One of the things that makes paid search a challenge is the degree to which minor changes in any given campaign element can add up to significant differences in the health of a given campaign. For
example, positional changes (say, moving from #3 to #1 on a SERP) create enormous, sometimes exponential differences in click volume. Subtle changes in ad copy, keyword (positive and negative)
selection, match type settings, landing page design, and call to action language can create large improvements (or declines) in campaign profitability. This is why obtaining optimum campaign
performance requires search marketers to practice a regime of continual testing.
Search engines do the same thing, of course. They make subtle changes, test, revise, and test again in a
continuous process of optimization. Subliminal changes made to SERPs that individual humans can barely notice can cause significant aggregate swings in financial performance. Moving elements a couple
of pixels this way or that can improve or degrade aggregate revenue and profitability.
advertisement
advertisement
In early August, Google revealed several tweaks of this type, the most notable being a reduction in
space between the right-hand column (in which paid search listings appear) and the organic results. In the same time frame, Google shortened the blue horizontal bars residing at the top and bottom of
its SERPs. Interestingly, at Didit, we noticed a small but noticeable (between 0.7% and 0.96%) increase in click-through rates for paid listings immediately after these changes were made. Hardly a
revolutionary change but nothing to sneeze at either, given how Quality Score rewards higher CTRs with reduced bid prices for the same position, and given that Google's paid clicks decreased
approximately 2% in Q2 v. Q1 2009 (according to its quarterly earnings announcement). The summer of 2009 also brought us Google's "caffeine" update, a subject I covered several weeks ago for
MediaPost. Subtle changes included what appear to be a favoring of video over image results, plus an added emphasis on keywords within domain names. Again, not exactly a revolution -- but a fairly big
deal for marketers running highly competitive campaigns.
Little things add up, and whether you're doing SEM in-house or with an agency, it's wise to apportion a share of your budget for
testing purposes. It's impossible to come up with a fixed percentage because each marketer is different, but 15% is a good benchmark to start with. Think of this 15% as R&D dedicated to finding ways
to make your campaigns more efficient. Obviously, if you can eke out a 15% efficiency improvement, your testing program will have paid for itself. More efficient campaigns translate to lower bidding
prices, thanks to the influence of Quality Score. A proper test regime will include keyword expansion, creative testing (especially on the power keywords driving the bulk of your traffic), campaign
reorganization, landing page testing, and offer testing.
This kind of extensive, never-ending testing may strike old-line advertising types as overkill, but it's the only way to ensure that
you're not leaving money on the table. Scientific testing of every campaign element and real-time (or at least near real-time) adjustments are what makes SEM different from prior forms of advertising
and marketing. Unless you're testing contiunally, you'll eventually fall behind your competitors, and that's one thing you can't afford in today's lean and mean marketing economy.