Ad Inventory Optimization – Breaking through the closed walls

Table of Contents Hide Frequency cappingDefaultsReal time optimizationSegmentationShort tail vs long tailControl Every content network has walked this rope at least once – ad inventory optimization. Direct sales team can…

Every content network has walked this rope at least once – ad inventory optimization. Direct sales team can only sell 40-50% of your ad inventory at any given time. The rest is filled with different ad networks, big and small. The math is very simple; push those ad networks which pay you higher CPM all the time. However, the truth is, no single ad network pays highest CPM all the time. It varies a lot depending on geography, time, page/site context, readers, etc. And to top it all, with strict frequency capping and default ads deployed,ad networks leave you playing with pennies.


Indirect sales can result in 30-40% of total ad revenue for any mid-size to large publisher. Today, for any smart publisher it’s must to optimize their indirect sales channel to boost the revenue. This becomes more challenging as almost allad networks are closed wall and have not (neither have intentions to) come out with APIs giving power to publishers.

Frequency capping

According to Wikipedia, Frequency capping is a term in advertising that means restricting (capping) the amount of times (frequency) a specific visitor to a website is shown a particular advertisement. This restriction applies to all websites that serve ads from the same advertising network.

In a competitive market where marketers are demanding more accountability, it’s obvious that ad networks have to optimize their ad delivery to boost CTR. However, this turns the balance in favor of ad networks. Publishers have little or no insight of how the pricing is structured for low CPM ads, and it converts into revenue loss.


Nothing frustrates publishers more than defaults. Just like it’s impossible for you to sell your 100% ad inventory in direct sales, so is for thead networks as well. Even Google Adsense, which today has the biggest pool of advertisers, can fill your ad inventory up to a certain limit.

Most ad networks allow you to specify tags from different ad networks to serve ads in case they don’t have any for your users. Publishers deploy static chains of ad networks pushing ads from other ad networks in case one defaults. These static chains are cumbersome to deploy or change. Publishers have no or little power to use their real time data to modify the chain whenever they want.

Some ad optimization networks like Pubmatic are offering dynamic daisy chains to publishers where default ad impressions are collected and then routed to the highest paying ad network.

Real time optimization

Yesterday’s home runs don’t win today’s games.” Babe Ruthoptimize baby

The goal of optimization is to maximize the yield of ad networks in real time. As pointed out earlier, this becomes daunting as ad networks are not ready to come out with their APIs. Yes, even the self-proclaimed “Don’t be Evil” giant shield its Adsense reports behind its own walls. You need either some grease-monkey script or advance data mining techniques to get into their systems and let you evaluate their performance based on CPM, click-through, conversions, defaults, etc. in real time and serve ads from best performing ad network for that case.


This for once, I will admit, Adsense is much better than any other ad network. They let you segment your ad zones very precisely. Their multiple-channel and url-channel tracking put publisher in driving seats; helping him optimize his inventory to any level. But again, the real challenge is to optimize Adsense with other ad newtorks and push an ad network only if it’s supposed to yield maximum output.

Ad optimization networks like Yield Build and Pubmatic helps you to some extent, but they are also tied with few ad networks only. Moreover, their system is best build for publishers doing 25 million and above pageviews per month leaving out small to mid-size publishers.

Monetizing international traffic is also a gem in the rough. Identify top countries sending you traffic and tie up with local ad networks to serve ads in those regions. No matter how good your current sets of ad networks are, they can never beat the local players.

Website targeting v/s page targeting – It’s no brainer that niche verticals pays you higher CPM (no I am not talking about sites dedicated to celebs). Every ad network identifies the page it’s serving ads on basis of context of the page, website or the user its showing ads for. If you have a large website divided into different channels, it’s worth finding out whichad networks are performing best of which sections or channels. Targeting the ad networks to the niche group will be instrumental in pulling higher eCPM than large, broad and general segment.

Short tail vs long tail

Thanks to Internet, today, life cycle of a content piece is more than few days. A content piece can continue to make money for anything between 12-18 months even after it’s published. Many contextualad networks performed (both in terms of CPM and CTR) much better than others when they receive traffic from search engines or various links pointing to your site.


boiler room
Who has the control between you and networks? Irrespective of what your account manager has told you, the truth is that you are standing on sidelines. Remember, no APIs means no control. If you really want to be in driving seat, you need to crash their ad network walls and get the reports in real time and make them compete with each other for every single ad impression.

Deploying an ad optimization engine is like passing away your control from one demon to another. This needs to be an in-house deployment giving you powers to tweak even the smallest things.

Not defying the credentials of Ad Optimization Networks
By voicing out these points, I am not defying the credentials of the already running ad-optimization platforms, if they work well for you nothing like that, I have heard great stories from people who use them but as the saying goes, one size never fits for all.

What’s your opinion?

[Guest article by Ankit Maheshwari, Co-founder of InstaBlogs and blogs at InstaMedia]

Sign Up for NextBigWhat Newsletter