Analyzing Business Performance Of Good Harvest Organic Farm And Market

Problem definition and business intelligence required

In this report, we analyse the business performance of a newly established enterprise by the name Good Harvest Organic Farm and Market. The enterprise started its operations about two years ago and their main business proposition and operations entails growing a range of quality fresh organic produce and selling them direct to the local community through its home delivery service.

We are provided with the data from the enterprise which is inclusive of a whole year of trading. This is the second year of business and the business is still in a start-up phase. The reported high Cost of Goods (COGS) is reportedly consistent with the Organic industry.

The main challenges in the business are revenue (i.e. lead generation/new business), Cost of Goods (COGS margins) and average sales.

Using the one year data from the enterprise, we sought to analyze (conducting both descriptive and predictive analytics) so as to gain insight that can be shared with the CEO of the company. In particular we sought to answer the following research questions;

1. What are the top/worst selling products in terms of sales?
2. Is there a difference in payments methods?
3. Are the differences in sales performance based on where the product is located in the shop? How does this effect both profits and revenue?
4. Is there a difference in sales and gross profits between different months of the year?
5. Are their differences in sales performance between different seasons?

Each and every business problem stated above was addressed using appropriate statistical tests. The first step was to ensure that the variables under study are well scaled. The screenshot below presents the first few values of the data analyzed: 

The boxplot shows that median returns are vastly different and that the downside of worst case is limited to total sales not exceeding $50, while the upside in best case yields returns up to almost $20,000. It is however important to note that there are quite a number of outliers both in the worst performing products and best performing products.  

To answer this question, we performed independent t-test. In the dataset there are 4 different payment methods namely; cash payment, credit, Visa and MasterCard payments. Cash and Credit were categorised into one group and an independent t-test was performed test whether there is a significant difference in the total cash received between the two payment methods. Independent t-test is normally used to compare two independent (unrelated) groups (Derrick, et al., 2017). Visa card and MasterCard were also grouped together and an independent t-test performed to compare the total cash received from the use of the two payment methods (John , 2006). We sought to test the following hypothesis;

H₀: There is no significant difference in the total cash received between the cash and the credit payment methods

H₀: There is significant difference in the total cash received between the cash and the credit payment methods

Table 2: t-Test: Two-Sample Assuming Equal Variances

	Cash	Credit
Mean	412.1755	604.6356
Variance	20811.55	42140.48
Observations	359	354
Pooled Variance	31401.02
Hypothesized Mean Difference	0
df	711
t Stat	-14.5002
P(T<=t) one-tail	3.14E-42
t Critical one-tail	1.647
P(T<=t) two-tail	6.27E-42
t Critical two-tail	1.963306

Results from an independent samples t-test showed that total cash received from cash payment method (M = 412.18, SD = 144.26, N = 359) was significantly lower when compared to the credit payment method (M = 604.64, SD = 204.28, N = 354), t(711) = -14.50, p < .001, two-tailed. The difference of 192.46 was large (effect size, d = 1.08). 

H₀: There is no significant difference in the total cash received between the Visa and the MasterCard payment methods

H₀: There is significant difference in the total cash received between the Visa and the MasterCard payment methods.

Table 3: t-Test: Two-Sample Assuming Equal Variances

Selected analytics methods and technical analysis

	Visa	MasterCard
Mean	576.3144	152.5472
Variance	50355.11	12000.98
Observations	353	53
Pooled Variance	45418.44
Hypothesized Mean Difference	0
df	404
t Stat	13.49813
P(T<=t) one-tail	7.81E-35
t Critical one-tail	1.648634
P(T<=t) two-tail	1.56E-34
t Critical two-tail	1.965853

Results of the two-independent samples t-test shows that mean total cash received significantly differs between Visa payment method (M = 576.31, SD = 224.40, n = 353) and MasterCard payment method (M = 152.55, SD = 109.55, n = 53) at the .05 level of significance (t = 13.50, df = 404, p < .05, 95% CI for mean difference 365.02 to 485.48). On average Visa payment method tend to have higher total cash received than MasterCard payment method.

To test this research question, we performed analysis of variance (ANOVA) test. ANOVA tests the null hypothesis that the means of two or more populations are equal (Moore & McCabe, 2003). This test is used to assess the importance of one or more factors by comparing the response variable means at the different factor levels (Howell, 2002). As mentioned, the null hypothesis states that all population means (factor level means) are equal while the alternative hypothesis states that at least one is different (Gelman, 2005). The hypothesis tested in this analysis is;

H₀: There is no significant difference in the mean total sales between the different product locations in the shop

H₀: There is significant difference in the mean total sales between the different product locations in the shop.

