Introduction

Start-Tech Academy
 Sales Analytics

Hiring

Inventory Management
What is it?
Budgeting

& more…….

Start-Tech Academy
 Sales Analytics

What is it?

Data Information
Insights

Start-Tech Academy
 Sales Analytics
Organize and
Collect data
process data

Calculate and Visualize metrics

Descriptive analytics
Monitor KPI’s - dashboards

Identify
patterns

Forecast sales Predictive analytics

Train ML
models
Identify factors
Predict sales Prescriptive analytics
impacting sales

Start-Tech Academy
 Sales KPIs

KPI Definition Calculation

Conversion rate The percentage of leads or opportunities that are Converted/ Leads
converted into paying customers
Sales Sales cycle length The average length of time it takes to complete a Avg(Sales close time-
process sales process from start to finish sales start time)

Lead response time The amount of time it takes for the sales team to First contact time – lead
respond to a new lead gen time

Start-Tech Academy
 Sales KPIs

KPI Definition Calculation

Sales volume The total amount of money generated by a Amount of money
salesperson in a given period of time generated/ unit time
Salesperson Win rate The percentage of sales opportunities that a Closed sales/ total
specific salesperson successfully closes opportunities

Quota attainment The percentage of a salesperson's assigned quota Sales/ Target

that they have achieved

Start-Tech Academy
 Sales KPIs

KPI Definition Calculation

Marketing- The number of leads that are generated through Total leads
generated leads marketing efforts

Cost per lead The cost of acquiring each new lead, typically Total leads/ Total
Marketing calculated as the total marketing spend divided marketing spend
by the number of leads generated

Return on A measure of the efficiency of marketing efforts, Profits/ Marketing

investment (ROI) calculated as the profit generated from marketing spends
divided by the cost of the marketing efforts

Start-Tech Academy
 Sales KPIs
KPI Definition Calculation
Revenue The total amount of money generated from sales

The total amount of money made from sales after Revenue - cost of
Profit taking into account the cost of goods sold and other goods sold and other
expenses expenses

Financial Cost of goods The cost of producing or acquiring the products or

sold services that are being sold

The difference between the revenue generated from

(Revenue - cost of
Gross margin sales and the cost of goods sold, expressed as a
goods sold)/ Revenue
percentage

Start-Tech Academy
 Sales KPIs
KPI Definition Calculation
Customer The rate at which new customers are acquired New customers
acquisition rate over a given period of time acquired per unit time

Customer retention The percentage of customers who continue to Continuing customers/

rate use the company's products or services over time total customers
Customer
Customer lifetime The estimated value of a customer to the Net present value of all
related value business over the course of their relationship monetary exchanges
with the company

Customer A measure of how happy and satisfied customers Net promoter score
satisfaction are with the products or services they have (NPS)
received

Start-Tech Academy
 Sales Model
Whale Model
Earnings

Sale made Stops paying

What is it? Time

First contact

1. Reduce acquisition cost

2. Increase earnings per customer

3. Increase customer lifetime

Start-Tech Academy
 Sales Model
Objective Data question Analysis
Reduce acquisition How can my past lead conversion • Decision trees
cost data help me identify which leads
are more likely to convert? • RFM Analysis

Increase revenue What products can be pitched to • Market Basket

per customer a customer who has previously Analysis
• Price Bundling
Analytical tools purchased our products?

Increase customer How can past data help me find

lifetime which customer is more likely to Logistic regression
leave and stop paying us?

Sales Forecasting How can past sales data can help • Time series
me predict the future sales? forecasting
• Linear regression

Start-Tech Academy
Decision trees
To create segments and identify which segment is more likely to convert

leads
Marketing team Sales agents

Name Email Phone

Why John [email protected] 7878567
Jack [email protected] 8651412
Jason [email protected] 9542413
Name Email Phone Job title Income Age
John [email protected] 7878567 CEO $93000 59
Jack [email protected] 8651412 Manager $46000 42
Jason [email protected] 9542413 Executive $28000 28
Decision trees
Income > $50000
400, 75% con
Age > 30 300
600, 60% con
1000 leads
360 Income < $50000
40% con Old
200, 30% con Contacted 1000
400 Converted 400
60
Age < 30
New
400, 10% con
Contact only 400
40 Convert 300
 Market Basket Analysis & Lift
Market basket Analysis is used to find products which customers tend to
purchase together

What & Why

Start-Tech Academy
 Market Basket Analysis & Lift
We need data of all items bought by each customer

Lift
Consumer’s basket –
Market Basket Analysis – Finding insights from
list of items bought
basket data
by a consumer

𝐴𝑐𝑡𝑢𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑡𝑖𝑚𝑒𝑠 𝑎 𝑐𝑜𝑚𝑏𝑖𝑛𝑎𝑡𝑖𝑜𝑛 𝑜𝑐𝑐𝑢𝑟𝑠

𝑳𝒊𝒇𝒕 =
𝑃𝑟𝑒𝑑𝑖𝑐𝑡𝑒𝑑 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑡𝑖𝑚𝑒 𝑡ℎ𝑒 𝑐𝑜𝑚𝑏𝑖𝑛𝑎𝑡𝑖𝑜𝑛 𝑜𝑐𝑐𝑢𝑟𝑠 𝑖𝑓 𝑖𝑡𝑒𝑚𝑠 𝑖𝑛 𝑐𝑜𝑚𝑏𝑖𝑛𝑎𝑡𝑖𝑜𝑛 𝑤𝑒𝑟𝑒 𝑖𝑛𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡

Start-Tech Academy
 Sales Forecasting
Sales forecasting is the process of estimating future sales for a business

Forecast can help in planning for future

Forecast suggests rising demand Forecast suggests a slump –

What & Why – hire more sales agents Manage costs

Start-Tech Academy
 Sales Forecasting
Time series forecast – Find patters in historical sales data

Time Jan-22 Feb-22 Mar-22 Apr-22 May-22 Jun-22 Jul-22 Aug-22 Sep-22 Oct-22 Nov-22 Dec-22 Jan-23
Sales 6184 6105 6831 6430 6734 7118 7165 7436 8509 8213 9008 9309 ??

