Companies look toward ever-more telling analytics
As analytics companies and agencies release better ways to measure media effectiveness, two of the
hottest buzzwords of 2008 are "integration" and "attribution." The days of the "last click wins" mentality are quickly coming to a close, and marketers are eager to measure each of their campaign
Attribution modeling is a flexible method for finding the value in all your marketing activities and determining their influence in your marketing mix. If you work
closely with your analytics provider, no doubt they've spoken to you recently about their "attribution" capabilities. What marketers need to keep in mind, however, is that these analytics tools
provide a framework for tracking and measurement, neither of which are fully effective without an attribution model based on your unique goals.
This type of tracking will balance the
scales between the tactics that "score the goals," such as PPC, SEO and CRM, and those activities that create "assists," such as display media, emerging media like widgets and gadgets, and even
There are two basic forms of attribution: click-based and view-based. Click-based attribution is the one many of us are familiar with. It's based on tracking
user clicks on a variety of advertisements and watching those clicks turn into some kind of action, such as a sale. But you can make this metric work for you, and it's worth your while to do so.
Here's how:Last Clicks:
Within a realistic time window (be it 30 days or seven days), look at all clicks from your various activities that touched the customer.Balanced Scorecard:
When a conversion event takes place, you can distribute the revenue that was created equally to each click-generating activity. For example, say you collected
$200 in revenue from a transaction, but five click events transpired within an established buy-time window. Each click event then gets $40 of "revenue credit" to be passed back to its "Activity
For identification of the activities that create long-term benefits for your business, perform a quick a/b test between those customers who were touched
by a specific marketing activity and those who were not. You can discern this by looking for changes in metrics such as average order value, number of items in the cart, conversion rate and visits per
visitor. These "hidden value" metrics are helpful in seeing the true return on your budget investment.
View-based attribution is a little trickier, and there will always be a margin of
error, but you can reduce the error when you have a common tracking platform across all activities - specifically one that tracks impression data, such as DoubleClick or Atlas. This type of
attribution allows you to understand the significance of your display advertising impressions as they relate to your overall marketing mix. Set a time window:
click-based attribution, view-based attribution requires a defined measurement period. However, when measuring impression- or view-based data, you need to make its time period separate from the
click-based aspect of your campaign. Typically, it will be shorter than the click-based model; if your click-based window is 14 days, limit your impression- or view-based window to seven days or
Keep in mind that some conversion events may be coincidental. Start with a longer view-based time window, as you will be able to learn more quickly what is truly affecting your
conversions. For example, look for aspects that affect your metrics, such as users deleting their cookies (comScore estimates that 30 percent of users delete their cookies monthly) or the user being
driven by media not measured by your online analytics suite.
By delving into the real life cycle of your campaign - rather than living and dying by a single aspect - you will see a
clearer, more accurate picture of how your marketing dollars affect your ROI. After all, you may have clicks that look like the "goal scorers," but behind many high-converting search terms or e-mail
promotions is a team of impressions that helped you make the sale.Dustin Engel is director of strategy and research at Range Online Media. (firstname.lastname@example.org)