Anyone that has studied economics at any level will have heard about Price Discrimination. I first learnt about Price Discrimination when I was just 17 years old as part of my A-level economics course. It is an essential part of any basic microeconomics course. It is also the key to understanding how media owners can maximise ad revenue. For this blog post, I will first describe what Price Discrimination is and secondly describe how to use it as a real-world programmatic media pricing strategy.
Price Discrimination (The Theory)
It’s been a while since I studied Economics, what is Price Discrimination?
Price discrimination is simply a pricing strategy where a seller sells similar goods or services at different prices to different buyers. The price charged to each buyer depends on their willingness to pay. Sometimes people use the term differential pricing for the same thing. See here for Wikipedia definition.
See below demand curves (remember those?) for ad inventory. The first chart shows the typical situation, i.e. a technology-led solution with no coherent sales strategy and a single exchange price (bid floor) and multiple exchanges. This is contrasted with the ideal scenario, where a price discrimination sales strategy enabled by a single exchange set-up with dynamic bid-floors. It can clearly be seen that the firm earns more money where it can set a price for each buyer.
Two things need to happen if a price discrimination strategy is to work. Firstly, the Publisher must be able to find out the maximum each buyer is willing to pay. Secondly, the Publisher must prevent reselling and arbitrage, i.e. allowing buyers from selling to each other.
The first condition sounds really difficult to achieve in any practical way… at first. Even looking beyond digital media, how can any firm simply ask customers how much they are willing to pay then charge them that amount?
The answer to that depends on the nature of what is being sold and the market structure. There is always a way of finding a mechanism or proxy to calculate willingness to pay. Probably, the easiest type of market to operate a price discrimination strategy is in a market for perishable goods. The reason is that the product being sold is restricted by time and time is used as a mechanism to tease out willingness to buy.
The best example is the travel industry. All travel tickets have a fixed expiry time, a flight ticket has a zero value one second after the gates close. The value of a plane ticket depends on three main variables: type of seat, day/time of flight, days/time from departure. Most of us have compared these three variables when booking a flight and when this happens we are working out our willingness to pay!
The second condition sounds difficult too. How does a firm stop an intermediary from buying at a low price and selling at a higher price (i.e. arbitrage)? This is achieved by strictly allocating the product to the buyer. Again, the easiest way to do this is with perishable products. For example, most tickets for most events or travel are restricted to the buyer. In many cases, the firms selling these tickets have lobbied lawmakers to make it illegal (or at least very difficult) to resell tickets. This is the case for everything from festival tickets to airline tickets. It should be no surprise that changing a name on a flight ticket (even when it is to correct a misspelling) is so difficult and costly.
Does price discrimination really work?
Price discrimination has been around a very, very long time. That should give some indication of how useful a strategy it is. The first serious academic work on it was done by the French civil engineer and self-taught economist Jules Dupuit in the 1840s. However, Dupuit only provided the first rigorous academic analysis to what was already common practice. There are many examples of price discrimination strategies in practice, for examples the use of coupons, premium pricing (i.e. economy, regular and premium prices), student discounts, pensioner discounts and volume discounts, etc.
However, the best single (and most advanced) use of price discrimination has got to be in transport market, particularly the market for plane tickets. This has become so sophisticated there is now the concept of Dynamic Pricing, where pricing automatically changes according to complex business rules and real-time market demand.
What about implementing price discrimination in the programmatic digital display market?
Prior to the current programmatic ecosystem, a salesman could negotiate in person or over the telephone to extract the highest price they could for whatever media they were selling. In this, albeit it unscientific way willingness to pay could be understood and exploited by the media owner.
In the programmatic ecosystem, there are literally thousands of media buyers. Since your audience is made up of individuals, not all of your site’s audience are equally valuable to these thousands of advertisers. Therefore, it makes sense to charge differently to different advertisers.
The huge problem (for Publishers) in today’s’ programmatic market is that it the market is dominated by large ad tech players that operate on an arbitrage basis. I call them “End-to-End Ad Platforms” (see more here). These are ad platforms that both buy ad media from Publishers and sell ad media to Agencies or Advertisers. The business model of an End-to-End Platform is to arbitrage between the lower sale price of the Publisher and the higher buying price of an Agency. They don’t want Publishers to know how much agencies/advertisers are willing to pay because if this market information is known to Publishers it will destroy the ad platforms business model. This is one of the main reasons why there is so little transparency in programmatic markets. The result is a digital media market that is inefficient and ineffective. It is hardly surprising digital publishers are struggling while digital marketing spend increases year-on-year. Publishers using an End-to-End platform (e.g. DFP with AdSense or AdX) will not be able to operate price discrimination sales strategy effectively.
One of the main reasons why there is such a problem in digital media is that programmatic adoption was seen as a technology issue, rather than a market/sales strategy issue. Effective programmatic ad sales are not simply about adopting technology. The focus should be on achieving strategic goals through technology combined with complementary sales operations. Maximising advertising yield may have become more complex with the advent of programmatic but it still relies on price discrimination via simple tactical goals:
1. Maximise the number of potential buyers
2. Maximise selling opportunities
3. Strictly control buyer access
4. Vary price according to buyer willingness to pay
5. Continually optimise 1-4. (in real-time, impression-by-impression)
This may sound a little vague, so see below seven practical recommendations of how this can be achieved.
ADUNITY is a publisher focused ad technology company based in the UK with offices in both London and Bucharest. We deliver trust, transparency and advanced technology in the programmatic ecosystem. AdUnity provides transparent trading platforms for publishers (no black boxes) and does not operate an arbitrage model; just a flat revenue share. We comply with the latest programmatic standards (OpenRTB 2.4) and format standards (HTML5, VAST 2.0, VPAID 2.0, MRAID 1.0, Native Ads API 1.0). AdUnity works with the largest media and agency groups in Central and Eastern Europe and Turkey.
B2B publishers do not tend to use any monetisation ad technology because they sell all their ad inventory on a…
Today AdUnity announces the launch of the new privacy protected AdUnity Enterprise Ad Server for publishers (EAS). By default, the…
We have invested heavily in developing our privacy-by-design martech and ad-tech technology as required by GDPR (Articles 5, 11 and…
As part of our mission to get the advertising industry ready for GDPR AdUnity is pleased to announce that we…