Search Engine Guide: Can You Trust Your Paid Search Reports?

Share this article:
Paid search allows advertisers to match sales to corresponding ad expense, buying more of what works and less of what doesn't. But do you really know what's working? You don't if your data are inaccurate. Reviewing performance data from search marketers over the past three years, our firm has found significant data errors in almost 25 percent of the programs.


So can you trust your search numbers? Conduct a data audit to see. The effort is modest. After collecting information from internal systems, from your search marketing agency (if any) and from the engines themselves, someone on your staff with decent Microsoft Excel skills can perform an audit in a few hours. Here's how:


Test orders. Placing test orders verifies that your system properly credits orders to search. Side benefit: You'll re-experience your checkout, leading to usability improvement ideas.


Open up Internet Explorer. Clear cookies. Take a phrase you are advertising on Google broad match ("widgets") and Google search for it with a test ID appended ("widgets test_1"). Click through on your ad and place a small order with multiple items. (Hint: "4444444444444448" is a bogus Visa number that passes checksum tests.) Next, delete your cookies and restart IE. Go directly to your site (don't click through on an ad) and place another order.


Repeat this sequence using Yahoo, generating two more orders. Repeat this sequence using FireFox for four more orders, for a total of eight.


Check only the first, third, fifth and seventh orders that appear in your paid search order detail report. Ensure item quantities, prices, searched phrase and advertised phrase are correct.


Order audit: Pick a day. Obtain a search order detail report for that day, showing timestamp, order total, items and IP for every order that day credited to paid search. Obtain a list of all Web orders from your site that day, showing timestamp, IP, order total, items and - most important - marketing channel to which your site credited the order (unknown, search, e-mail, affiliate, portal, etc.).


Compare the two lists. Does the search order detail report contain orders that your Web site did not see or orders that your site assigned to other channels? Are some orders being counted multiple times? Is search taking credit for orders beyond a reasonable click-to-order interval?


Expense audits. Pick a calendar month. Obtain authoritative ad costs for the month from actual Google and Yahoo invoices. Obtain authoritative daily ad costs for the same period from the engines themselves (easily obtained from their public interfaces - log in yourself or ask your agency). Finally, obtain monthly costs from your weekly search performance reports from your agency or in-house team.


Compare these three sets of costs in aggregate and by day. For the month, they should differ 1 percent or less. Remember, data from the engines are authoritative. Your other reports must tie to those.


Running such audits semiannually ensures that your search marketing data are solid. Additional audit tips are available at www.rimmkaufman. com/whitepapers/audit.


For more articles from The Direct Marketer's Essential Guide to Search Engine Marketing, visit www.dmnews.com/search .


A PDF of the guide is available at: http://www.dmnews.com/pdffiles/semguide.pdf


Share this article:

Follow us on Twitter @dmnews

Latest Jobs:

Featured Listings

More in Digital Marketing

News Byte: CX Scores to Take Their Place Beside Price Listings

News Byte: CX Scores to Take Their Place ...

E-commerce aggregator PriceGrabber will begin offsetting price info with service expectations.

Data Byte: Interactive Ad Revenues Exceeding TV for the First Time

Data Byte: Interactive Ad Revenues Exceeding TV for ...

At nearly $43 billion, interactive advertising revenues exceeded broadcast for the first time in 2013.

Marketers: Data Rich and Knowledge Poor

Marketers: Data Rich and Knowledge Poor

While advertisers have become incredibly data-savvy, the most difficult challenge remains causally linking that data to outcomes that really matter.