Efficient Frontier Insights

More Reasons Why Biddable Display is Similar to Search

As Efficient Frontier continues to dig deeper into the world of biddable display advertising, we're finding a lot of similarities to search. I have just posted an article on this subject on AdExchanger. Feel free to read and tell us your thoughts.

Justin Merickel

UK Search Engine Report: Q4 2009

 

In addition to releasing the Q4 U.S. Search Engine Performance Report yesterday, we also released it's sister publication for the UK. The UK Report covers trends for the sector based on data from our UK Customer Index. Highlights from the quarter include:

• Sequential Quarter on Quarter (QoQ) growth is up 27% and Year on Year (YoY) growth is up 16%. 
• Click through rates on all engines increased year on year (YoY).
• Google’s share of clicks rose two percentage points to 86.9% and while Google’s share of spend rose one percentage point QoQ to 87.6%, Yahoo! gave up one percentage point in spend.
• Bing held relatively stable at 3.3% share of clicks in Q4 vs. 3.7% in Q3. Bing also remained steady at 4.3% share of spend.
• UK also expects 15-20% range in SEM spend growth for the full year of 2010 (based on current economic conditions).

Download the full report here for more analysis.

U.S. Search Engine Report: Q4 2009

Yesterday, we released our Q4 2009 State of Search Report that covers trends for the sector. Highlights from the  quarter include:

• Both sequential Quarter on Quarter (QoQ) and Year on Year (YoY) growth for Search was up 6%.    
• Spend in the retail sector grew 17% YoY and 46% QoQ.
• Search ad impression volumes that result from retail queries spiked 90% YoY and built on four quarters of positive sequential growth.
• Google’s share of clicks rose to 74.4% and reversed a downward trend in share since the launch of Bing.    
• Yahoo! was the biggest loser in paid click share giving up a full 3.4% since Q3, moving from 24.4% to 21%.    
• Bing held steady at 4.6% in Q4 vs. 4.7% in Q3.
• After a strong Q4 and continued recovery in the overall economy, we expect 15-20% in SEM spend growth for the full year of 2010.

Still want more stats and analysis? You can download the full report here.

Attention search marketers: The Efficient Frontier Q4 2009 Search Engine Performance Report is coming soon...

Since the beginning of this year, three teams from two continents have been working feverishly on the latest U.S. and UK Search Engine Performance reports. We already have some interesting conclusions to our analysis, so stay tuned! We plan to release it on Tuesday, January 19th. If you want this report to be delivered to you via email, sign up for the Search Engine Performance Report at the bottom of our homepage: www.efrontier.com

Multi metric bidding: The importance of tailoring bidding for all parts of your conversion funnel

Search marketing is all about getting conversions at the right price. But valuing those conversions is the art and science that separates a good SEM strategy from a bad one. A lot of attention that is paid to the conversion funnel that leads to an ad click but not enough is paid to the conversion funnel after the click. This insight is essential to determining what you’re willing to pay for those ad clicks. This post is a companion to a recent post Dr. Siddarth Shah on how such a multi metric strategy can actually be implemented.

For most search advertisers the initial conversion after the click is just another step in the funnel. Beyond applications there are approvals, beyond leads there are signups and enrollments, beyond purchases there is lifetime value. Generally speaking, there are initial conversions that lead to some ultimate conversion. All search advertisers can (and should) track how each click progresses from initial conversions to ultimate conversions. With this information you can begin to figure out how much each step is worth and how it influences RPC.

Ideally you would focus on ultimate conversions: like approved applications, confirmed subscriptions, or enrolled students. Unfortunately these ultimate conversions are often infrequent or have a long lag. So, initial conversions at the top of the funnel are needed as a proxy for the ultimate conversion. This is essential for tail terms which get infrequent traffic but are the best opportunity for incremental lift. For example if your conversions are applications and approvals, the applications will be more common than approvals and so will be a useful proxy metric for approvals.

The graph below shows the number of keywords with initial and ultimate conversions for a typical client. Light squares are initial conversions only; dark squares are ultimate conversions. Including the initial conversions more than doubles the number of keywords with 'revenue coverage'. The trick of course is deciding how much weight to place on these early funnel proxy conversions.

Keyword Coverage Map

Each square represents one keyword.

Do those light green cells indicate potential for ultimate conversions and if so what is the likelihood of conversion? If an application has a 50% chance of leading to an approval it’s worth half of an approval. So if you have a $10 cost-per-approval target you might value an application at $5. This is pretty easy if the conversion rate is always the same.

Naturally it’s not that easy. There is a huge variation in post-click conversion rates across keywords. In general an application or a lead from a brand term is much more likely to convert than one from a non-brand term. A targeted term is more likely to convert than a general term. There are any number of variations you can see on specific terms depending on your business and brand – for example keywords containing ‘fast’ will have a different conversion profile to ones containing ‘cheap’ or ‘best’.

Below you see another example from one of our clients that shows a stark variety in post-click conversion rates. This highlights the danger of assigning a fixed value to first level conversions. Some approvals are more valuable than others.


Valuing all applications at a flat $5 would not give you the best yield for your SEM spend. In Regions with low conversion rates you would be over-valuing leads. To get a more accurate proxy value you need to be more granular. Group your campaigns according to conversion rates after the click.

For some businesses there is an added layer of complexity. There are not only different conversion rates but different values for the ultimate conversions. A final conversion from California may have more value than a final conversion from Nevada for example. If this is the case just repeat the process just mentioned. It increases the number of metrics and estimates you need to make but is essential to making your spend efficient.


This presents a more complicated performance measurement and bidding strategy but also represents a significant opportunity to improve your budget allocation. Really digging into your post-click conversion data is essential to leveraging your SEM efficiently. For even more detailed techniques for managing your post-click conversion data see the recent post on Search Engine Land by Dr. Siddarth Shah, Director of Analytics here at Efficient Frontier: The What, Whys & Hows Of Multiple Metric Optimization.

Shay O'Reilly
Business Analyst