Goodwill and Caliber of an Ad Server

Ad Servers perform two functions that produce value

The number one function is to produce Profit. Given the potential difference between Sales Price and Cost the server finds the Ad Set that is expected to produce the most Gross Profit (Revenue – Cost of Revenue). The cumulative effect of this ad server function is seen in the company Income Statement.

 

The second function is to distribute sales to customers (and sometimes distribute revenue to suppliers) to produce Goodwill. Goodwill is an intangible Asset of the company. It usually represents the sustainability of the current Gross Profit production for some time in the future. These Ad Server effects are seen on the company Balance Sheet under the headings of Goodwill and or intangible assets. Sustainability and Goodwill are the inverse of Risk.

 

CEO Risk Aversion Θ

The CEO of the company has some Aversion to Risk of Sales Θ_Sales , in the range {0,1}. A CEO with high Aversion will want to serve ads to reduce Risk of Sales by distributing sales across as many customers as possible. They know change is an integral part of “at will” pricing. This CEO wants the business to be resilient to change. This CEO would like the Profit to remain steady even when negative changes occur. This CEO might “purchase” Supply Caliber or reduce the Risk of Supply by exchanging cash with a supplier for the rights to traffic (an up-front purchase of Goodwill). Accountants call this an intangible asset on the Balance Sheet, but I will call it profit sharing supplier Goodwill (see below).

 

The low Aversion CEO may prefer Profit because they discount the risks to sales. This CEO may believe the Sales team will immediately replace lost customers (big change). This CEO may believe that change will be slow to occur.

 

Dual Optimization

If the CEO were investing in the Company then their Aversion of Risk is part of their calculation of Goodwill and company value. They would calculate the worth of customer Goodwill over some time period τ as:

GW = Profit  Θ_Sales / Risk | τ

GW = Profit  Θ_Sales  Caliber | τ

where Caliber is the Caliber of the distribution of Sales

Caliber = ∑ – p log p

where p is percentage sales by customer

 

It seems a CEO (investor) would like the ad server to perform both optimization of Profit and Goodwill given their Aversion to Risk Θ. The ad server now has a new optimization problem. Find the ad set that produces Maximum of:

max –> Profit + GW

max –> Profit + Profit Θ_Sales Caliber

max –> Profit ( 1 + Θ_Sales Caliber)

 

The largest ad server on the planet, Google Adwords, shows 3 ads at the top of most search listings. These top three ads produce most sales and establish a minimum Caliber for each ad set which reflects a given Aversion to Risk of their CEO. The Caliber of these three ads is about 1.5 bits.

 

Factors that determine the max –> Profit( 1 + Θ_Sales Caliber) of an Ad Set

• CEO (investor) Aversion to Risk Θ_Sales
• Style of placements – flat vs prominence – p(performance | placement) – CTR matrix
• The gradient of the Revenue ∇(Price x Performance) of eligible ads
• Cost of traffic
  • Fixed margin m or profit sharing suppliers
    • this requires production of Supplier Goodwill and new Balance Sheet asset entry
    • Risk of Supply and Caliber of Supply enters into optimization
    • Find max –> Profit [1 + Θ_Sales Caliber(sales) + Θ_Supply Caliber(supply)]
    • What is CEO Aversion to Supply Risk Θ_Supply? This may be different than Aversion to Risk of Sales.
  • Variable cost – open market traffic buying
    • Produces no “Supplier Goodwill”
    • Find max –> Profit [1 + Θ_Sales Caliber(sales)]

Algorithm to find max

Let’s take the most simple optimization where costs are variable and no Supplier GW is involved. The strategy is to pick a good beginning Ad Set and “Jiggle” the set by swapping placement of two ads to increase Goodwill. Recalculate the Equity with each Jiggle.

1. Find the Ad Set with maximum expected Profit.
2. Calculate: Beginning Equity = Profit + Goodwill of the Ad Set
3. Find the most efficient change in ad rank of two eligible ads to produce the maximum Goodwill per Profit lost.
   1. Calculate: New Equity
   2. If New Equity > Beginning Equity:
      1. Accept changes to Ad Set.
      2. Goto step 3
   3. If New Equity < Beginning Equity
   4. Show ads