Squeezing More SEM Information From Web Analytics

Whether you're working on search engine marketing (SEM) in-house or at an agency, your ability to prove your worth is only as great as your ability to pull that proof from your analytics program. Today, let's examine two tips for more efficiently extracting SEM-based information from analytics programs.

Watching URLs as Well as Keywords

Everyone watches referring traffic sources to gauge the amount of search-driven visits to your site. Another popular report is the Referring Keywords report. That's appropriate because you need to be able to tell what types of terms are driving qualified traffic, not just driving traffic.

Watching the top landing pages is a similarly important, but is a less popular method to assess a campaign's success. If you can isolate the referring keywords that drive visits to specific URLs, you have a great deal of intelligence to use for site improvement.

In Google Analytics, begin at the Content | Top Landing Pages report. In the report body, click a URL or search for a specific page using the Find Page search box at the end of the report. Click the URL in the report and you'll see the full report on that URL's performance. In the Analyze dropdown box on that full report, select Entrance Keywords. The resulting report will show you the specific search phrases that brought people not just to the site, but to this specific URL.

To find the same data in Omniture SiteCatalyst, start at the Content | Pages report. Click the green magnifying glass icon next to the page in the report, which results in a dropdown box. Select Break down by: Traffic Sources | Search Keywords.

In addition to tracking efforts across your entire URL set, watching this data is a very helpful way to detect and examine the effectiveness of site links on a results page for a given query. For example, for two distinct queries that each produces a site-links result, one query might result in 90 percent of all clicks going to the main link, while the other query might result in a fairly even split among the main link and the remaining site links.

Export More and Wonder Less

When people analyze their SEM campaigns, a common frustration about various analytics packages is their inability to export complete data sets. In Google Analytics, for example, you can export any report to various formats such as CSV (define) and XML (define), but the reports can show only 500 rows. This can make it very frustrating to analyze a keyword report with thousands of rows, because many people still export 500 at a time and stitch them together in Excel.

Bypassing the export limits and pulling all the data from a Google Analytics report is very easy.

Go to the Google Analytics report that you want to export. Take note of the total number of rows of data in your report. For example, a keyword report might say, "Search sent 40,195 non-paid visits via 12,240 keywords." In this case, the important number is 12,240.

Go to the URL field of your browser. At the end of the URL, add the following characters:

&limit=12240

The number after the equal sign should match the number of rows you want to export.

After typing those characters, hit "Enter" or otherwise tell the browser to fetch that URL. (In Firefox or IE, for example, you could click the green right-arrow at the right side of the URL field instead of hitting "Enter." In Chrome, click the blue right-facing triangle.)

Export the data as you normally would, and your export document should have all the data you requested.

Here are some important notes about this technique:

• Some people recommend simply using the phrase &limit=50000 to ensure that they capture all the data in a report. This works too, as long as your report has no more than 50,000 rows, of course.



• After you append the &limit=X parameter to your URL and hit Enter, your screen will not look any different. This confuses a lot of people, because they expect all 12,000 rows to suddenly fill the browser window. This does not happen. The data appears only on export, not in the on-screen report.

I wish I could take credit for finding this, but I read it on a blog about a year ago and can't remember where it was. So to whoever found it, thanks.

Conclusion

Don't be satisfied with aggregated keyword data; instead, split it down to the URL level. If you think you know exactly what URL maps to each query term, you might be surprised to see the facts. And feel free to share additional metrics or analytics tips in the comments section, below. I'll post the best ones in a future column.