Test 1 : What is the Sales Forecast for Dec ?
100
200
300
400
500
600
700
800
900
1000
1100
0
200
400
600
800
1000
1200
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Test 2 : What is the Sales Forecast for Dec?
2000
1900
1800
1700
1600
1500
1400
1300
1200
1100
1000
0
500
1000
1500
2000
2500
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Test 3: What is the Sales Forecast for Dec ?
100
200
300 300
400
500
600 600
700
800
900
0
100
200
300
400
500
600
700
800
900
1000
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Test 4 : What is the Sales Forecast for Dec ?
100
50
200
100
300
150
400
200
500
250
600
0
100
200
300
400
500
600
700
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Test 6 : What is the Sales Forecast for Dec ?
300
900
0
100
200
300
400
500
600
700
800
900
1000
Oct Nov Dec
Is your answer 600-700-800 ?
Not sure ….Difficult .. isn't it ?
What is the learning ?
More the data, its easy to forecast !
If there is a pattern…its easy to forecast !
If there is seasonality…its easy to forecast !
Forecasting : 3 Simple Factors
Use these 3 factors to
forecast the sales
Season
Time series
Trend
Time Series and Patterns in Time series
Time Series: The repeated observations of demand for a
service or product in their order of occurrence.
There are five basic patterns of most time series.
Horizontal. The fluctuation of data around a constant mean.
Trend. The systematic increase or decrease in the mean of the series
over time.
Seasonal. A repeatable pattern of increases or decreases in demand,
depending on the time of day, week, month, or season.
Cyclical. The less predictable gradual increases or decreases over
longer periods of time (years or decades).
Random. The un-forecastable variation in demand.
Forecasting : 3 Simple Factors
Use these 3 factors to
forecast the sales
Season
Time series
Trend
Time Series: The repeated observations of demand for a
service or product in their order of occurrence.
Easy to Forecast !
More the data .. More predictability..
Better Forecasts
Not Sure!
Forecasting : 3 Simple Factors
Use these 3 factors to
forecast the sales
Season
Time series
Trend
A trend is a movement in a particular direction
Trend lines in charts
• A trend is a movement in a particular direction
• A trend line is a straight line connecting multiple
points on a chart.
• The magnitude of the slope of a trend line, or
steepness, indicates the strength of the trend.
• Trendline can be used to forecast future !
Creating Trend line in excel : Step 1
Select the graph and
right click
Select Add Trend line
A window opens with types
of trend line to select from 6
Options
Select the default option for
trend line – Linear
Click OK
Creating Trend line in excel : Step 2
You can see the trend line now
Creating Trend line in excel : Step 2(Contd.)
Formatting Trend line in excel : Step 3
Right click on the trend line to
format it.
Select style, color and weight of
your choice
Formatting Trend line in excel : Step 3 (contd.)
Right click on the trend line to
format it.
Select style, color and weight of
your choice
By using this forecast Option you
can extend the trend line
forward or backward to the
number of periods as desired
Here, I selected 6 periods
(Months - in this case) forward
Forecast Options - Trend line in excel : Step 4(Contd.)
You can Observe that the trend line got extended in to the future by 6 periods
This denotes if this trend continues, the sales of the country are likely to be like
this
Forecast Options - Trend line in excel : Step 4(Contd.)
By clicking “display equation on
Chart” and display r-Squared
value on chart, you will be able
to display them on the graph -
Next slide explains them
You can Set Intercept to any
value by clicking it and adding
value (0 is default).
As this Option is not used
often in simple trend lines ,
explanation is beyond the
scope of this slide set.
Options - Trend line in excel : Step 4(Contd.)
This is the equation generated by excel that holds
the mathematical relationship of Sales and months.
Equation and R Squared values – Step 4 (Contd.)
Equation
This is the equation generated by excel that holds
the mathematical relationship of Sales and months.
This equation will be unique to each data set
This equation can be used to compute sales of any
future month
to put it simply , here “y” denotes the sale (which
is on Y axis) and x denotes the month
So lets compute 13th month projection, using this
equation
y= ((17.311)*13)+296.39 that equals 521.43
So the likely sales projection for 13th month is
521.43
This is the R-Squared value for this trend line.
R Squared value
This is the R-Squared Value for this trend line.
This value denotes the reliability of the sales
projections.
R squared value will range between 0 and 1
If the R squared value is 1, then the trend is most
predictable and reliable.
The reliability of trend line goes up if the R-
Squared value is nearest to 1
Let us see the next example to understand it
better.
R Squared value
Take a look at the trend of sales and R Squared value
R Squared value = 1
Here the R Squared Value is 1.
Just take a look at the sales progress.
With every passing month, this territory is adding
$100 to the previous month.
So, going by the trend, you can be almost sure, that
the 13th month sales are .
Remember, Trend lines and Forecasts means you are
presuming the existing market conditions are not going
to change radically.
