
August
7790
8,067
277
76,773
0.036
7,954
164
27,05
0
0.021
7,863
73
5,275
0.009
Septemb
er
8125
8,079
46
2,147
0.006
8,040
85
7,270
0.010
8,073
52
2,754
0.006
October
8062
8,075
13
178
0.002
8,051
11
124
0.001
8,064
2
4
0.000
Novemb
er
7920
8,044
124
15,442
0.016
7,985
65
4,282
0.008
7,949
29
831
0.004
Decemb
er
8323
8,100
223
49,724
0.027
8,154
169
28,48
8
0.020
8,248
75
5,600
0.009
January
7869
8,054
185
34,154
0.023
8,012
143
20,33
7
0.018
7,945
76
5,751
0.010
Februar
y
8378
8,119
259
67,264
0.031
8,195
183
33,56
1
0.022
8,291
87
7,505
0.010
March
7795
8,054
259
67,039
0.033
7,995
200
39,96
1
0.026
7,894
99
9,855
0.013
April
2021
7638
7,971
333
110,712
0.044
7,816
178
31,84
5
0.023
7,689
51
2,627
0.007
May
8149
8,006
143
20,338
0.018
7,983
166
27,64
7
0.020
8,057
92
8,455
0.011
June
8137
8,033
104
10,918
0.013
8,060
77
5,950
0.009
8,121
16
256
0.002
July
8260
8,078
182
33,121
0.022
8,160
100
10,01
4
0.012
8,232
28
773
0.003
August
7756
8,014
258
66,361
0.033
7,958
202
40,79
0
0.026
7,851
95
9,071
0.012
Septemb
er
6916
7,794
878
771,033
0.127
7,437
521
271,4
23
0.075
7,103
187
34,98
7
0.027
October
8652
7,966
686
471,051
0.079
8,044
608
369,0
67
0.070
8,342
310
95,97
0
0.036
Novemb
er
9576
8,288
1,288
1,659,6
28
0.135
8,810
766
586,3
80
0.080
9,329
247
60,89
0
0.026
Decemb
er
6832
7,997
1,165
1,356,2
64
0.170
7,821
989
978,3
64
0.145
7,331
499
249,4
49
0.073
January
6720
7,741
1,021
1,042,9
93
0.152
7,271
551
303,1
18
0.082
6,842
122
14,95
5
0.018
Februar
y
7560
7,705
145
21,030
0.019
7,415
145
20,94
4
0.019
7,416
144
20,60
4
0.019
March
6496
7,463
967
935,501
0.149
6,956
460
211,2
69
0.071
6,680
184
33,89
0
0.028
April
2022
5,970.
57
3,477.
82
1,336.
02
Average
352.6
4
246,320
0.05
214.7
2
100,3
27
0.03
89.6
2
18,41
8
0.01
4.58
2.78
1.16
Source: Data Process Result (2022)
Single exponential smoothing uses 3 types of dumping factor, namely 0.8; 0.5 and 0.2. The value
of this dumping factor with α of 0.2 each; 0.5 and 0.8. Single exponential smoothing uses a total of 36
months of data from April 2019 to March 2022. Thus, the predicted demand forecasting is April 2022.
The april 2022 prediction results have differences from each different dumping factor. The predicted

145
Airlangga Journal of Innovation Management, Vol.3, No.2, October 2022
value of the request is used with the smallest MAPE value proof of the three MAPE values used. The
smallest MAPE value at the dumping factor is 0.2 with a predicted demand of 1336 cups of arabica
coffee. The conversion of 1336 cups into kilograms using the assumption that every 1 kg can be used
for 50 cups of arabica coffee. Thus, the predicted demand for 1336 cups of arabica coffee with an
estimated need for coffee bean capacity of 26.7 ~ rounded to 27 kg of arabica coffee.
Table 7.
Error Rate Comparison
Forecasting
alpa=0,2
alpha=0,5
alpha=0,8
Month
April 2022
5970.6
3477.8
1336.0
MAD
352.6
214.7
89.6
MSE
352.6
100327.1
18418.2
MAPE
4.6
2.8
1.2
Source: Data Process Result (2022)
The comparison of the smallest error rate is at the dumping factor of 0.2 with a α accuracy rate
of 0.8. MAD value of 89.6; and an MSE value of 18418.2. Thus, it can be expressed that the smaller
the value of the α will decrease the MAD and MAPE values, while the MSE value will increase.
