There is a growing need to automate pay per click performance review, since most of activities are routines and error prone. While Google and Yahoo provide ton of data, someone has to evaluate data to see whether account performance is in line with expectation.
If you are a serious search engine analyst, you probably would have spent many hours monitoring pay-per-click performance.
Basically, you monitor performance for two reasons. First, you want to identify external factors that affect your business, including keyword bid escalation by known competitors, offerings by new competitors, bidding strategy changes by other bidding tools as well as seasonal changes. The second reason is that you want to track the impact of the changes you've made to your own campaign including bid pricing, ad rewrites, landing page changes, campaign restructuring and redeploying the scope of your keywords.
From our experience, this task of performance monitoring is very demanding as performance or trend changes can be subtle, unpredictable and easy to miss. Unless you review and analyze your data daily, by the time you see the change it may be too late to respond effectively.
Let's examine two separate examples of trend changes. The first occasion was due to internal changes. One of our clients was in the lead generation business, selling these leads to business owners. Some customers had complained about the limited information in the leads and our client changed the signup form to include twice as many questions. Unfortunately, however, they did not tell us about the change. We have a daily routine to check basic performance criteria such as daily spending, click counts and conversion count. Since daily fluctuation between this client's conversion counts is proportionately quite large (between 2 and 8), it took us almost a week to realize the resulting changes were significant.
The other example resulted from external factors. In this case, we did not confirm the pattern for a couple of weeks. In this case the daily conversion count of the campaign ranged from 0 to 2, making it hard to identify trends within short time frames. Even a week of very poor performance may not be enough to confirm a significant trend.
Drawing on this kind of experience, we've added new features to ClickSweeper, our pay-per-click performance management tool that detect real performance trends almost as soon as they occur, drawing on pattern-detection analysis of earlier time periods.
We don't know of this kind of feature being implemented in any vendors or major bid management applications as of yet. However, as more professional marketers and agents are pressured to improve profitability and the performance, this kind of feature will become essential to keep the cost of review minimum while having a reliable review process with a timely detection.
© Adcidia™ 2010 (Article99.com)
