Whats up with that smart pricing?

7 Responses to Whats up with that smart pricing?

1. ezman says:

   People have this strange stigma about smart pricing. I've worked in the industry for 10 years and been involved at an algorithmic and technical level on smart pricing. It's gets very complicated very quickly, but the basis is actually very simple.
   
   First thing to understand is that traffic from different sources (publishers, websites, etc) converts for advertisers at different levels. (Conversion Rate or CVR).
   
   So, to solve this problem, ad network "smart-price" which is a way of leveling out the amount an advertisers pays for clicks of varying quality.
   
   Here is a simply example.
   
   Advertiser is bidding $1.00 cpc on the term "buy widget".
   
   Publisher A's traffic sells one widget for every 20 clicks (5% conversion rate). This means the advertiser is paying an average of $20 for every sale they get from publisher A.
   
   Now, enter Publisher B. Publisher B's traffic coverts at a lower rate -- let's say that only 1 in 40 people actually buy a widget from advertiser.
   
   This would mean that the advertiser is paying $40 for a sale generated from publisher B. So, the ad network "smart prices" clicks from publisher B down by 50% to maintain the publishers eCPA of $20.
   
   Why do they do this? Simple. If they didn't smart price traffic sources that covert at a rate less than average, the advertiser would lower their bid to compensate. In the above example, assuming that both publishers had the same volume, the advertiser would lower their bid to $.75 to maintain their eCPA at $20.
   
   This would cause publisher A's earning to decrease even through their traffic converts better and publisher A would then be better off going to another ad network that didn't have Publisher B in it, causing lower bids.
   
   So, all traffic is "smart priced" to either the network average or the "best of network". This keeps upward price-pressure in the market and helps the network maintain high ppcs and high converting traffic sources.
   
   Make sense?

   Reply
2. admin says:

   Yes, thank you for this detailed explanation.
   
   And welcome aboard ;)

   Reply
3. BirdOPrey5 says:

   Thanks for the explanation but this still isn't fair to publisher B... Publisher B runs the same ads as A, takes the same chance of annoying their own users as A, and otherwise runs the same risks as A for less gain (half in your example.) It's the advertiser's job to turn a click into a sale, not my job- my job is to display the ad- nothing more.

   Reply
   • ezman says:

     Thanks for the explanation but this still isn't fair to publisher B... Publisher B runs the same ads as A, takes the same chance of annoying their own users as A, and otherwise runs the same risks as A for less gain (half in your example.) It's the advertiser's job to turn a click into a sale, not my job- my job is to display the ad- nothing more.

     Yes, thats the common misconception of publishers. It's the publishers job to drive qualified visitors to an advertiser so that they make a purchase since the advertiser is ultimately paying you for those sales / leads.
     
     If you're placing ads near navigation items and getting accidental clicks, that doesn't provide value to the advertiser. If you are blending ads in too much and people are clicking on ads thinking they are navigation items, that also doesn't provide value to the advertiser either.
     
     The money that you're paid for showing ads need to come from somewhere, and that somewhere is from someone selling a product / service, etc.
     
     You can still find ad networks that don't smart price -- they are the ones with the 2 cent PPCs because they fall into the above trap of continually losing high quality traffic sources which results in declining ppcs.
     
     Adsense, on the other hand, most likely smart-prices by the most granular level with statistically significant data available. So, I imagine they have a base rate for an entire Adsense account, then by domain, then by individual ad placement or page / size.
     
     All this would be split by geographical regions as well. So, if you change your ad placements and you see increased clicks and earnings for a few days (or longer / shorter depending on volume) and then see click remain consistent but earnings drop (a PPC drop), you are most likely being smart priced. Try removing the likelihood of accidental or misleading clicks, leave it for a week or so (again, dependent upon volume) and you'll likely see PPC rise again.
     
     The trick here is to find the right balance.
     
     If you're curious how much traffic Adsense probably needs to make smart pricing decisions, consider that they probably have somewhere in the neighborhood of 75% conversion pixel coverage and then use an online statistical significance calculator to see how many clicks from each country you would need to send them for them to be 85% confident in a decision.

     Reply
   • adsensjm says:

     Thanks for the explanation but this still isn't fair to publisher B... Publisher B runs the same ads as A, takes the same chance of annoying their own users as A, and otherwise runs the same risks as A for less gain (half in your example.) It's the advertiser's job to turn a click into a sale, not my job- my job is to display the ad- nothing more.

     
     Publisher A could have a site for ready to buy products. (most likely to convert) Publisher B could have a site for information about the products (not likely to convert - browsers)
     
     Makes sense for publisher B. to be smart priced.

     Reply
4. admin says:

   Thank you for your insights ezman,
   
   helped me a lot to better understand the whole topic of Smart Pricing.
   
   Great contribution.

   Reply
5. Blandt says:

   Make sense ... It seems to be a good thing then :) Please keep your insight coming ezman

   Reply