Table 4: Analysis of variance (ANOVA) for the total sales versus product location

Total Sales ($)
	Sum of Squares	df	Mean Square	F	Sig.
Between Groups	134299725.02	4	33574931.256	37.176	.000
Within Groups	929333380.82	1029	903142.255
Total	1063633105.84	1033

We conducted a one-way between subjects ANOVA so as to compare the mean total sales received based on the locations of the product. Table 4 above presents the ANOVA summary. As can be seen, there is significant differences in the mean total sales of the products based on the location of the product at the p<.05 level for the five conditions [F(4, 1029) = 37.18, p = 0.000].

Table 5: Mean total sales based on product location

	N	Mean	Std. Deviation	Std. Error	95% Confidence Interval for Mean
Lower Bound	Upper Bound
Front	155	572.75	1430.657	114.913	345.74	799.76
Left	376	218.22	427.614	22.053	174.86	261.58
Outside Front	12	3384.37	4719.347	1362.358	385.84	6382.90
Rear	180	536.07	1072.153	79.914	378.38	693.77
Right	311	239.89	553.004	31.358	178.19	301.59
Total	1034	369.96	1014.719	31.556	308.04	431.88

Looking at table 5 above, we see that products located at outside front had the highest mean total sales while products located on the left side of the shop had the lowest mean sales.

Just like in analysis 3, this analysis also utilized ANOVA to test whether there is a significant difference in sales and gross profits between different months of the year. Two hypothesis were coined out of this analysis. The hypothesis are as follows;

H₀: There is no significant difference in the mean total sales between the different months of the year.

H₀: There is significant difference in the mean total sales between the different months of the year. 

Table 6: Analysis of variance (ANOVA) for the net sales versus months

Net_Sales
	Sum of Squares	df	Mean Square	F	Sig.
Between Groups	1399993.227	11	127272.112	1.303	.221
Within Groups	34584296.355	354	97695.752
Total	35984289.582	365

A one-way between subjects ANOVA was conducted to compare the mean net sales received based on the month of the year. Table 6 above presents the ANOVA summary. As can be seen, there is no significant differences in the mean net sales of the products based on the month of the year at the p>.05 level for the 12 conditions [F(11, 354) = 1.303, p = 0.221].

Table 7: Mean net sales based on month of the year

Results and findings

	N	Mean	Std. Deviation	Std. Error	95% Confidence Interval for Mean
Lower Bound	Upper Bound
January	31	946.63	350.196	62.897	818.18	1075.09
February	29	1043.04	230.818	42.862	955.24	1130.84
March	31	1044.35	373.638	67.107	907.30	1181.40
April	30	1059.15	309.943	56.588	943.41	1174.88
May	31	1033.86	332.570	59.731	911.87	1155.85
June	30	899.49	221.123	40.371	816.92	982.06
July	31	975.96	242.643	43.580	886.95	1064.96
August	31	990.78	300.051	53.891	880.72	1100.84
September	30	971.00	278.452	50.838	867.02	1074.97
October	31	1009.77	336.438	60.426	886.36	1133.17
November	30	1154.40	302.646	55.255	1041.39	1267.41
December	31	1045.45	405.614	72.850	896.67	1194.23
Total	366	1014.26	313.986	16.412	981.99	1046.54

As can be seen in table 7 above, there is no much difference between the net sales for the month with highest mean net sales and that with the lowest mean net sales. November had the highest net sales (M = 1154.40, SD = 302.65) while June had the lowest net sales (M = 946.63, SD = 350.20).

H₀: There is no significant difference in the mean gross profits between the different months of the year.

H₀: There is significant difference in the mean gross profits between the different months of the year.

Table 8: Analysis of variance (ANOVA) for the gross profits versus months

Profit Total
	Sum of Squares	df	Mean Square	F	Sig.
Between Groups	35370.948	11	3215.541	3.867	.000
Within Groups	294370.006	354	831.554
Total	329740.954	365

We performed a one-way between subjects ANOVA to compare the mean gross profits based on the month of the year. Table 8 above presents the ANOVA summary. As can be seen, there is significant statistical differences in the mean gross profits based on the month of the year at the p<.05 level for the 12 conditions [F(11, 354) = 3.867, p = 0.000]. 

