Overview

This report presents an analysis of sales performance based on 500 sales opportunities from the previous year. The goal is to identify factors influencing sales outcomes and provide recommendations for improving sales performance.

Data Sources

The data used for this analysis was sourced from the “AnalysisProblem1_Data (Problem1 Data)” worksheet. Each row represents a single sales opportunity or ‘deal’.

Methodology

Data Cleaning

The data underwent several cleaning steps to ensure accuracy and consistency:

  • Repeated rows were removed.
  • Discount values greater than or equal to 1 were corrected.
  • Outliers in the DaysBetweenEvents variable were identified and removed.
  • ContactsInvolved values greater than or equal to 100 were scaled down.
  • Source values were standardized.
  • TechSales values were converted to binary.
  • Missing values were dropped.

Data Validation

The cleaned data was validated to ensure its integrity and reliability: - Summary statistics were generated to understand the distribution of key variables. - Outliers were identified and removed from the analysis. - The t-test was performed to compare means between winning and losing deals. - Logistic regression was conducted to identify significant predictors of sales outcomes.

Analysis Abstract

Hypothesis Testing: DaysBetweenEvents and Win Rates

A t-test was conducted to test the hypothesis that fewer DaysBetweenEvents leads to better win rates.

Logistic Regression Analysis

Logistic regression was performed to identify significant predictors of sales outcomes. The analysis revealed several significant and marginally significant coefficients.

Won-Lost Deals Boxplot

Table: Days Between Events When By Outcome Stats

Outcome deal_count deal_pct mean sd minimum q25 q50 q75 q95 maximum
Lost 269 53.8 10.12 2.42 3.87 8.46 10.28 11.70 13.96 16.00
Won 223 44.6 9.38 2.54 1.81 7.84 9.38 11.17 13.15 18.85

Insights

  • By performing a t-test we know that the diference between means by outcomes is highly significative on data where higly divergent ouliers were removed.
  • By looking at the statistics, we can infer that less Days Between Events leads to better win rates.

Logistic Regression to find current and new relationships

Table: Significant and Marginally Significant Coefficients from the Regression

Estimate Std. Error z value Pr(>|z|) probability
DaysBetweenEvents -0.1235920 0.0381989 -3.235487 0.0012144 0.4691413
ContactsInvolved 0.1145582 0.0633506 1.808321 0.0705566 0.5286083
SourceMarketing -0.5010243 0.2688501 -1.863582 0.0623804 0.3773000
SourceSales -0.5987929 0.2450630 -2.443424 0.0145486 0.3546199

Insights

DaysBetweenEvents:
  • Estimate: -0.123592
  • Pr(>|z|): 0.00121 (significant at the 1% level)
  • Probability: 0.4691413
  • Interpretation:
    • A decrease in the number of days between events is associated with a higher likelihood of winning a deal. Specifically, for each additional day, the probability of winning decrease by approximately 46.9%. This is a strong and highly significant predictor. Confirming what we saw with the T-test.
Contacts (Involved):
  • Estimate: 0.114558
  • Pr(>|z|): 0.07056 (marginally significant)
  • Probability: 0.5286083
  • Interpretation:
    • Each additional contact involved in a deal increases the probability of winning by approximately 52.8%. While this predictor is not statistically significant at the conventional 5% level, it is marginally significant and suggests a potential positive impact on deal outcomes.
Source (Marketing):
  • Estimate: -0.501024
  • Pr(>|z|): 0.06238 (marginally significant)
  • Probability: 0.3773000
  • Interpretation:
    • Deals sourced from the Marketing department have lower odds of being won compared to those sourced from Customer Success, with a reduction of approximately 37.7% in the odds of winning. This finding is marginally significant, suggesting that marketing-sourced deals might require additional support or different strategies.
Source (Sales):
  • Estimate: -0.598793
  • Pr(>|z|): 0.01455 (significant at the 5% level)
  • Probability: 0.3546199
  • Interpretation:
    • Deals sourced from the Sales department have significantly lower probabilities of being won compared to those sourced from Customer Success, with a reduction of approximately 35.4% in the probability of winning. This is a significant finding, indicating a need to investigate the reasons behind this lower success rate and possibly provide additional support or training for sales-sourced deals.

Summary of Findings:

  • DaysBetweenEvents: The fewer the days between meetings, the better the win rates. This is a highly significant finding.
  • ContactsInvolved: Having more contacts involved in a deal shows a positive effect on winning, though it is only marginally significant.
  • SourceMarketing: Deals sourced from Marketing have lower odds of winning, which is marginally significant.
  • SourceSales: Deals sourced from Sales have significantly lower odds of winning compared to those sourced from Customer Success.

Recommendations:

  1. Reduce Days Between Events:
    • Training sales reps to minimize the days between meetings with customers could improve win rates.
  2. Increase Contacts Involved:
    • Encouraging sales reps to involve more contacts from the customer’s side might have a positive impact.
  3. Reevaluate Sourcing Strategies:
    • Investigate why deals sourced from Marketing and Sales have lower win rates and consider improvements or additional support for these sources.
  4. Further Analysis:
    • Explore potential interactions between variables and consider additional factors that might influence sales outcomes.

These findings provide actionable insights that can help guide sales strategies and training efforts to improve overall sales performance.