(a) Dumping Factor 0,8
(b) Dumping Factor 0,5
(c) Dumping Factor 0,2
Figure 2.
Comparison of Single Exponential Smoothing Criteria Dumping Factor
Source: Data Process Result (2021)
Comparison of single exponential smoothing values based on dumping factor 0.8; 0.5 and 02
have an increased predicted value at dumping factor 0.5 and dumping factor 0.2. Meanwhile, the

146
Airlangga Journal of Innovation Management, Vol.3, No.2, October 2022
dumping factor of 0.8 is predominantly declining. Thus, dumping factor has an influence on forecasting
using the single exponential smoothing method.
Research Discussion
Based on the results of the study, it is stated that the supplier criteria that must be met by the
supplier need to have the highest weight on quality, cost, and delivery. Meanwhile, the expected
consistency with the smallest consistency value close to the value of 0 is delivery, quality, and
felxybility with an inventory of arabica coffee beans of 27 kg every month with an estimated amount
per cup provided 1336 cups of arabica coffee for customers at the TKP Coffee Shop. AHP context
findings answer (Ahmad et al., 2022), where the importance level of the Quality criteria is equal, that
is, the value obtained is 0.37. This means that the quality criteria are the main benchmark of the various
findings and results of this study. Therefore, we provide an assessment of subcriteria as an advantage
over the preliminary findings. It is stated that from the Quality criteria have sub-covers no broken coffee
beans, the color of the coffee beans is reddish brown and smells like medicine. These three sub-criteria
are the responsibility of the supplier to have a consistent commitment to the sustainability of the TKP
Coffee Shop. In addition, the Cost criterion with the sub-criteria of negotiated price, wholesale price,
price including expedition costs is of further weight interest with the dominant value of wholesale price.
This finding is in the same direction as (Baroto & Utama, 2020) with the same price role – equally
having importance. This means that the price has an impact on the sustainability of the TKP Coffee
Shop. Therefore, the higher the compensation given regarding prices by suppliers, it will be a high
opportunity to increase profits and increase the supply of arabica coffee for visitors. This is why the
role of demand prediction needs to be calculated with the Single Exponential Smoothing method.
The calculation selected with the smallest MAPE is 1.2 with a predicted demand of 1336 cups or
27 kg of arabica kopo seeds in April 2022. This prediction becomes a dominating answer and in the
same direction as the findingswith the same price role – equally having importance. This means that the
price has an impact on the sustainability of the TKP Coffee Shop. Therefore, the higher the
compensation given regarding prices by suppliers, it will be a high opportunity to increase profits and
increase the supply of arabica coffee for visitors. This is why the role of demand prediction needs to be
calculated with the Single Exponential Smoothing method. The calculation selected with the smallest
MAPE is 1.2 with a predicted demand of 1336 cups or 27 kg of arabica kopo seeds in April 2022. This
prediction becomes a dominating answer and in the same direction as the findings (Fauzi et al., 2020),
where the MAPE value becomes the selected parameter at the smallest MAPE value. In addition to
MAPE's interests, the use of dumping factors has a high chance of estimating the best amount of
inventory.
The dumping factor used was 0.8; 0.5 and 0.2. So, dumping factor experiments answer the
findings (Rincón et al., 2020), that the dumping factor used was 0.4; 0.2 and 0.3 do not necessarily
contribute to the prediction of high demand, however, this plays more of a role in predicting security in
purchasing products so as not to exceed the costs in the TKP Coffee Shop. The use of dumping factors
is a critical agenda that needs to be deepened in the inventory management model. Where the role of
dumping factor plays a full role in the estimation of demand predictions. Therefore, in further research
after determining the criteria for suppliers, it can be carried out the implementation of dumping factors
with more diverse exponential forecasting models. Thus, it can be found the most fully contributing
pattern to the already estimated demand.