Table 9: Mean gross profits based on month of the year

	N	Mean	Std. Deviation	Std. Error	95% Confidence Interval for Mean
Lower Bound	Upper Bound
January	31	33.0187	41.52022	7.45725	17.7890	48.2484
February	29	23.4317	18.62369	3.45833	16.3477	30.5158
March	31	19.3377	16.22981	2.91496	13.3846	25.2909
April	30	19.6233	12.79108	2.33532	14.8471	24.3996
May	31	20.2316	20.39691	3.66339	12.7500	27.7133
June	30	19.3213	15.05624	2.74888	13.6992	24.9434
July	31	28.8258	17.17982	3.08559	22.5242	35.1274
August	31	34.4823	20.72196	3.72177	26.8814	42.0831
September	30	43.0557	35.75048	6.52711	29.7062	56.4051
October	31	46.2616	38.93676	6.99325	31.9795	60.5437
November	30	43.2477	49.52855	9.04263	24.7534	61.7419
December	31	37.2877	29.33486	5.26870	26.5276	48.0479
Total	366	30.7098	30.05661	1.57108	27.6202	33.7993

The average gross profits for the month of June was 19.32(the lowest in all the months), this was statistically different from the October’s gross profit which was 46.26 (the highest in all the months). 

Using analysis of variance (ANOVA), we sought to answer this research question. There are four seasons (factors) making it prudent to use ANOVA.

The following hypothesis was tested;

H₀: There is no significant difference in the mean net sales between the different seasons.

H₀: There is significant difference in the mean net sales between the different seasons.

Table 10: Analysis of variance (ANOVA) for the net sales versus seasons

Net_Sales
	Sum of Squares	df	Mean Square	F	Sig.
Between Groups	487761.038	3	162587.013	1.658	.176
Within Groups	35496528.544	362	98056.709
Total	35984289.582	365

We performed a one-way between subjects ANOVA to compare the mean net sales based on the season. Table 10 above presents the ANOVA summary. As can be seen, there is no significant statistical differences in the mean net sales based on the season at the p<.05 level for the four conditions [F(4, 362) = 1.658, p = 0.176].

Table 11: Mean net sales based on season

	N	Mean	Std. Deviation	Std. Error	95% Confidence Interval for Mean
Lower Bound	Upper Bound
Summer	91	1011.02	338.322	35.466	940.56	1081.48
Autumn	92	1045.64	336.464	35.079	975.96	1115.32
Winter	92	956.02	257.435	26.839	902.71	1009.33
Spring	91	1044.67	313.798	32.895	979.31	1110.02
Total	366	1014.26	313.986	16.412	981.99	1046.54

The season with the highest average net sales was autumn (M = 1045.64, SD = 336.46) while winter had the lowest mean net sales (M = 956.02, SD = 257.44). Despite the observed differences, statistically the differences were not significant at 5% level of significance.

Analysis of the company’s one year data showed some interesting results that we believe are very insightful to the CEO towards executing his duties as the overall overseer of the company. The key results were as follows;

• Water, vegetable, dairy products and drinks were among the best performing products of the company while juicing, herbal teas, spices, snacks, salad greensand stock sauces were among the worst performing products within the company.
• Credit card and Visa payment methods were the payment options that received huge cash collections.
• Location of the product in the shop matters a lot in terms of the sales performance.
• Month of the year does not significantly affect the net sales of the products however it significantly affects the gross profit.
• Season does not significantly affect the net sales of a product.

We found out that month of the year does not affect the net sales but affects the gross profits; this suggests that there are months when the production costs goes high (either through the materials or labor), the CEO should therefore find out whether there are months that the cost of goods just go up and see on how to solve this. If for instance, the cost of materials rise some months then the CEO could come up with a policy to make purchases before that months comes.

The CEO should also focus on how the products are displayed in the shops as this was found to significantly impact on the net sales.

References

Derrick, B., Toher, D. & White, P., 2017. How to compare the means of two samples that include paired observations and independent observations. The Quantitative Methods for Psychology, 13(2), p. 120–126.

Gelman, A., 2005. Analysis of variance? Why it is more important than ever. The Annals of Statistics, Volume 33, p. 1–53.

Gelman, A., 2005. Analysis of variance? Why it is more important than ever. The Annals of Statistics, p. 1–53.

Howell, D., 2002. Statistical Methods for Psychology. p. 324–325.

John , A. R., 2006. Mathematical Statistics and Data Analysis.

Moore, D. S. & McCabe, G. P., 2003. Introduction to the Practice of Statistics. p. 764.