The Publisher Opportunity in Search and Contextual, Part 2

The nuts and bolts of a highly-tuned publisher keyword campaign requires technology to track the positive outcomes deriving from every keyword-driven media click. The campaign’s objective is to maximize the holistic profit generated from the keyword traffic buy.

The positive outcomes can be derived from factors outlined in Part 1, but include maximizing page views per visit, as well as conversions to other profit-generating behaviors. Many publishers haven’t assigned values to a new newsletter subscription, but they generally know the value of a paid online or offline subscription.

Running a keyword campaign holistically may require that different departments at the publisher pool resources as well as metrics. Failure to consolidate metrics will result in a subpar campaign.

Yield managers seeking to drive online page views in specific site categories must collaborate with the offline circulation manager as well as the customer-acquisition manager seeking to build site registrations. Once the success metrics can be folded into a single formula, every page of the site must be pixeled if third-party systems and technologies are to be used.

The campaign manager’s goal is to isolate all the elements of a particular click that make it profitable and buy more clicks with similar profiles. Running the campaign based purely on averages is much more straightforward, but given that keyword inventory is auctioned in real-time and that APIs (define) exist to allow for niche selection of traffic sources, it would be folly not to buy the best identifiable traffic.

If we were to look at the advertising arbitrage aspect of this holistic ad buy (ignoring for the moment the additional profit-generating behaviors of a visitor), the break-even — and therefore the campaign’s profit — will depend on three factors: publisher revenue per thousand pageviews (RPM); the price at which inventory is available on a CPC (define) basis, and the click’s stickiness.

For example, if the publisher site is earning an RPM of $40 within a particular section of the site on which additional inventory could be sold, we can solve for the number of page views per visitor that need to be achieved to reach break-even at a given CPC (of inbound clicks). Alternatively, we could calculate the break-even CPC given a page view-per-visit metric.

For example:

• RPM = $40
• Earnings per page view = $.04 (divide the RPM by 1,000)
• Page views per visit = 2.5
• Break-even CPC = $.10

In an auction-bidding environment, not much inventory may be available at 10 cents, which is the break-even in this calculation. Plus, you obviously want to incorporate some profit.

The solution may initially seem to focus on keywords with lower bid prices and higher page views per visit. Clearly, if the number of page views per visit is higher among certain keywords from certain engines, your ability to earn money directly from your campaign is higher.

I urge you to use technology to go far beyond this level of simplistic campaign optimization. What other factors predict a high page view per visit? Tying back each user session generating page views to each inbound click will build a more complex and powerful model to re-allocate spending where it does the most good. Run an analysis to determine if the predicted number of page views per visit varies by any of the following factors:

• Syndication method. Search and contextual networks deliver different kinds of traffic. Which visitor type is better for your stickiness factor? Keep in mind that the contextual portions of the ad network may perform better with different ad copy than the search campaign. Searchers are in a different state of mind and surfers may respond to different messaging in a contextual targeting environment.
• Time of day. Is there a dayparting effect for site visitors? If so, you can choose either to cherry-pick the very best times of day or alter your bidding strategy to coincide with the optimal dayparts.
• Days of the week. The same way that dayparting changes typical expected stickiness, due to a difference in audience makeup, days of the week (in particular weekdays vs. weekends) may generate a different stickiness level.
• Geography. It’s rare to find publishers (and, for that matter, any search marketers) taking advantage of the dramatic differences in click quality (predicted stickiness and conversion) caused by differences in geography. By finding hot geos, one can bid more aggressively and still make a profit. Remember: just because you’re running geo campaigns, it doesn’t mean that running a simultaneous national campaign is a bad idea. A large chunk of searchers and surfers aren’t identified by geography in the search engine’s ad server. At the right price point, even national campaigns may work and deliver more volume due to the broader audience.

The same methodology to get the ad-driven profitability out of a keyword campaign should be expanded to include other success metrics once the value of these behaviors has been established. For example, if a free newsletter subscription is deemed to be worth $3 and the conversion rate of one segment of traffic is 0.5 percent, then the value of that click can be calculated at one and a half cents.

In this case, one could clearly not run a campaign exclusively based on newsletter subscriptions, but adding the penny to a bid could result in a higher position or greater rotation in a contextual campaign. Similarly, any conversions to more valuable offline paid subscriptions, even if highly infrequent, might add an incremental penny or two to the allowable CPC.

If you have an interest in a spreadsheet illustrating a model such as the one described above, I’d be willing to share a simplified version with you; just contact me.

Over the next year, I expect more old-school publishers to take advantage of their prestigious positions as brand advertising magnets and use online keyword marketplaces to build traffic and revenue. Doing so will require breaking down internal fiefdoms and adopting a holistic budget strategy, but those publishers heeding my words will prosper, whereas others will find themselves losing share of spend to more savvy publishers able to both establish themselves as magnets for brand dollars and demonstrate the acumen to take advantage of the huge disparity in CPMs between contextually-utilized remnant inventory and the premium ad space.

Join us for SES London February 19-21 and for training classes on February 22.

Want more search information? ClickZ SEM Archives contain all our search columns, organized by topic.