Question

Robert Mondavi has great pride in his wine, in fact, he is a legend up north for his technical improvements and marketing strategies brought worldwide recognition for the wines od the Napa Valley in California. He recently dabbled in advertising is a product on Facebook. While he wants to try the “social thing” – you suggest that “the relationship between advertising display ads on Facebook and sales is positive & somewhat linear- but the story isn’t that simple…”

Below is the data.

Problem:

Ad Exp.     Sales

$2,500       $95,000

$3,000       $110,000

$3,500     $124,000

$4,000      $138,000

$4,500      $143,000

$5,000       $147,00

$5,500     $149,00

$6,000      $155,000

$6,500      $169,000

$7,000      $185,000

$7,500      $187,000

$8,000      $192,000

$8,500      $195,000

$9,000      $198,000

Why is the story not so simple? What is going on here? (make sure you research “display ads” on Facebook prior to answering).

Answer 1

The normal charting of the data and the change in the sales due to change in ad expense (delta) is shown below.

X fAd exp A1 B UA A C E G M. N C S T Ad exp Sales Delta 2500 95000 3000 110000 30 Delta Sales 28 35 250000 A000 138000 6 4500 143000 10 30 147000 7 5000 200000 5500 149000 4 25 155000 6000 12 10 6500 169000 28 150000 11 185000 20 7000 32 187000 12 7500 4 13 8000 192000 10 15 100000 14 8500 195000 6 15 9000 198000 6 10 16 50000 17 18 19 C 2500 3000 3500 4000 4500 5000 5500 6000 6500 7000 7500 8000 8500 9000 10 12 14 22 23 24 25 26 27 28 Sheet1209 Sheet567 (2) Sheet1210 .. Sheet1205 Sheet1206 Sheet1207 Sheet1208 READY

From the sales chart, the data looks linearly related. However the delta describes a different picture. We can see that the relationship between sales and ad expenses are not always linear. This means that there are other factors in play.

One of the explanations for these can be that there are various stages of threshold that needs to be crossed in order to reach the next level of sales. Till then the overall trend seems to mature. This is why the deltas form a (somewhat) sinusoidal pattern.

The other factors that needs to be considered are the audience types, impact of other advertisers, the bid amount and the quantity. Especially the bidding process on Facebook or social media is pretty dynamic. Due to this, the level of exposure that may happen for a given budget can be quite random. Since the exposure is random, the level of sales also depend on the exposure. This is why there may be deeper story here.

For example, spending 2500 can give you 50000 visitors or 75000 visitors. If we consider the same level of conversion rate, they may give different values in terms of number of sales and thus the sales values.