147
Airlangga Journal of Innovation Management, Vol.3, No.2, October 2022
5. Conclusion
This study produces conclusions according to the goals that have been formulated. The
conclusion of this study is (1) the performance of suppliers selected by TKP Coffee Shop with the
highest criteria includes Quality, Cost, and Delivery based on the highest weight. (2) The level of
consistency of supplier performance selected by TKP coffee shop with the smallest criteria, namely
Delivery, Quality and Flexybility. (3) the prediction of demand for using the smallest MAPE with a
capacity of 1336 cups of arabica coffee converted to a capacity of 27 kg arabica coffee beans in April
2022. The implications of this study are expected to be carried out in determining the dumping factor
is experiment with a dumping factor decision-making model that is adjusted to the needs of the TKP
Coffee Shop. In addition, it can be carried out the determination of demand strategies for a sustainable
increase in demand. The suggestion of this study for researchers can then determine the estimated
capacity of safety supplies and an economical ordering model. On the other hand, it can consider a
business model to increase demand capacity and assess how important the criteria that have been
determined by the AHP method are in this finding.
Acknowledgments
Thank you to the TKP Coffee Shop and Coffee Suppliers for participating in this research. We
would like to thank the alma mater of Adhi Tama Institute of Technology Surabaya, Sebelas Maret
University, Mojosari Nganjuk Institute of Technology and Kadiri University for providing academic
knowledge for research provisions.
6. Reference
Ahmad, Ipov, Amelia, M, D., A, K., R, G., & Edward. (2022). Penggunaan Metode Fuzzy Ahp Dan
Topsis Pada Pemilihan Supplier (Studi Kasus: PT. SS). Seminar Nasional Ke- IV Universitas
Tarumanegara, 437–444.
Alfian Pradana, J., Komari, A., & Dewi Indrasari, L. (2020). Studi Kelayakan Bisnis Tell Kopi Dengan
Analisis Finansial. Industri Inovatif : Jurnal Teknik Industri.
https://doi.org/10.36040/industri.v10i2.2855
Amelia, Nurviana, Muliani, F., & Nuri, B. (2019). Forecasting Annual Coffee and Rubber Production
in Aceh Using Exponential Smoothing. Regular Proceeding 3rd ISIMMED, 3–10.
Baroto, T., & Utama, D. M. (2020). Integrasi AHP dan SAW untuk Penyelesaian Green Supplier
Selection. SENTRA: Seminar Nasional Teknologi Dan Rekayasa, 38–44. http://research-
report.umm.ac.id/index.php/sentra/article/view/3895
BPSStatistik. (2017). Statistik Kopi Indonesia. In BPSStatistik-Indonesia (Ed.), BPSStatistik (2nd
ed.). BPSStatistik.
Deina, C., do Amaral Prates, M. H., Alves, C. H. R., Martins, M. S. R., Trojan, F., Stevan, S. L., &
Siqueira, H. V. (2021). A methodology for coffee price forecasting based on extreme learning
machines. Information Processing in Agriculture, xxxx.
https://doi.org/10.1016/j.inpa.2021.07.003
Desha Aguslian Bermano, R., & Gustian, D. (2021). Sistem Pendukung Keputusan Pemilihan Supplier
Di Tentera Coffee Corp Dengan Menggunakan Metode Analytical Hierarchy Process. Jurnal
Rekayasa Teknologi Nusa Putra, 5(1), 13–21. https://doi.org/10.52005/rekayasa.v5i1.97
Duong, L. N. K., Wood, L. C., & Wang, W. Y. C. (2018). Effects of consumer demand, product lifetime,
and substitution ratio on perishable inventory management. Sustainability (Switzerland), 10(5).
https://doi.org/10.3390/su10051559
Edelmann, H., Quiñones-Ruiz, X. F., & Penker, M. (2022). How close do you like your coffee? -
Examining proximity and its effects in relationship coffee models. Journal of Rural Studies,

148
Airlangga Journal of Innovation Management, Vol.3, No.2, October 2022
91(January), 24–33. https://doi.org/10.1016/j.jrurstud.2022.02.007
Endra, R. Y., & Laurina, O. (2021). Aplikasi Prediksi Penjualan Kopi dengan Metode Single
Exponential Smoothing untuk Mengetahui Produk Kopi Terlaris. EXPERT: Jurnal Manajemen
Sistem Informasi Dan Teknologi, 11(2), 129. https://doi.org/10.36448/expert.v11i2.2212
Fauzi, E., Putra, W. E., Ishak, A., & Astuti, H. B. (2020). Pendugaan Model Peramalan Harga Ekspor
Kopi Indonesia. AGRITEPA: Jurnal Ilmu Dan Teknologi Pertanian, 7(1), 22–30.
https://doi.org/10.37676/agritepa.v7i1.1002
Harijanto, B., Ekojono, & Iqbal, M. M. (2020). Pengembangan Sistem Peramalan Produksi Kopi
Menggunakan Metode Triple Exponential Smoothing (Studi Kasus Pabrik Kopi Duoningrat).