What & Why Machine Learning models – Relationship of sales with other factors

Advertising spend Per capita income Fuel price

Sales

Start-Tech Academy
 Linear Regression
Find relationship between sales and other factors

Advertising spend Spend on QM Sale discount?

What & Why Sales

If I double my ad spend, what happens to sales?

Start-Tech Academy
Linear Regression

Advertising spend Sales

Predictor variables Target variable

Independent variables Dependent variable

How
Linear relationship (single predictor variable)
𝑦 = 𝑎𝑥 + 𝑏
𝑆𝑎𝑙𝑒𝑠 = 𝑎 × 𝑎𝑑𝑣𝑒𝑟𝑡𝑖𝑠𝑖𝑛𝑔 𝑠𝑝𝑒𝑛𝑑 + 𝑏

Multiple predictors –
𝑦 = 𝑎1 𝑥1 + 𝑎2 𝑥2 + 𝑎3 𝑥3 + ⋯ + 𝑎𝑛 𝑥𝑛 + 𝑏

Start-Tech Academy
Linear Regression
Sales

500
𝑒2 𝑒4 𝑒5
400
𝑒1 𝑒3
300

200

100

How 2
Ad. spend

𝑆𝑎𝑙𝑒𝑠 = 𝑎 × 𝑎𝑑𝑣𝑒𝑟𝑡𝑖𝑠𝑖𝑛𝑔 𝑠𝑝𝑒𝑛𝑑 + 𝑏

𝑒1 + 𝑒2 + 𝑒3 …. Will not work

𝑀𝑖𝑛𝑖𝑚𝑖𝑧𝑒 [𝑒12 + 𝑒22 + 𝑒32 + 𝑒42 + 𝑒52 ]

• This is called minimizing sum of squared errors

Start-Tech Academy
 Linear Regression

Coefficients
R
Significance
P–value
square––tells
how
–F –the
iftells
well
that
a’sifour
and
particular
themodel
b’s
predictor
of this
explains
predictor
equation
variables
the
variable
can
𝑦 to
=in𝑎
variations
be
hasused
significant 𝑥relationship
1 + of
predict
1value 𝑎2sales
𝑥2 +values.
target 𝑎with
3 𝑥3 sales
+ ⋯ + 𝑎be
Should 𝑛 𝑥small.
𝑛+𝑏

Use these
Adjusted
Rule of thumb
Rto
square
create
– p-value
significance
– the
if there
final
should
are
Fequation
should
be
more
lessbe
than
than
less
one
0.1,
than
predictor
0.1,
for
For more
eg. 𝑆𝑎𝑙𝑒𝑠
for -more
than = 890%
variables,
than
90% × 𝑎𝑑𝑣𝑒𝑟𝑡𝑖𝑠𝑖𝑛𝑔
confidence
use confidence 𝑠𝑝𝑒𝑛𝑑 − 250
adjusted R Square

Result

Start-Tech Academy
Help John

John
Sales Analytics Manager at STA Retailers

STA Retailers owns hundreds of grocery stores across the

country. John’s team invests a large amount of time and
effort in analyzing different localities to find a location for
new stores . John has recently completed a course in
Analytics and Machine Learning, and he thinks that STA
can use Machine Learning to use past data to create data-
driven process for selecting new store locations.

Start-Tech Academy
Retention

Increasing customer lifetime by preventing churn

When the customer pays over a period of time, churn impacts sales revenue

Sales team shares responsibility, tries to retain customers

What Time Series forecasting to ML models to predict

predict overall churn probability of churn

Logistic regression can help predict probability of churn

Pitch alternate options/ offers to the customer to retain the customer

Logistic Regression

Uses historical data to train, same as linear regression

Linear Regression – Logistic Regression –

predicts continuous
What variable
predicts categorical
variable

Output of logistic regression corresponds to probability of belonging to one

of the categories
 Logistic Regression
Avg. Bill Value Relationship Age No. of service calls Churned?
120 34 0 0
90 6 3 1
135 23 1 0
… … … …

Example Use this data to train our logistic regression model

Avg. Bill Value Relationship Age No. of service calls Churn prob
150 12 2 0.75
80 29 0 0.32
 Logistic Regression
Predicted Positive Predicted Negative
Actual Positive True Positive False Negative
Actual Negative False Positive True Negative

True Positive - the model correctly predicted the churn.

False Positive - the model incorrectly predicted the churn.
Confusion Matrix True Negative - the model correctly predicted the retention.
False Negative - the model incorrectly predicted the retention.

Accuracy = (True positive + True Negative)/ Total cases

Predicted Positive Predicted Negative Accuracy =

Actual Positive 63 27 (63+122)/250 =
185/250 =
Actual Negative 38 122 92.5%
Help John
Tim
Sales Analytics Manager at STA Telecom
STA Telecom provides 4G and 5G services all over India. STA currently
offers services including fixed-line broadband, and voice services.

Tim has been tasked to reduce customer attrition. He has recently

completed a course in Sales Analytics, and he thinks that STA can use
Machine Learning to use past data to create data-driven process for
reducing customer churn.

Start-Tech Academy