Europet Questions
Week 9
3/05/13
•
Europet Class Lecture
Sales = f (advertising, etc) o y = a1 + b1*x1 o a = intercept o b = beta coefficient o x = $ spent
Presentations
• does adv pay for itself?
• R2 low = low correlation
• Needs more relevant variables
• Increased R2 to 85%
• Tv > radio beta coeff for radio <2 so insignificant
• Sales vs volume r2 < .85 no relationship between variables
• Second regression sales vs tv and radio grps – r2 < .85 no relationship
Spreadsheet
• GRP: gross rating points tv adv budgets evaluated in terms of ratings o Every time reach 1% of audience = 1 point (multiplication of both = grp)
REGRESSION
• Beta coefficients o Tv coeff = 1,219
1219 =Unit rise in sales per each unit rise in GRP o Radio = 520
• Each $1 grp increase in tv budgeting get $1219 pike in sales
• Increase tv by 1 grp push up sales by 1219 sales units (dollars)
• Substitute $300 for 1 tv grp (given) know short term effect of tv adv spend 300 sales lever goes forward 319 times o y = $2,214,238.5 + 1,219 * $100
= $2,336,138.5
• Sales to dollars if spend 100 grp spent on tv
• 1219 = increase in store sales for each additional grp o total sales increase 100 $1219 o lagged effect (mf=3) $3657 multiplication factor given o margin of 30% $1097
(profit = 30% of lagged) adv paid for self in terms of profit? given margin 30% o less cost of GRP = $300 o profit on 1 GRP = $797
1097 + (300) o ADV ROI = 266% profit / less cost
• RADIO o Beta: 519 o Increase in avg store sales for each additional grp = o Total sales increae = 519 o Lagged = 1557 o Margin 30% = 467 o Less cost of grp = 25 o Profit on 1 grp = 442
o
•
•
•
Adv roi = 1768%
Beta on tv/radio both grp = 792
Radio beta = .035
+/ diff variables in right combo, can make tv/radio coeff go up and down o don’t know if fuel, temp, holiday, etc impact adv sales o limits of using regression to answer questions about “does adv cause sales”
what is better procedure than regression?
• “correlation is not
3/05/13
•
Europet Class Lecture
Sales = f (advertising, etc) o y = a1 + b1*x1 o a = intercept o b = beta coefficient o x = $ spent
Presentations
• does adv pay for itself?
• R2 low = low correlation
• Needs more relevant variables
• Increased R2 to 85%
• Tv > radio beta coeff for radio <2 so insignificant
• Sales vs volume r2 < .85 no relationship between variables
• Second regression sales vs tv and radio grps – r2 < .85 no relationship
Spreadsheet
• GRP: gross rating points tv adv budgets evaluated in terms of ratings o Every time reach 1% of audience = 1 point (multiplication of both = grp)
REGRESSION
• Beta coefficients o Tv coeff = 1,219
1219 =Unit rise in sales per each unit rise in GRP o Radio = 520
• Each $1 grp increase in tv budgeting get $1219 pike in sales
• Increase tv by 1 grp push up sales by 1219 sales units (dollars)
• Substitute $300 for 1 tv grp (given) know short term effect of tv adv spend 300 sales lever goes forward 319 times o y = $2,214,238.5 + 1,219 * $100
= $2,336,138.5
• Sales to dollars if spend 100 grp spent on tv
• 1219 = increase in store sales for each additional grp o total sales increase 100 $1219 o lagged effect (mf=3) $3657 multiplication factor given o margin of 30% $1097
(profit = 30% of lagged) adv paid for self in terms of profit? given margin 30% o less cost of GRP = $300 o profit on 1 GRP = $797
1097 + (300) o ADV ROI = 266% profit / less cost
• RADIO o Beta: 519 o Increase in avg store sales for each additional grp = o Total sales increae = 519 o Lagged = 1557 o Margin 30% = 467 o Less cost of grp = 25 o Profit on 1 grp = 442
o
•
•
•
Adv roi = 1768%
Beta on tv/radio both grp = 792
Radio beta = .035
+/ diff variables in right combo, can make tv/radio coeff go up and down o don’t know if fuel, temp, holiday, etc impact adv sales o limits of using regression to answer questions about “does adv cause sales”
what is better procedure than regression?
• “correlation is not