Seminar Informatika Aplikatif Polinema, 2020.
Indrasari, L. D. (2020). Penerapan Single Exponential Smoothing (SES) dalam Perhitungan Jumlah
Permintaan Air Mineral Pada PT. Akasha Wira International. JATI UNIK: Jurnal Ilmiah Teknik
Dan Manajemen Industri, 3(2), 87–98.
Jawak, J. B. W., & Sinaga, C. J. S. (2020). Aplikasi Analytical Hierarchy Process (Ahp) Dalam Memilih
Pemasok Pada Ksu Pom Humbang Cooperative. Jurnal Sains Dan Teknologi: Jurnal Keilmuan
Dan Aplikasi Teknologi Industri, 19(2), 123. https://doi.org/10.36275/stsp.v19i2.207
Mario, M., Nurhafidzh, D., & Koesdiningsih, N. (2018). Analisis Pemilihan Pemasok Menggunakan
Metode Analytical Hierarchy Process ( AHP ) untuk Pengambilan Keputusan Pemilihan Pemasok
( Studi Kasus pada Kopi Arjuna di Desa Cibodas Lembang Jawa Barat ) Supplier Selection
Analysis Using Analytical Hierarchy Pro. Prosiding Manajemen, 904–916.
Nguyen, N. B. T., Lin, G. H., & Dang, T. T. (2021). A two phase integrated fuzzy decision-making
framework for green supplier selection in the coffee bean supply chain. Mathematics, 9(16), 1–
21. https://doi.org/10.3390/math9161923
Novanda, R. R., Sumartono, E., Asriani, P. S., Yuliarti, E., Sukiyono, K., Priyono, B. S., Irnad, I.,
Reswita, R., Suryanty, M., & Octalia, V. (2018). A Comparison of Various Forecasting
Techniques for Coffee Prices. Journal of Physics: Conference Series, 1114(1).
https://doi.org/10.1088/1742-6596/1114/1/012119
Onainor, E. R. (2019). Sistem Pendukung Keputusan Berbasis Metode Analytical Hierarchy Process
(AHP) Dalam Pemilihan Biji Kopi Berkualitas. Journal of Information Technology and Computer
Science (INTECOMS), 4, 105–112.
Rincón, R. D., Palacios, W., & Paipa, H. O. (2020). Comparison of statistical forecasting techniques
for Colombian coffee demand in South Korea. Journal of Physics: Conference Series, 1448(1).
https://doi.org/10.1088/1742-6596/1448/1/012023
Saputra, A. F., & Novita, I. (2021). Sistem Penunjang Keputusan Penentuan Supplier Terbaik Dengan
Metode Analytical Hierarcy Process (Ahp) Dan Simple Additive Weighting (Saw) Pada Xy
Coffee and Roastery. IDEALIS : InDonEsiA JournaL Information System, 4(2), 205–214.
https://doi.org/10.36080/idealis.v4i2.2850
Siregar, E., Nazir, N., & Asben, A. (2019). The Analysis of Strategic Partnership to Supply Mandailing
Arabica Coffee for Export Quality Markets. IOP Conference Series: Earth and Environmental
Science, 347(1), 012046. https://doi.org/10.1088/1755-1315/347/1/012046
Sungkawa, I., & Megasari, R. T. (2011). Penerapan Ukuran Ketepatan Nilai Ramalan Data Deret Waktu
dalam Seleksi Model Peramalan Volume Penjualan PT Satriamandiri Citramulia. ComTech:
Computer, Mathematics and Engineering Applications.
https://doi.org/10.21512/comtech.v2i2.2813
Tirkolaee, E. B., Dashtian, Z., Weber, G. W., Tomaskova, H., Soltani, M., & Mousavi, N. S. (2021).
An integrated decision-making approach for green supplier selection in an agri-food supply chain:
Threshold of robustness worthiness. Mathematics, 9(11), 1–30.
https://doi.org/10.3390/math9111304
Zhu, X., Zhang, F., Deng, M., Liu, J., He, Z., Zhang, W., & Gu, X. (2022). Exponential Smoothing and
ELM to Predict Multi-Factor Landslide Displacement. Remote Sensing, 14(3384), 1–20.