Take a look at the Equation.. It is y=100x
This means Y, the next month sales (13 th Month
sales)
y= ((100)*13) that equals 1,300
So ,the likely sales projection for 13th month is 1,300
Remember the Forecast Test 1 ?
Now that you know what is R Squared value,
It’s Time to understand Correlation
By the way, r is called Correlation Coefficient
and
R Squared value is called Ccoefficient of Determination.
You need not remember these Hi-Fi terminology
It’s good enough to understand that if R Squared value is
near value of 1, the forecast is more accurate!!
In both the examples, There is a very good correlation between the
months progress and sales progress
With every month increase , there is a gain or loss of 100
In both the cases, R Squared value is equal to 1
This is also a perfect Correlation!
The R Squared value , if measured must be 1 !!
Some time ago, Wal-Mart decided to combine the data
Once combined, the data was mined extensively and many
correlations appeared.
Some of these were obvious; people who buy gin are also likely to
buy tonic. They often also buy lemons.
However, one correlation stood out like a
sore thumb because it was so unexpected.
Those queries revealed that, between 5pm and 7pm,
customers tended to co-purchase beer and diapers.
It seems , they found out that parents who wants to babysit
and also watch football and drink beer do not want to be
disturbed by their babies!
Trend line types
1. Logarithmic
2. Polynomial
3. Power
4. Exponential
5. Moving Average
Do Remember, that for
every trend line type you
select, you will get different
equation and R Squared
value
Equation and R Squared Values- various trend lines
Rule of thumb to use type of Trend line
1.Linear trend line : Use it if data values are increasing
or decreasing at a steady rate.
2. Logarithmic trend line : Useful when the rate of
change in the data increases or decreases quickly
and then levels out.
Rule of thumb to use type of Trend line
3. Polynomial trend line : Used when there are data
fluctuations like the sales following seasonal trends
Rule of thumb to use type of Trend line
4. Power trend line : Use with data that has values that
increase at specific rate at regular intervals.
Rule of thumb to use type of Trend line
5. Exponential trend line : Use when data values
increase or decrease rates that are constantly
increasing.
Rule of thumb to use type of Trend line
6. Moving average trend line : Use it when uneven
fluctuations are in data values
Rule of thumb to use type of Trend line
Best fit Trend line
• We have learnt that if R2 Value is near to 1, the reliability of
trend line is better.
• So, now we need to use a trend line from the five available
trend lines in excel menu to arrive at the most appropriate
one to ensure that our forecast is most reliable.
• The other trend line left out is Moving average for which
,you will neither get the equation nor the R2 value.
• Simplest way to find the best fit trend line is to check every
trend line’s R2 Value and use the trend line with highest R2
value ( Which is nearest to 1- out of 5 types)
Forecasting : 3 Simple Factors
Use these 3 factors to
forecast the sales
Season
Time series
Trend
Measuring Seasonality is simple
Sales Year 1 Year 2 Year 3 Year 4 Year 5 Year 1 Year 2 Year 3 Year 4 Year 5 5 Year Avg.
Jan 205 293 380 468 565 6.2% 6.7% 7.0% 7.1% 7.2% 6.9%
Feb 296 387 482 571 663 8.9% 8.9% 8.8% 8.7% 8.5% 8.7%
Mar 368 453 544 635 795 11.1% 10.4% 10.0% 9.7% 10.2% 10.2%
Apr 396 483 568 657 770 11.9% 11.1% 10.4% 10.0% 9.9% 10.4%
May 388 480 571 669 789 11.7% 11.0% 10.5% 10.2% 10.1% 10.5%
Jun 248 338 425 524 613 7.5% 7.7% 7.8% 8.0% 7.9% 7.8%
Jul 214 301 387 484 578 6.4% 6.9% 7.1% 7.4% 7.4% 7.1%
Aug 245 332 426 525 621 7.4% 7.6% 7.8% 8.0% 8.0% 7.8%
Sep 241 322 407 505 597 7.2% 7.4% 7.5% 7.7% 7.7% 7.5%
Oct 244 334 422 515 611 7.3% 7.6% 7.7% 7.8% 7.8% 7.7%
Nov 265 348 446 542 629 8.0% 8.0% 8.2% 8.2% 8.1% 8.1%
Dec 217 299 393 483 564 6.5% 6.8% 7.2% 7.3% 7.2% 7.1%
Year 3,327 4,370 5,451 6,578 7,795 100.0% 100.0% 100.0% 100.0% 100.0% 100.0%
So, these three factors can help us forecast
Use these 3 factors to
forecast the sales
Season
Time series
Trend
Here is a tool I created for you… using all 3 factors
Please download it in the downloads section or click the link below
Forecasting tool
Pharma Factors to consider
• Price variations – Better to forecast on units than
value
• Sales Closings and Incentives – Cyclical patterns ?
• Campaigns – Special campaigns ?
• Trade offers - Seasonality?
• Market – Localized Seasonality ? ( Local festivals in
zones e.g. Diwali, Puja etc.)