CPA vs ROI Optimization…What’s The Best Practice?

March 12, 2010

Whenever I’m asked to audit a PPC account, my first question to the client is: how do you measure performance? It’s a simple question and the usual response is CPA (cost per acquisition) or ROI (return on investment). If the answer is CPA, then there are some follow-up questions: how many products or services are offered? And if more than one, then do they cost the same, or are they worth the same to the advertiser? What I’m getting at is that CPA is usually is a limitation of tracking – the advertiser cannot associate true revenue with the conversion. Managing to an ROI, and moreover, managing to maximize profit margin for each keyword will open up new doors for PPC efficiency.
Clients that manage to CPA generally have several products or services. To see the underlying issue with CPA optimization, consider the situation where we have two possible conversions: one worth $20 and one worth $40. If we optimize to an avg CPA of $30, we may be only selling $20 products and losing money. With ROI optimization, this is not a possibility. If CPA optimization is the only possible route, then there are a few best practices to find a good CPA to set as a goal. If the transactions occur online, then calculate the average order value and subtract the avg variable cost to the advertiser – this is the max CPA. If the online conversion is a lead which requires offline sales, then multiply this CPA by the average offline conversion rate to define the eCPA (effective CPA). Read the rest of this entry »



Biases as a Barrier to Good Business Decisions

February 18, 2010

It has come to my attention that when selecting a PPC provider, a good amount of marketers have what I would refer to as a “blind bias”. The blind bias refers to comments such as “we like Agency X because they seem to be an industry-leader” or “we want to work with Agency Y because we are familiar with them” or “we really like Company Z because they really dominate the PPC space.” I have to wonder, what do these comments even mean? While there might be something to be said for the idea of familiarity as it invokes a certain amount of trust, there isn’t much to be said for allowing basic hearsay to drive crucial SEM decisions. I call this type of bias blind because that’s what it is. There are no numbers or authentic experience attached to these claims, yet they seem to drive the PPC selection process for many advertisers. Read the rest of this entry »


PPC vs SEO

January 20, 2010

Why pay for clicks when you can get them for free?

This question often comes up when facing relatively inexperienced advertisers deciding between PPC and SEO. While I cannot refute any argument based solely on this logic (organic clicks are free and paid clicks are not), I can make a case for PPC even when SEO is at its best – which certainly is not the case for the majority of sites out there.
There are three distinct advantages that paid listings have over organic listings, and as a result these are the three most compelling arguments for integration of SEM into any marketing scheme:

1. If you don’t use SEM, your competition will poach your traffic. Read the rest of this entry »


It’s your Algorithm!

November 6, 2009

Most SEM professionals will agree that a solid SEM program strategy is multi-faceted. The program should focus not just on bidding, but also on Quality Score, data gathering and reporting, and aggressive testing of all aspects of the program. That being said, the great promise of bid automation remains a viable, debatable, and important topic.

There is no doubt that autobid algorithms are important to a well-rounded SEM bidding strategy but they can also be detrimental. Success with autobid depends on a number of factors. Obviously, one crucial factor is how well constructed the algorithm is. Perhaps more important, though, is the level of customizability made available by the algorithm. Does it treat core and long-tail terms similarly, or is it specialized for success with a particular kind of keyword? Is it customizable in terms of defining quantity of data analyzed, statistical relevancy, convergence speed, and data tolerance leading to a bid change? Can you influence bid security, day parting, data aggregation? Find out more!